Geoimage 2007 Brochure



Geoimage 2007 Brochure
Satellite Imagery Applications
© DigitalGlobe 2006
Now an official reseller of QuickBird Imagery
Very High Resolution Satellites
VHR Satellite Applications
How to order VHR Data
ALOS Pakistan Case Study
INDIAN Satellites
RADAR Satellites
Geophysical Data Processing
Advice on DEMs
Advice to Farmers
Exploration case history
Example imagery at scale
Dear Colleagues,
Welcome to this 2007 update of the GEOIMAGE Satellite Imagery Brochure. This year’s brochure is not intended as a complete update of the 2005 Brochure
and copies of this book are still available from your closest GEOIMAGE office or digitally online.
The big news for GEOIMAGE this year is our appointment as an official distribution partner of DigitalGlobe enabling us to resell DigitalGlobe’s high resolution
satellite imagery to clients throughout Australia, New Zealand, Papua New Guinea and the South Pacific.
DigitalGlobe is the only company operating a constellation of sub-metre commercial imaging satellites - QuickBird and WorldView-1. QuickBird is the
world’s highest resolution multispectral satellite delivering 0.6m pan-sharpened and 2.4m 4 band multispectral imagery and WorldView-1 is the highest
resolution, most agile commercial satellite ever flown, delivering 0.5m panchromatic imagery. WorldView-1 was successfully launched on Sept 18th, 2007
and is expected to be fully operational in early 2008.
DigitalGlobe’s updated and growing ImageLibrary contains over three hundred million square kilometres of satellite and aerial imagery suited to countless
applications for people who map, view, navigate and study the Earth.
In keeping with our policy of offering the widest range of satellite imagery for all applications, GEOIMAGE will continue to meet the demands of the marketplace
with data from all the commercially available satellites.
For over 19 years, GEOIMAGE has been providing and processing spatial information and our team includes highly qualified staff that use their technical
skills and experience to help you obtain your goals. We pride ourselves on keeping up to date with the latest that technology has to offer and will continue
to evolve with the ever changing needs of the spatial industry.
We look forward to hearing from you soon.
Bob Walker - CEO
A common theme that we have tried to address in this brochure is how to decide what is the best satellite imagery for your application. This is a complex question
and we hope that you will use the information in this booklet in conjunction with our 2005 Brochure to help you decide. If you have any questions concerning any
of the datasets please ring one of our staff for advice. GEOIMAGE will perform free data searches of your area of interest taking into account all your constraints. All
we ask is that you give us as much information about your requirements, including if possible - pixel resolution, panchromatic or multispectral, time contraints (will
archive data be sufficient or is new capture required), the area of interest and the application. The area of interest can be either a MapInfo TAB or ArcView Shape
file, the corner coordinates of a rectangle in lat/long or the centre of an area and a radius. The results of the archive searches are presented as image quick looks
and mapups usually in MapInfo and these will be emailed to you with a price quotation for any alternative solutions and a time table for delivery.
min new capture 64sq km
25.40 - 28.60
30.30 - 33.50
min archive 49sq km
20.20 - 35.20
min new capture 100sq km
20.20 - 35.20
21.20 - 37.60
35km by 35km
70km by 70km
35km by 35km
27km by 27km
142km by 141km
23km by 23km
370km by 370km
60km by 60km
1.81 - 3.53
60km by 60km
1.07 - 1.51
60km by 60km
2.83 - 5.70
60km by 60km
0.69 - 151
60km by 60km
3 and 4
0.69 - 151
60km by 60km
3 and 4
1.42 - 3.07
60km by 60km
60km by 60km
0.01 - 0.04
SPOT 2, 4
min archive 25sq km
25.40 - 28.60
60km by 60km
180km by 180km
180km by 180km
180km by 180km
0.01 - 0.03
6deg by 5deg
6deg by 5deg
6deg by 5deg
*Pricing arbitrarily based on 100 sq km
Actual costs will be dependent on how a clients area of interest relates to the standard paths of the satellites except in the case of the high resolution satellites.
All pricing subject to change.
0.5m resolution WorldView-1 Addis Ababa 05 October 2007. © DigitalGlobe
On 24 September 1999, a new era in commercial satellite remote sensing
commenced with the successful launch of Space Imaging’s IKONOS satellite from
Vandenberg Air Force Base, California. IKONOS was the world’s first high-resolution
satellite to offer one-metre commercially available Earth imagery. This was soon
followed by the launch on 18 October 2001 of the QuickBird satellite by DigitalGlobe
which still offers the highest resolution multispectral imagery available in the
commercial market. On 26 June 2003, OrbView-3 was launched by ORBIMAGE
however very little data over Australia is available from this satellite. The Australian
archive from the QuickBird and IKONOS sensors continues to increase and data
may be purchased from the existing archive or from new capture.
DigitalGlobe was founded in 1992 as an Earth imaging
and information company. That same year, it was the
first company to receive a high-resolution commercial
remote sensing license from the U.S. Government under
the Land Remote Sensing Policy Act. With the launch
of its QuickBird satellite in October 2001, DigitalGlobe
became the world’s highest resolution commercial
satellite imagery provider. Today, DigitalGlobe employs
more than 350 people and is leading the industry into
the next generation of commercial satellite imaging
with the planned launch of its WorldView satellite.
DigitalGlobe’s investors include Ball Aerospace &
Technologies, Corp., Hitachi Ltd., Morgan Stanley and
Telespazio S.P.A/Eurimage Investment.
Because DigitalGlobe was founded by geographic
information systems (GIS) and mapping users, the
company understands its customers’ unique needs
for geospatial information products for a number of
mapping and planning applications. While DigitalGlobe’s
formidable leadership has ensured solid financing,
world-class management and legendary marketing
vision, the company’s customer-friendly culture,
positions DigitalGlobe as the industry’s “go-to” provider
of geospatial information products.
WorldView-1 was successfully launched on 18th
September 2007, becoming the world’s first commercial
satellite to snap pictures of the Earth at 50 centimetre
resolution. The WorldView system includes more efficient
image processing systems and multi-satellite collection
planning, shorter tasking timelines, and an expanded
network of remote ground terminals. DigitalGlobe’s
imaging constellation combining WorldView and
QuickBird will be capable of collecting more than 4.5
times the imagery of any current commercial imaging
system. WorldView-1 alone will be capable of collecting
nearly 750,000 square kilometres (290,000 square
miles) per day of half-metre imagery.
450km orbit
680km orbit
Sun-synchronous orbit with
Equatorial crossing 10:30 a.m.
Sun-synchronous orbit with
Equatorial crossing 10:30 a.m.
In-track and cross-track pointing viewing angle.
Maximum off-nadir angle 30deg
Satellite Accessible ground swath 544km centred on the satellite
ground track
In-track and cross-track pointing viewing angle.
Maximum off-nadir angle 51deg
Accessible ground swath 900km
centred on the satellite ground track
Revisit 1 - 3.5days depending on latitude at 0.7m
Satellite orbit repeats every 140days.
Revisit 1-3 days dependent on latitude.
3days at 1m resolution at 40deg latitude.
1.5days at 1.5m resolution at 40deg lat.
4days at 1m resolution at Equator.
1.7days at 1.5m resolution at Equator.
Panchromatic 0.61m GSD (Ground Sample Distance) at
Black and White 450 to 900 nanometres.
Panchromatic 0.82m GSD (Ground Sample Distance) at
Black and White 445 to 900 nanometres.
Band4:Near Infrared 760-900 nanometres.
Multispectral 3.28m GSD (Ground Sample Distance) at
Band1:Visible Blue 445-516 nanometres.
Band2:Visible Green 506-595 nanometres.
Band3:Visible Red 632-698 nanometres.
Band4:Near Infrared 757-853 nanometres.
Data collected at 11bit.
Data collected at 11bit.
Nominal swath width is 16.5 km at nadir.
Nominal single image is 15 km by 15 km.
Nominal swath width is 11.0km at nadir.
Nominal single image is 10.5km by 10.5km.
Multispectral 2.44m GSD (Ground Sample Distance) at
Band1:Visible Blue 450-520 nanometres.
Band2:Visible Green 520-600 nanometres.
Band3:Visible Red 630-690 nanometres.
Image Area: The area of Interest (AOI) is a customer
specified geographic area. The AOI may be:
1. a rectangle in geographic coordinates,
Swath 2. an ESRI shape file supplied by the customer or reseller.
Minimum dimension is 5km by 5km. Areas wider than
Width and
15 km East-West or longer than 100 km North-South
Area Size may be collected as separate satellite images but are
processed and delivered together.
Minimum purchase from archive is 25 sq. km
Minimum area for new capture is 64 sq. km
Image Area: The area of Interest (AOI) is a customer
specified geographic area. The AOI may be:
1. a rectangle in geographic coordinates,
2. an ESRI shape file supplied by the customer or reseller.
Minimum dimension is 5km by 5km. Areas wider than
10.5 km East-West or longer than 100 km North-South
may be collected as separate satellite images but are
processed and delivered together. Areas of Interest larger
than 100 x 100 km are collected, processed, shipped
and invoiced as separate orders.
Minimum purchase from archive is 49 sq. km
Minimum area for new capture is 100 sq. km
GeoEye was formed as a result of the ORBIMAGE
acquisition of Space Imaging, which was completed in
January 2006. ORBIMAGE bought Space Imaging after
the latter put itself up for sale in early 2005, primarily
because it had lost its competitive position in the U.S.
remote sensing industry when it was not awarded an
important National Geospatial-Intelligence Agency (NGA)
NextView contract. GeoEye currently markets imagery
from the IKONOS satellite and will launch GeoEye-1 in
the first or second quarter of 2008.
Launched 18 September 2007
Resolution (Pan) at nadir (m)
.46 (.50)*
.46 (.50)*
.41 (.50)*
Resolution (Multi) at nadir (m)
Multispectral bands
New Bands red edge,Coastal,
yellow, NIR2
Schedule launch date
Orbital Height (km)
Nominal Swath Width (km)
* Imagery sold to commercial users must be resampled to 0.50m pixel size.
0.5m resolution WorldView-1 Yokohama 05 October 2007. © DigitalGlobe
DigitalGlobe’s QuickBird satellite provides the largest swath width, largest on-board
storage, and highest resolution of any currently available commercial satellite.
QuickBird is designed to efficiently and accurately image large areas with industryleading geolocational accuracy. The QuickBird spacecraft is capable of acquiring
over 75 million square kilometres of imagery data annually (over three times the size
of North America), allowing DigitalGlobe to populate and update its ImageLibrary at
unprecedented speed.
DigitalGlobe provides customers with flexibility in Product Levels, Ordering Options,
Delivery Options, and Licensing Options and these are discussed below.
QuickBird Imagery Products are available in three different product types.
> Basic Imagery is the least processed of the QuickBird product suite and is corrected
for radiometric distortions, internal sensor geometry, optical distortions, and sensor
distortions. Basic Imagery is neither geo-referenced nor mapped to a cartographic
projection. Basic Imagery is provided with the QuickBird sensor model and is intended
for sophisticated photogrammetric processing such as orthorectification. Basic Imagery
is a scene-based product, meaning that it can only be ordered in scene increments i.e.
16.5km by 16.5km at nadir. Basic Imagery is available as either black and white or
multispectral products.
Spatial and Spectral Resolution
Black &
450 to 900
Pixel Resolution
61 to 72 cm
2.44 to 2.88 m
27552 to
27424 pixels
6888 to 6856 pixels
450 to 520
520 to 600
630 to 690
Near IR
760 to 900
Scene Size
16.5 by 16.5 km (nadir)
Positional Accuracy
CE 90%
23 m
14 m
Black &
450 to 900
60 or 70 cm
> carry out other image processing steps such as pan-sharpening, mosaicing, etc
> DSM generation from stereo IKONOS imagery
> supply of digital data on DVD or for FTP download
> supply of normal or pseudo-stereo hardcopy prints
to obtain these or the client supplying them.
> Standard Tasking – For customers who require a new collection with a collection
window that allows more collection attempts than Priority Tasking. DigitalGlobe suggests
a 90 day collection window to ensure enough time to collect imagery which meets your
> Priority Tasking – For customers who require a competitive advantage over orders
placed as Standard Tasking order type and an accelerated delivery timeline.
> Rush Tasking – For customers who need their imagery collected on short notice and
delivered very quickly. Rush Tasking has the highest order priority and fastest delivery
timelines among all the commercial tasking types. Customers will receive an image
collected via Rush Tasking even if it does not meet the cloud cover or image quality
specifications. Those images greater than 20% cloud cover or Marginal or Poor image
quality will receive a 50% discount off of the originally quoted price.
Note that different tasking order types imply differences in the collection window, the
maximum number of acquisition attempts and the processing timelines.
Order Size
520 to
600 nm
630 to
690 nm
Near IR
760 to
900 nm
2.4 or
2.8 m
Natural Colour
60 or 70 cm
Colour Infrared
60 or 70 cm
4 band
60 or 70 cm
number of
CE 90%
23 m
14 m
> Orthorectified Imagery is a terrain corrected product, which is radiometrically
calibrated, corrected for sensor and platform-induced distortions, and mapped to a
cartographic projection. This product is GIS-ready and can be used as an image base
map for a wide variety of applications where a high degree of absolute accuracy is
required. Orthorectified Imagery is an area-based product, meaning that the product is
defined by your area of interest without reference to scenes. Orthorectified Imagery is
available as either black and white, multispectral, colour, or pan-sharpened products
which are identical to those for Standard Imagery.
Orthorectified imagery is supplied at various scales and accuracies and is dependent
on DigitalGlobe either having accurate ground control and DEMs, using contractors
Product Type
All Ortho
Order Size
450 to
520 nm
Positional Accuracy
> acquire the data and orthorectify to your datum/projection
Spatial and Spectral Resolution
Product Type
> arrange for the capture of new imagery
> ImageLibrary Orders – Imagery to suit your needs may have already been collected.
ImageLibrary orders feature fast delivery timelines and a smaller minimum order size
(25sq km) than tasked orders for Standard Imagery. Please contact GEOIMAGE for
more information.
> Standard Imagery is a geo-referenced product, which is radiometrically calibrated,
corrected for sensor and platform-induced distortions, and mapped to a cartographic
projection. Standard Imagery is provided with image metadata and is intended for a
wide variety of applications. Standard Imagery is an area-based product, meaning that
the product is defined by your area of interest without reference to scenes. Standard
Imagery is available as either black and white, multispectral, colour, or pan-sharpened
products. Standard Imagery products have a coarse DEM applied to minimise terrain
distortions. This degree of correction is relatively small, so this product is not considered
orthorectified. Alternatively, customers may choose Ortho Ready Standard Imagery which
does not have any terrain corrections, making this product suitable for orthorectification
by the customer. GEOIMAGE recommends its customers purchase the Ortho
Ready Standard Imagery Product.
> archive searches for QuickBird & IKONOS data covering your area of interest
cloud cover
1 scene
64 sq km
1 scene
64 sq km
100 sq km
100 sq km
Basic/Standard 10,000 sq km 10,000 sq km
Rush Tasking ImageLibrary
1 scene
100 sq km
1 scene
25 sq km
100 sq km
2,500 sq km
single pass
>24 hours
from order
10,000 sq km
10,000 sq km 10,000 sq km
10,000 sq km
>48 hours
>48 hours
All products
from order
from order
15 to 365
days from
7 to 365 days 1 to 14 days
All products start collection
from start
from start
date; 90 days collection date collection date
0 to 20%
0 to 20%
0 to 20%
As collected
As collected,
All Ortho
0 to 20%, 0% 0 to 20%, 0%
Basic/Standard 0 to 25 deg
0 to 25 deg
0 to 40 deg
As collected
Ortho (1:5000
0 to 15 deg
0 to 15 deg
0 to 15 deg
and 1:4800)
All other orthos 0 to 25 deg
0 to 25 deg
0 to 25 deg
All products
All QuickBird Tasking Orders must pass two feasibility studies prior to acceptance of
the order:
> Physical Feasibility assesses the number of times that QuickBird has physical access
to the target based upon the parameters the client provides. Items that affect physical
feasibility include off-nadir angle (wider angles will have more possible captures than
narrower angles), latitude (QuickBird has increased access to locations at higher latitudes),
collection windows (the larger the collection window, the more access QuickBird will
have) and cloud cover forecast. Note that orders over a 30deg off-nadir angle will
require a special review.
> Competitive Feasibility assesses DigitalGlobe’s ability to collect your order based upon
other orders in the system. Items that affect competitive feasibility include orders already
in the system and orders that have a higher relative priority.
Pan-sharpened enhanced natural colour QuickBird imagery at 1:5k scale Port Hedland Port area 06 May 2002. © DigitalGlobe 2002
Black & White (Panchromatic) Products
DigitalGlobe QuickBird black & white products enable superior visual analysis
based on 61cm resolution (at nadir) and 11-bit collected information depth.
The panchromatic sensor collects information at the visible and near-infrared
wavelengths (450nm -725nm). The output Ground Sample Distance (GSD) of black
& white products varies with product level. Basic Imagery products are delivered at
the GSD in which the data were collected (ranging from 61cm at nadir to 1.14m at
45deg off-nadir). For Standard and Orthorectified Imagery products, the customer
has the choice to resample to either a 60cm or 70cm GSD.
Bundle (Black & White and Multispectral) Products
The QuickBird satellite collects both multispectral and black & white (panchromatic)
imagery concurrently, therefore customers have the option to order both black
& white and multispectral products for the same area. When a customer selects
the ‘Bundle’ option, the products will be processed to the same product level, the
same product parameters and delivered as two distinct products (one containing
black & white imagery and one containing all four multispectral bands) with two
sets of associated Image Support Data (ISD) files.
Colour Products (3-band)
QuickBird Imagery Products are available in two 3-band colour product options:
• Natural Colour (using blue, green and red bands)
• Colour Infrared (using green, red and near-infrared bands)
These pan-sharpened products combine the visual information of three
multispectral bands with the spatial information of the black & white band. Colour
products are available as options for Standard and Orthorectified Imagery, but
not for Basic Imagery. Customers may choose between a 60cm or 70cm GSD.
Currently, Colour products are resampled using only the 4x4 cubic convolution
or the 8 pt sinc methods.
Port Hedland Port area. QuickBird Panchromatic imagery. 06 May 2002.
Approximate scale 1:10k © DigitalGlobe 2002
Multispectral Products
DigitalGlobe QuickBird multispectral products provide four discrete non-overlapping
spectral bands and 11-bit collected information depth. The multispectral products
cover the visible and near-infrared wavelengths in four bands. All four bands
are delivered as one file when this product is ordered by the customer. The
output Ground Sample Distance of the Multispectral Product varies with product
level. Basic Imagery products are delivered at the GSD in which the data were
collected (ranging from 2.44m to 4.56m at 45deg off-nadir). For Standard and
Orthorectified Imagery products the customer has the choice to resample to
either a 2.4m or 2.8m GSD.
Port Hedland Port area. QuickBird Colour Images.
Top image: natural colour. Bottom image: Colour Infrared. Pixel size 0.6m 06 May
2002. Approximate scale 1:10k © DigitalGlobe 2002
Pan-sharpened Products (4-band)
QuickBird Imagery Products are also available in a 4-band pan-sharpened product
option. These products combine the visual information of four multispectral
bands (blue, green, red and near-infrared) with the spatial information of the
panchromatic band. Pan-sharpened products are available as options for Standard
and Orthorectified Imagery, but not for Basic Imagery.
Port Hedland Port area. QuickBird Multispectral image. Pixel size 2.4m 06 May
2002. Approximate scale 1:10k © DigitalGlobe 2002
Customers may choose between a 60cm or 70cm GSD. Currently, pansharpened products are resampled using the DigitalGlobe pan-sharpening
kernel. The 4x4 cubic convolution and 8 pt sinc methods are available but are
not recommended.
Pan-sharpened enhanced natural colour QuickBird imagery at 1:5k scale Port Hedland Mangroves 06 May 2002. © DigitalGlobe 2002
Each order is defined by an order polygon which delineates your area of interest.
There are several ways to specify your order polygon. GEOIMAGE recommends that
you use Lat/Long coordinates in decimal degrees, based on the WGS84 ellipsoid for
all of these methods.
1. Specify the upper left and lower right coordinates if your area is rectangular.
2. Specify a center point and a height and width to define your area.
3. Email a shapefile to GEOIMAGE - .shp, .shx, .prj and .dbf files must be supplied
and the shapefile must contain only one polygon.
4. Email the coordinates in an ASCII text file (.gen), ArcInfo generate file format to
The order polygon can be defined with a minimum of 3 vertices and a maximum of 99
vertices. Due to terrain distortion, off-nadir viewing angles, and uncertainties in polygon
accuracy, we suggest that you buffer your order polygon by 200m to ensure that you
receive your entire area of interest. The minimum length for a side of an order polygon
is 5km for Basic and Standard Imagery and 10km for Orthorectified Imagery.
Off-nadir angles may be specified by the customer and in the case of BASIC and
STANDARD imagery is 0 to 45deg, within the following constraints by product type.
There must be at least a ten degree difference between the minimum and maximum
off-nadir angle specified. Please note that orders for Standard Imagery collected at
greater than 30deg are resampled to a resolution very different from the collection
The revisit time for QuickBird depends on the latitude of your area of interest and the
maximum off-nadir angle that you select. For example, an area of interest at 40deg
latitude would have a revisit time of approximately 7 days at 0 to 15deg off-nadir
Revisit time in days
Ground sample
distance (cm)
0 - 15 deg
0 - 25 deg
0 - 45 deg
and a revisit time of approximately 4 days at 0 to 25deg off-nadir. The revisit time
directly affects the amount of time DigitalGlobe requires to collect your imagery. Orders
specifying a maximum off-nadir angle of 0 to 25deg can be collected faster than orders
specifying a maximum off-nadir angle of 0 to 15deg. The off-nadir angle also affects
the ground resolution of the imagery so the maximum angle of capture is a trade off
between faster capture and larger pixel size (and more height distortion).
QuickBird collects data with an 11-bit dynamic range i.e. values in the range 0 to
2047. The 11-bit data allows more shades of grey than 8-bit data and allows details
to be seen in areas such as shadows. Because computers cannot read 11-bit data,
DigitalGlobe delivers the data in either 8-bit Dynamic Range Adjusted (DRA) or 11-bit
data stored in a 16-bit integer file with no stretching.
The DRA process as applied to QuickBird imagery is a visual enhancement applied
in two parts: colour correction and contrast enhancement. The enhancement is
strictly visual and does not affect the geographic location of the pixels. This product
is recommended for users who don’t have the tools to apply visual enhancements
to QuickBird imagery but is not recommended for those users intending to perform
scientific analysis or spectral classification on the data.
DigitalGlobe offers flexible licensing options to meet the client’s needs based on the
number of client groups requiring access to the data. Clients must select the license
type and identify the organisations that will be using the data at the time of order
placement. The Base Licensing options includes up to five permitted customer groups
working on a joint project as defined by the DigitalGlobe Product End User License
Agreement. The License is not transferrable and non exclusive. Other Licensing options
are available for larger groups of clients. Please check with GEOIMAGE for details.
Customers are permitted to create derivative works for internal use, including
reformatting QuickBird Imagery Products into different formats or media from which
they are delivered, modifying the QuickBird Imagery Products through manipulation
techniques and/or the addition of other data and making copies of the resulting
bundled image product.
Customers are permitted to use the imagery data provided by DigitalGlobe to prepare
vector maps so long as the created vector maps do not include DigitalGlobe’s imagery
data, and to further distribute the created vector maps.
GEOIMAGE downloads QuickBird data from DigitalGlobe by FTP transfer, processes
the data to your requirements and supplies the data on CD, DVD or puts it on the
GEOIMAGE FTP site for you to download.
Product levels, options, ordering, delivery and licensing of the IKONOS data are
described in our 2005 Brochure and on the GeoEye website at
Major differences between QuickBird and IKONOS are:
> GeoEye maintains a network of Regional Affiliates (with their own ground
stations) who are involved in the collection, processing, marketing and
distribution of imagery within their capture cones. These groups have a higher
pricing compared to imagery captured by GeoEye using onboard recorders and
downloaded into the US.
> Licensing is different and is detailed in the IKONOS Product guide which
is available as a PDF on the GeoEye website at
> IKONOS is currently the only VHR satellite imagery that can be programmed
to capture in stereo mode for the production of Digital Surface Models.
> The strip width of the IKONOS imagery at nadir is 11km however in new capture
mode the satellite is capable of imaging multiple adjacent image strips during
the same overpass with the number of strips being dependent on the northsouth dimensions of the area of interest.
Because of the ability of capturing several strips of adjacent imagery on the one
overpass and programming conflicts, new captures from IKONOS are sometimes
faster than from QuickBird. Ask your GEOIMAGE contact to get a timing feasibilty
assessment from both operators.
IKONOS Product Levels at a Glance
Positional Accuracy
Elevation Angle
Stereo Option
Visual & interpretive applications
15.0 metres*
60˚ to 90˚
Standard Ortho
50.0 metres**
25.0 metres
60˚ to 90˚
Basic mapping projects
25.4 metres
11.8 metres
60˚ to 90˚
Regional, large area mapping and
general GIS applications
10.2 metres
4.8 metres
66˚ to 90˚
Transportation, infrastructure,
utilities planning, economic development
4.1 metres
1.9 metres
72˚ to 90˚
High positional accuracy for urban applications
2.0 metres
0.9 metres
75˚ to 90˚
Detailed urban analysis, cadastral
& infrastructure mapping
* Exclusive of terrain effects ** May be up to 75 metres CE90 in undeveloped areas with high terrain relief (e.g. Andes or Himalayan mountain ranges)
Pan-sharpened natural colour QuickBird Imagery at 1:10k scale Townsville Port Area. © DigitalGlobe 2006
For certain applications, high resolution satellite imagery offers a cheaper
alternative to or has a number of advantages over aerial photography. The reasons
for this are based on the following attributes of the satellite imagery:
> four spectral bands from visible blue to near infrared, providing the equivalent
of both colour and colour infrared photography.
> digital radiometric data which provides the ability to undertake spectral processing
(e.g. spectral classification) and modeling (e.g. crop yield estimating).
> imagery captured on a stable platform, therefore no distortion will occur around
the edge of a scene, as with an aerial photograph.
> a 1-3 day repeat revisit time, allowing regular updates and multi-temporal
analysis of the same area.
> can be calibrated or matched so that large area seamless mosaics can be compiled.
> captured as 11bit rather than 8bit giving better dynamic range than aerial
photography and airborne scanners.
> does not have the variable brightness within a single image that is usually
associated with photos.
> satellite operators may already have data in archive over your area.
> image capture can commence within 3 days of placing an order.
Local and State Government Applications
High resolution imagery provides powerful tools for urban and regional government
planning and management applications, including image base cadastral and map
updates, transportation analysis, asset management, environmental planning,
crime mapping and analysis, stormwater management, public safety and disaster
management, validating vegetation maps produced at coarser scales, and enabling
Internet based property information access.
Monitoring Progress of Development
Most local government areas along the coastal sections of Australia are undergoing
rapid urban development. Using multi date high resolution imagery, local
governments are able to monitor development and enable better management
and town-planning practices. Some private businesses are now realising the
benefits of being able to instantly examine an area on the desk top to make
informed business decisions without the necessity of visiting a site. For example
surveying/engineering firms that in the past would have visited a site before quoting
to provide services, are able to quickly assess the job context without leaving
the office. For any job where a geographical context is required high resolution
imagery can be a great help and time saver. In this age where time equals money,
the investment in imagery will quickly pay for itself.
Example imagery (above)is 1m pan-sharpened natural colour IKONOS of Cairns
captured in August 2004, overlaid with cadastral boundaries in yellow. Image
© GeoEye 2004.
11 May 2001
19 April 2002
Vegetation Clearing
Local governments use high resolution satellite imagery to monitor vegetation
clearing in remote or difficult to access areas. By using simple subtraction
techniques, local governments can pin-point areas where clearing has occurred
and cross check this against the permits issued to identify illegal clearing. State
governments have used IKONOS imagery in court cases to win successful
prosecutions by proving that clearing has taken place before the date that the
permit was issued or after the property was purchased by the current owner.
The imagery has been able to prove which owner, past or current, performed
the illegal clearing.
Example imagery is 4m multispectral IKONOS, courtesy of the Maroochy Shire
Council, and shows clearing of remnant vegetation between 2001 and 2002.
Imagery © GeoEye 2001-2
Real Estate Applications
The high resolution colour imagery from IKONOS and QuickBird is ideal as a tool
in the sales and promotion of real estate. Imagine being able to show a potential
client all the available houses for sale from the comfort of you own office. During
the process you would be able to highlight the advantages of each property,
showing the site’s proximity to beaches, golf courses, schools and shopping
centres, as well as main roads, busways and railway tracks. Similarly for the
promotion of a new housing development, the ability to put the new area into its
geographical context showing all the existing development and services would
be an ideal selling strategy.
Example image (above)is approximately 1:5000 scale over Sovereign Island, Gold
Coast. IKONOS image collected 31 May 2005. © GeoEye 2005.
Pan-sharpened natural colour QuickBird Imagery at 1:20k scale Townsville Port Area. © DigitalGlobe 2006
Mapping and Monitoring Crops
The application of high resolution satellite imagery in the agricultural arena is only
limited by the imagination. This case study, kindly provided by CTF Solutions, shows
the application of IKONOS data when identifying problems with farming practices in
agricultural crops.
Image1 is an IKONOS NDVI image of a barley crop. Blue areas show high NDVI and
healthy crops while the green-yellow areas are lower values of NDVI and less healthy
crops, or sparser croppings, and red areas are non-crop areas. Investigations into
the striping pattern in the field eventually led to the conclusion that it was caused by
an unequal distribution of fertiliser by the fertiliser spreader. When the spreader was
fixed, the problem disappeared. Images 2 a,b and c show the same paddock sampled
at different resolutions and indicates that to solve cropping problems of this nature
requires high resolution 1metre or less imagery. © GeoEye 2003.
Mine Site Monitoring
Aerial photography is often flown over mine sites to provide a base map for the
mining operation, and for monitoring and planning purposes. Imagery is also used
for environmental assessment of the site before, during and after mining activities
have commenced. Where mine sites are developed in remote, difficult to access or
dangerous areas, such as in China, Mongolia, the Middle East, Brazil or Laos, and
where it is difficult to obtain aerial photography, IKONOS or QuickBird imagery offers an
ideal substitute. The ability to produce Digital Elevation Models from IKONOS imagery
is also ideally suited to mining applications.
Example image (above) is 0.6m pan-sharpened natural colour QuickBird over the
Super Pit, Kalgoorlie. © DigitalGlobe 2005.
Shallow Water Mapping
Both IKONOS and QuickBird capture imagery in the blue portion of the visible
spectrum, which allows both these satellites to have shallow water penetrating
capabilities. The shallow water penetration depth is dependent on a number of
factors, including off-nadir angle of capture, collection azimuth and turbidity of
the water.
Above - A 15kmx20km natural colour Landsat image window providing an
overview of the Moreton Bay area.
Left - QuickBird natural colour image showing seagrass beds in Moreton Bay.
© DigitalGlobe 2004.
This QuickBird example is representative of the high resolution satellite imagery being
used for shallow water mapping in coastal estuaries and coral reefs in Australia,
the Pacific Islands and the Caribbean by the Centre for Remote Sensing and Spatial
Information Science at the University of Queensland. The spatial resolution, spectral band
placement and radiometric sensitivity of this imagery are suited for producing maps
of benthic cover types (e.g. seagrass, sand, macro-algae) and their level of cover and
condition. Maps containing this information are commonly required by coastal resource
QuickBird natural colour Pan-sharpened imagery at 1:10k scale. Kalgoorlie Super Pit. © DigitalGlobe 2006
management agencies for baseline inventory, design of marine protected areas and
monitoring the effects of extractive industries, natural disasters (cyclones, tsunami) and
human activities. The QuickBird image above showing the eastern banks of Moreton Bay
in Queensland, Australia, was captured at 09h56 on 31 July 2004 which was 59min after
a 1.7m high tide. The imagery was purchased by the ARC Linkage Project “Integrating
Natural Vision and Remote Sensing” and Coastal CRC and was processed by Prof. Stuart
Phinn from the University of Queensland.
Data Integration
Triple Plate Junction plc (TPJ) is a successful
gold mineral exploration company which is
listed on the Alternative Investment Market in
London. The company is the operator of a joint
venture with Newmont at the Pusamcap Project
in North West Vietnam (Newmont 49%: TPJ
51%). The joint venture is currently exploring
two exploration and one prospecting license in a
highly prospective alkaline volcanic and intrusive
complex. Numerous Fe-Ox-Cu-Au structures are
currently being exploited by artisanal miners and
deeper level Alkalic Porphyry style mineralisation
have been discovered.
The high resolution QuickBird imagery has been integrated with detailed airborne
radiometrics and magnetics, as well as factual geological mapping and geochemical
sampling. It has provided a critical backdrop for all these data sets and is particularly
important in outlining structural detail and areas of potential alteration as well as subtle
textural detail indicating different lithologies. This has enabled detailed geological
interpretation and rapid identification of potential exploration target areas.
The Pusamcap region is centered on the
Pusamcap Mountains, approximately 330 km
north-west of the capital of Vietnam, Hanoi..
This region has been poorly explored due to the very rugged and mountainous terrain
with steep sided gullies and valleys rising over 2,300 metres from 400 metres in the
adjacent valley floors. The area is commonly cloud covered most of the year. Landsat
7 ETM+ and Shuttle Radar Topography Mission (SRTM) 90m Digital Elevation Model
(DEM) data had been previously used but lacked sufficient detail for the next stage of
the planned exploration activities.
New capture QuickBird imagery over the Pusamcap exploration project was ordered in
May 2005 and successfully captured on 21st December 2005. This was surprisingly
fast considering the rainy season in Vietnam extends from May to October. An Aster
scene over the prospect area was also used to produce a 15m DEM that was used in
conjunction with surveyed ground control points (gcp’s) to orthorectify the QuickBird
The QuickBird image area was 16km by 31km and was captured during a single
overpass on 21st December 2005. The image on the left is a natural colour image
with bands 1 2 3 in BGR. The image shows the bimodal nature of the image statistics
with the bright lowland areas with high haze and the darker forested area on the
highland area. Normal contrast enhancement that uses whole area statistics results
in a compromise with poor image detail in both areas. The right image incorporates
two different techniques to improve the overall contrast in the image. The green
band incorporates a percentage of the infrared band together with the visible green
to better highlight the variation in vegetation in the image. Secondly a local area
stretch has been applied that seeks to maximise the contrast using statistics of local
regions in the image rather than the statistics of the whole image. This technique
is particularly useful where there are a few spatially distinct areas of different cover
types or where there is high level wispy cloud and shadow.
Imagery © DigitalGlobe 2005.
Left - processed QuickBird image, Middle - RTP airborne magnetics,
Right - geological compilation map.
Use of Data
The imagery was particularly important in the early stages of the project for accurate
location of drainage for BLEG sampling as well as location of numerous zones of
artisanal mining. A detailed airborne magnetics and radiometrics survey carried out
by Newmont on TPJ’s behalf in April/May 2006, in combination with initial rock and
soil geochemistry was also used extensively on a backdrop of the QuickBird image for
rapid follow-up and identification of exploration targets. The imagery has been especially
useful in the planning of field programs including logistics in very steep terrain as well
as providing a reference point for an initial environmental baseline study.
QuickBird image at approximately 1:15 000 scale with an overlay of coloured
10nT RTP magnetic contours showing the ease of location in ground followup of
anomalous areas. Imagery © DigitalGlobe 2005.
GEOIMAGE thanks Mr. Matthew Farmer of Triple Plate Junction PLC for
permission to use this case study.
Pan-sharpened natural colour QuickBird Imagery at 1:7.5k scale with RTP contours. © DigitalGlobe 2005
GEOIMAGE will perform free data searches of your area of interest taking into account
all your constraints. All we ask is that you give us as much information about your
requirements, including if possible - pixel resolution, panchromatic or multispectral,
timing requirements (will archive data be sufficient or is new capture required), the
area of interest, and the application.
The area of interest can be supplied to us either as a MapInfo TAB or ArcView Shape
file, the corner coordinates of a rectangle in lat/long or the centre of an area and a
radius. Our preference is for a MapInfo TAB or ArcView Shape file in WGS84 lat’long
projection. We then check for the availability of archive imagery on the DigitalGlobe and
GeoEye search engines. The advantages of archive imagery are that
> the data is cheaper compared to new capture
> it is available in a shorter time frame
> the cloud amount is already known (20% cloud on a new capture is seen as acceptable
and data must be purchased)
The most obvious disadvantage of archive imagery is that it does not show the most
recent changes.
At this stage we prepare a graphic of the area and send it to the client with the details
of cloud cover and dates of the archived imagery and pricing for the archived and
new capture.
to collect ground control which are then provided marked on the image in a MapInfo
TAB file with accurate XYZ coordinates. GEOIMAGE uses these points to orthorectify
the original raw data. GEOIMAGE has prepared a document which offers advice on
the collection of accurate ground control. In the case of new capture stereo imagery,
we recommend the client mark targets before the image capture and get such points
accurately surveyed.
Pricing includes a freight component which GEOIMAGE gets charged by the satellite
owners even though most transfers are by FTP.
Upon acceptance of the quote by the client, GEOIMAGE will accurately map up the area
and forward the final quote and copyright documents to the client for signature. Upon
receipt of these documents, GEOIMAGE places the order with the relevant supplier.
In the case of archived data, the time between the receipt of the final order and supply
of the final product to the client may vary from a few days to two weeks depending on
the time taken by the satellite operator to supply the data which is archived in the US.
In the case of new capture data, the time to supply can vary from less than a week to
several months and is dependent on the following factors1. other programming requests in the same area,
2. the amount of cloud in the area. The satellite operators will continue to capture imagery
until the cloud on the imagery is less than 20%. This level of cloud is recognised as
acceptable by all operators and the data must be purchased, and
3. the number of overpasses required to cover the area. The nominal swath width of
the QuickBird imagery is 16.5km at nadir and the width of the IKONOS imagery is
11km at nadir. The IKONOS satellite will often capture more than one swath in a single
satellite overpass.
GEOIMAGE keeps the client advised through the capture process and provides a
quicklook as soon as the capture is complete. Final data supply is on DVD or FTP
transfer in the datum and projection required by the client.
Results of an archive
search for QuickBird imagery
over the Brisbane City Council area.
The pricing will be based on the size of the area in sq km and most likely be for 3 and/or
4 band pan-sharpened 16-bit colour imagery. Various other products are available (see
Pages 4-6) however this product is usually the most cost effective. DigitalGlobe gives
the option of purchasing 3 or 4 band QuickBird imagery and most applications will only
require 3 band natural colour. Applications which involve the health of the vegetation
such as for environmental monitoring will benefit from the use of the near-infrared band
and 4 band data should be purchased for these.
> Pricing will normally be given for archived and new capture imagery remembering
that archive imagery is cheaper except in the case of IKONOS imagery that is less
than 6 months old at the time of the order.
> Pricing will also include orthorectification of the data. Many VHR images are collected
at significant angles to the vertical so correction
is necessary for height distortion (See Page 11).
This processed is termed orthorectification.
This is handled in three different ways and the
pricing will reflect this.
1. The data can be systematically orthorectified
using a satellite model and the ephemeris data
from the image capture and using the SRTM
DEM for the height control. Such imagery should
be accurate to within 20m to 30m excluding any
errors introduced by using a DEM with a 90m cell
size and z accuracies of several metres.
2. If the client can supply accurate ground control
points marked on existing imagery or which can
be readily identified on the imagery, then the data
can be orthorectified using either a client supplied
dem or the SRTM DEM. In such cases accuracies
of a few metres can be expected (excluding any
errors due to inaccuracies in the dem used).
Example of a target prepared by
3. The client may decide to use the systematically a client and the same target on
orthorectified imagery supplied by GEOIMAGE the VHR imagery.
Raw QuickBird imagery supplied
as 4 strips of imagery captured between
June and Sept 2002. The strips of imagery
were orthorectified to an airphoto base using
an ALS DEM supplied by the client.
Final natural colour mosaic
of the 4 QuickBird orthorectified
strips and supplied to the client as
individual strips of 16-bit imagery and as
an ECW mosaic of the full area. © DigitalGlobe 2003.
Pan-sharpened natural colour QuickBird imagery at 1:5k scale Brisbane area 17 November 2006. © DigitalGlobe 2006
Most data that GEOIMAGE purchases on behalf of clients requires some type of
preprocessing before we pass it on. Such processing might involve simply reading
the data from a proprietary “tape” format into a TIFF or ECW file that will load into
the client’s GIS or might be more complex and involve one of more of the following
> Geometric correction and orthorectification
> Image matching and mosaicing
> Image calibration
> Contrast enhancements
> Hardcopy and digital output in client specific formats
> DEM generation.
Some of these processes are explained below while others are treated elsewhere
in this booklet.
Example of mosaicing of four different dates of 0.6m pansharpened QuickBird imagery - Brisbane to Ipswich corridor,
Queensland. Image A - scenes before mosaicing and ImageB
- final mosaic. © DigitalGlobe 2006-07.
Geometric Correction and
Most types of raw satellite imagery require
some type of geometric correction or
rectification so that the image corresponds
to real world map projections and coordinate
systems. Geometric rectification improves the
horizontal positional accuracy of the imagery
by warping the imagery to match the client’s
vectors or accurate ground control, and is
suitable where the area is largely flat and the
imagery has been acquired from nadir (near
vertical) viewing.
together, GEOIMAGE often ‘cookie-cuts’ the resulting large mosaic into a series of
user-defined tiles, usually based on 1:50,000 map sheet boundaries. This allows
the client to use one smaller tile or file when working in a smaller discrete area to
avoid manipulating a large and often cumbersome file.
Sub-metre accuracies like this
For areas where there is undulating between image and cadastre are
topography, or the imagery has been acquired not possible without accurate
ground control and orthorectification
at a high angle to the vertical, or very high using an accurate DEM. IKONOS
accuracy is required, orthorectification is Gold Coast. © Space Imaging 2005.
necessary. Orthorectification is basically
rectification that incorporates a digital elevation model(DEM) to compensate for
topographic relief by allowing the vertical aspect to also be taken into account.
GEOIMAGE uses PCI OrthoEngine which has specific satellite sensor math models
to generate very precise orthoimages. Orthorectification is also usually required if
several images or scenes need to be mosaiced in order to ensure that the joins
are seamless. For both rectification and orthorectification, accurate ground control
is essential to produce geometrically accurate imagery.
Schematic diagram showing the
reason why orthorectification is
necessary to produce geometrically
accurate imagery.
GEOIMAGE often performs colour-balancing, mosaicing and tiling after
orthorectification of satellite imagery, particularly the higher resolution imagery
such as IKONOS, QuickBird and SPOT5. This is because IKONOS and QuickBird
image swaths are relatively narrow so many separate files are often required to
cover a large area of interest, but the resulting mosaic of the imagery creates a file
of several gigabytes in size which is difficult to use and must be cut up in smaller
and more manageable pieces.
Most of the VHR and medium resolution satellites now capture data such that there
is one high resolution panchromatic band and a lower resolution multispectral
mode where the pixel size of the multispectral data is a multiple of the pixel size of
the pan band. GEOIMAGE has perfected the technique of merging such datasets
to produce a colour image showing the better detail of the panchromatic image.
The client therefore receives a colour image, which is better suited to a range of
applications, such as vegetation mapping, or simply as a backdrop to vectors, with
the higher spatial resolution allowing the client to see more detail.
In most instances, the pan/multi imagery is captured simultaneously and is
co-registered and the imagery can be pan-sharpened or merged without prior
orthorectification. In cases where imagery from different satellites are to be pansharpened, such as merging the 25m multispectral Landsat5 imagery with 10m
panchromatic SPOT4 imagery, the images first need to be orthorectified to ensure
that they exactly co-register with each other. GEOIMAGE will recommend which
datasets can be pan-sharpened to ensure you get the best spatial and spectral
resolution possible.
Colour-Balancing, Mosaicing, Tiling
Colour-balancing, mosaicing and tiling are all services that make satellite imagery
easier to use, integrate into a GIS or make more aesthetically pleasing. GEOIMAGE
has many years experience in processing satellite imagery to integrate seamlessly
into our client’s existing systems, and is able to custom produce imagery formats
and products to our clients unique requirements.
Colour-balancing is the process of matching two adjacent and joining satellite
images, usually captured on different days, so that there is no discernable difference
in the colour between them. This process is vital to the process of mosaicing
images so that no join or seam is noticeable. Mosaicing requires that the imagery
is not only colour-balanced but also spatially co-registered to each other so that
the two or more images can be joined seamlessly. Once the images are mosaiced
Example of pan-sharpening of enhanced natural colour 0.6m QuickBird imagery
over an urban area. 1:5k scale. © DigitalGlobe 2003.
Pan-sharpened natural colour QuickBird imagery at 1:5k scale Brisbane area 16 June 2007. © DigitalGlobe 2007
PRISM Characteristics
Number of Bands
The Advanced Land Observing Satellite (ALOS) developed by the Japan Aerospace
Exploration Agency (JAXA) was successfully launched on January 24, 2006. The
satellite has three sensors i.e., two optical imagers(PRISM and AVNIR-2) and an L-band
Synthetic Aperture (PALSAR). The mission objectives of ALOS include cartography,
regional observation, and disaster monitoring. It has been designed to have a short
revisit capability such that in cases of natural disasters, ALOS will be able to capture
images of the disaster area with AVNIR-2 or PALSAR within a few days.
GEOIMAGE has been appointed as a Commercial Distributor of ALOS imagery
and authorised to sell Standard and Value-Added Products to clients in Oceania
(Australia, New Zealand, PNG and Pacific Islands), and Value-Added Products to
clients throughout the world.
Number of Detectors
Number of Optics
Base-to-Height ratio
Spatial Resolution
Swath Width
1 (Panchromatic)
0.52 - 0.77 micrometres
3 (Nadir; Forward; Backward)
1.0 (between Forward and Backward looking)
70km (Nadir only) /35km (Triplet mode)
Pointing Angle
Bit Length
28000 / band (Swath Width 70km)
14000 / band (Swath Width 35km)
-1.5 to +1.5 deg. (Triplet Mode, Cross Track)
8 bits
PRISM data is collected in one of 9 possible Observation Modes - of which Mode 1
- Triplet observation mode with 35km wide simultaneous forward, nadir and backward
views is expected to be the main one. This means that 2 cycles of capture of 46
day each are required to get complete PRISM coverage of any area. Mode 3 which
produced 70km wide nadir PRISM imagery (corresponding to an AVNIR-2 nadir scene)
is not expected to be a significant mode of operation.
Emergency response image of the Newcastle floods in June 2007. AVNIR-2 image
collected on 10 June 2007. © JAXA 2007. A points to the stranded coal carrier the
Pasha Bulker and B shows the interface between the brown flood waters and the
normal ocean waters.
Data Availability and Applications
The stereo PRISM data allows excellent quality high resolution DEMs to be developed
over large areas at relatively low cost (See page 16). Merging or pan-sharpening the
nadir PRISM data with AVNIR-2 imagery creates 2.5m high resolution multispectral
datasets including natural colour imagery.
The Standard Data Products are in CEOS format and includePRISM/
Uncompressed, reconstructed digital counts, with radiometric calibration coefficients
and geometrical correction coefficients which are appended but not applied.
Radiometrically calibrated data
Geometrically calibrated data
Uncompressed, reconstructed signal data, with radiometric calibration coefficients and
geometrical correction coefficients which are appended but not applied.
Range and azimuth compressed
Complex data on slant range
Multi-look processed images projected to map coordinates
GEOIMAGE recommends the Level 1B2R where orthorectification of the data is required
and level 1B2G (purchased as a TIF file) if only systematic rectification is required.
The ALOS mission features a systematic observation strategy which comprises prelaunch, systematic global observation plans for all three instruments. The strategy is
implemented as a top-level foreground mission and with a priority level second only
to that of emergency observations. Copies of these observation plans for PRISM and
AVNIR-2 are available at
ACRES is the Oceania Data Node for ALOS and as well as having a reception and
archiving role, will be distributing ALOS data to non-commercial customers. Customers
wishing to purchase ALOS data from ACRES will first need to register their proposed
use with ACRES so that they can validate and approve the use as non-commercial.
The Panchromatic Remote-sensing Instrument for Stereo Mapping (PRISM) is a
panchromatic radiometer with 2.5-metre spatial resolution. The instrument has three
independent optical systems for nadir, forward and backward looking to achieve
along-track stereoscopy. Forward and backward telescopes are inclined + and - 24
degrees from nadir to realise a base-to-height ratio of 1.0. It is expected that the
PRISM instrument will be capable of producing DEMs with errors of about 5m and
with a 10m grid spacing.
Greyscale ALOS PRISM at 1:15k scale Sydney area. © JAXA(2007) Distributed by RESTEC
Orthorectified 70km by 70km
10m resolution AVNIR-2 image & superimposed
35 by 35km 2.5m resolution orthorectified nadir PRISM image
over Sydney collected on 21st January 2007. © JAXA(2007)
The Advanced Visible and Near Infrared Radiometer type 2 (AVNIR-2) is a visible and
near infrared radiometer for observing land and coastal zones and provides better
spatial land coverage maps and land-use classification maps for monitoring regional
environment. The AVNIR-2 is a successor to the AVNIR onboard the Advanced Earth
Observing Satellite (ADEOS) launched in August 1996. Its main improvement over
AVNIR-1 is its instantaneous field-of-view (IFOV). The AVNIR-2 provides 10m spatial
resolution images compared with the 16m resolution of AVNIR-1. The higher resolution
was realised by improving the CCD detectors (AVNIR: 5,000 pixels per CCD, AVNIR-2:
7,000 pixels per CCD) and their electronics. Another improvement is a cross track
pointing function for prompt observation of disaster areas. The off-nadir pointing angle
of AVNIR-2 is + and - 44deg.
AVNIR-2 Characteristics
Number of Bands
Band 1 : 0.42 - 0.50 micrometres
Band 2 : 0.52 - 0.60 micrometres
Band 3 : 0.61 - 0.69 micrometres
Band 4 : 0.76 - 0.89 micrometres
Spatial Resolution
10m (at Nadir)
Swath Width
70km (at Nadir)
Band 1~3 : >0.25
Band 4 : >0.20
Number of Detectors
Pointing Angle
Bit Length
7000 / band
-44 to +44 deg.
8 bits
For more technical details on the ALOS imagery, including PALSAR, please see page
36 of the GEOIMAGE 2005 Brochure.
The AVNIR-2 data that we have seen to date is generally of very high quality although
there has been one instance where striping was present both along path and cross
path and a substitute scene had to be ordered. Some striping noise is evident in the
PRISM imagery and its removal has been identified as a future goal for JAXA/RESTEC
(See insert this page). Details of the geometric and radiometric accuracies of standard
products as assessed by JAXA are detailed at
> archive searches for data covering your area of interest
> acquire the data and orthorectify to your datum/projection
> DEM generation from triplet PRISM imagery
> pan-sharpening of the AVNIR-2 using PRISM data
> supply of digital data on DVD or for FTP download
> supply of normal or pseudo-stereo hardcopy prints
Data from the ALOS sensors has certain advantages over comparable satellite
imagery including1. The low cost of the data compared to other similar resolution data,
2. The presence of a Visible Blue band on the AVNIR-2 allowing the preparation of
natural colour imagery,
3. The triplet mode PRISM data has the capability to produce DEMs with a 10m cell
size and high Z accuracies,
4. The PRISM nadir imagery is captured as near as possible to vertical i.e. with
minimal height offset and
5. Quicklooks of new imagery are in the catalogue by the following day and there is
rapid turnaround on orders.
Disadvantages of the sensors are1. The PRISM images are quite noisy and there is some residual detector striping,
2. The nadir PRISM and nadir AVNIR-2 data are rarely captured simultaneously over
the 70km swath width of the AVNIR-2 sensor although there is some 35km wide nadir
PRISM imagery collected on the same overpass as vertical AVNIR-2 and
3. ALOS imagery is not captured on demand although this service may be introduced
after the initial 3 year capture program which finishes in 2009.
Like other very high resolution data, PRISM
imagery has Blooming and Smear artifacts
caused by highly reflective objects on the
ground surface (Image A). The Blooming
artifacts have a tail which is horizontal and
this compares with a path orientated tail on
similar artifacts in IKONOS and QuickBird
The PRISM imagery also suffers from
artifacts that can best be described as
“boxed noise” (Image B) and are due to
non-optimal on-board jpeg compression.
These artifacts are generally most visible
in areas of low texture.
Pan-sharpening is the process of transforming
a low spatial resolution multispectral image by
fusing it with a coregistered panchromatic image.
ALOS imagery lends itself to this technique
where the 10m resolution AVNIR-2 imagery
(B) is combined with the 2.5m resolution nadir
PRISM image(A) to give a 2.5m colour image
(C). Because of the normal acquision of triplet
mode PRISM where the path of the imagery is
only 35km wide and the 70km wide path of the
AVNIR-2, two overpasses of PRISM data are
required to achieve complete spatial coverage.
Merging of the nadir PRISM imagery with AVNIR-2
data which has been captured at off-nadir
angles should not be a problem as long as a
high resolution DEM is available. 1:20k scale.
Pan-sharpened natural colour ALOS AVNIR-2 at 1:15k scale Sydney area. © JAXA(2007) Distributed by RESTEC
DEM Generation
Lake Resources N.L. has been exploring for epithermal gold and porphyry copper-gold
deposits in Pakistan since its inception in 1997. One of the main prospects is the Koh-iSultan prospect in the Chagai region in northwest Balochistan. The Koh-i-Sultan tenement
covers the partially dissected remnants of a major Quaternary age compound andesitic
stratovolcano. Extensive zones of advanced argillic alteration with sulphur, barite and silica
deposits and stockwork-veined altered intrusives, are present. The area is prospective for
high sulphidation epithermal gold and porphyry copper-gold deposits. Anomalous levels
of gold and base metals are present in stream sediments associated with major alteration
zones in the vicinity of the Nawah Caldera at the south-eastern end of the Koh-i-Sultan
Complex. Rock chip sampling of altered areas returned significant gold values.
Ground control for the back and forward PRISM images was taken from the orthorectified
nadir PRISM imagery and the same points were also identified on the raw nadir image.
Independent of the source of the stereo pairs, PCI OrthoEngine’s methodology for creating
the DEM is similar and is described in detail by Toutin - http://www.ipi.uni-hannover.
de/html/publikationen/2005/workshop/032-toutin.pdf. The method involves selection
of several accurate GCPs in the stereo pairs, computation of a 3-D stereo model based
on the sensor parameters and refined using the GCPs, processing of epipolar images in
which the height parallax is maximised in the X direction, autocorrelation between the
epipolar pairs on a pixel by pixel basis and computation of XYZ cartographic coordinates
from the elevation parallaxes in a regular grid spacing.
The Problem
DEM generation was carried out using epipolar images that were produced at 5m
resolution to minimise the effect of noise in the PRISM data. With three obvious stereo
images, there are three possibilities for epipolar pairs and a combination of pairs was
found to produce a 10m DEM with almost perfect correlation and a low level of noise
spikes. This grid was run through the Geomatica DWCON program which prepares
the data for hydrological analysis and in particular fills small isolated depressions and
produces a flow accumulation image which was used to automatically generate vector
drainage lines. The infilled grid was used to generate 10m vector contours.
GEOIMAGE had previously supplied Landsat 7 ETM+ imagery and ASTER spectral imagery
and DEMs to Lake Resources as a base for regional exploration over the Koh-i-Sultan
prospect. To aid the next phase of exploration of the Nawah Caldera, more detailed
accurate base maps, including topographic contours, were required to plan drill access
into the area and to map the obvious zones of surface alteration and brecciation. The
highest priority area required to be covered was 11 by 11km surrounded by a 20 by
20km lower priority area.
The Solution
Initial thinking was that an IKONOS new stereo acquisition would be the ideal solution and
this would provide 1m resolution pan-sharpened colour imagery and a high resolution
DEM. It was found however that the area was located within a GeoEye regional ground
station and the cost of the IKONOS imagery was greater than double the normal
worldwide standard price.
A search of the ALOS archive showed that a triplet of cloud free ALOS PRISM imagery,
acquired on 17th August 2006, was available over the area and that cloud free vertical
AVNIR-2 imagery had been collected on 2 October 2006. This would allow the preparation
of 2.5m pan-sharpened imagery and a high resolution DEM. Note that although the AVNIR2 and PRISM imagery were not collected on the same date, the area is not subject to
any short term change (except perhaps snow and cloud) and because the imagery was
collected with similar ephemeris, image merging would not be a problem.
The extracted DEM is a surface model that normally includes the height of manmade
features such as vegetation and buildings. This is not a problem in this area but will be
a problem in areas of cultural disturbance and vegetation of variable heights.
The accuracy of the final grid is impossible to ascertain because of the lack of ground
control. Based on the local detail of the grid, the automatically generated drainage lines
and theoretical considerations related to the size of the pixel, the local noise is believed
to be of the order of 3-4metres.
GEOIMAGE would like to thank Jim Clavarino of Lake Resources N.L. for permission to
write up this case study. All the ALOS images are Copyright JAXA(2006) and distributed
The Methodology
The standard triplet of back, nadir and forward PRISM imagery and a non-standard
AVNIR-2 image (moved to the south to cover the PRISM data), were purchased at 1B2R
processing level. The data was read into PCI OrthoEngine using the newly developed
ALOS module.
Accurate ground control was only available for the immediate area of the Nawah Caldera
and so it was decided to systematically rectify the imagery using the Geocover 2000
Landsat Pan image for E and N control and the SRTM data for Z control and then to fine
tune the final orthorectified imagery using the client supplied ground control.
The AVNIR-2 data was orthorectified first and then the nadir PRISM scene was
orthorectified using ground control automatically generated in the Geomatica AUTOGCP
module. The orthorectified AVNIR-2 and nadir PRISM data were then pan-sharpened
using the Geomatica PANSHARP module.
Final orthorectified AVNIR-2 natural colour image and
superimposed greyscale nadir PRISM image.
© JAXA 2006
Pseudocoloured ALOS DEM
for the PRISM pair at a 10m cell size.
The elevations range from 770m to 2320m.
Box shows location of enlargements below.
Comparison of the 10m ALOS generated DEM (left) and the SRTM DEM (right) over
the Nawah Caldera.
1:5k scale
Port HedlandDistributed
May 2002. © DigitalGlobe 2002
Pan-sharpened natural
at 1:25k
© JAXA(2007)
Left image: Natural colour pan-sharpened ALOS imagery at 1:25k scale with drainage lines automatically generated from the 10m DEM.
Right image: Pseudocoloured ALOS 10m DEM with 20m black contours and red drainage lines automatically generated from the DEM
Stereo Hardcopy
One of the products that has been generated from the ALOS pansharpened imagery and the ALOS DEM is a set of pseudo-stereo
hardcopy prints at 1:25k scale. In these images, the elevation value for
each point is used to offset the image data at that point in an east-west
direction to simulate height distortion. The amount of distortion is usually
the same but in opposite directions to produce a left and right stereo pair.
For interpretation work however, it is usual to prepare a left image(with
twice the amount of height distortion) and a vertical image so that
interpretation is made on the undistorted vertical image.
Examples of the left stereo and vertical images
(superimposed) which were printed at 1:25k scale.
An example of photogeological
interpretation using a 1:25,000 scale
ALOS stereo pair is shown by the images
on the right. Part of the rim of a large
collapse caldera structure, partly infilled
by young sediments, can be seen on
the eastern side of the image. A smaller
caldera, about 1km in diameter can be
seen to the west. Extensive hydrothermal
(advanced argillic) alteration can be seen
as pale areas on the image (red crosshatching on the map), while unaltered
volcanics are dark (purple colours on
the map). Note the presence of reddish
young ferruginous sediments on the
ALOS image; these form low terraces
(shown as yellow on the map). The ALOS
stereo pair also enabled the recognition
of an important, previously unmapped
NNW-SSE oriented fault zone. Dr Colin
Nash of Colin Nash and Associates
Pan-sharpened natural colour ALOS AVNIR-2 at 1:25k scale Pakistan with 10m DEM contours from PRISM. © JAXA(2007) Distributed by RESTEC
One of the aims of the ALOS PRISM program was to produce DEMs (Digital
Elevation Models) with less than 5 metre errors and with 10 metre grid spacing
in support of the creation and updating of maps at a scale of 1:25,000. Software
vendors in the photogrammetric sector have been slow to develop modules
to process ALOS optical data because many rely on the availability of rational
functions to define the “camera model”. The rational functions have only been
released for PRISM Level 1B1 data where the data is still in separate CCD format.
This is not the case for PCI Geomatics OrthoEngine which uses Toutin’s satellite
model and which can process PRISM level 1B1 and 1B2R data. This software
processes the PRISM triplet as separate epipolar pairs and has been used to
produce the DEMs discussed here.
GEOIMAGE’s DEM generation procedure can be summarised as follows.
> orthorectify the nadir PRISM image using the best available ground control and
available DEM. The use of the SRTM DEM here is not a problem because of
the near vertical aspect of the nadir image.
> Select control points on the back/nadir/forward raw images based on the
orthorectified nadir image. Only points that can be identified on all three images
are used and obviously the more accurate the location of the points the better
the model.
> Epipolar images for each image pair are generated at 5m pixel spacing.
> Epipolar DEMS are generated at either a 5m or 10m spacing.
> If there is low level systematic noise in the DEMs resulting from poor control of the
model this is best corrected at this stage before the geocoding of the DEMs.
> The epipolar DEMs are geocoded and the final dem reading at any point on
the grid may be either based on an average of the individual DEMs or based
on the correlations score during the epipolar DEM generation.
> The final DEM is edited for artifacts in water, cloud and shadow areas or for
GEOIMAGE recently conducted a test DEM generation from a PRISM triplet over
Sydney to assess the accuracy of the DEM. The image used was collected on 21
January 2007 and was cloud
free. The NSW Government’s
PANRIIE SPOT 5 imagery was
used for X and Y control and
was used for Z control. The
ALOS DEM was generated at
5m cell size from an average
of the back/nadir and nadir/
forward DEMs. The ALOS DEM is a Digital Surface Model and includes the tops of
buildings and vegetation as shown in the images at the bottom of this page. The
comparison of approximately 1000 points between the NSW LANDS DTM and the
ALOS DEM showed a mean difference of 3.7m and an RMS error of 5m.
This error is in the same range as determined by K. Wolff, A. Gruen, ETH Zurich:
DSM Generation from Early ALOS/PRISM Data using SAT-PP and Junichi Takaku,et
al.,High Resolution DSM Generation from ALOS PRISM.
Many of GEOIMAGE’s client are working in relatively isolated areas where good
control is difficult to obtain and where the SRTM is the best DEM available. In such
areas, where we have used the SRTM for the Z control, it has been necessary
to do a considerable amount of massaging to take out high frequency noise but
believe the final DSM result is accurate to about 5m relative.
3m coloured contours derived from an ALOS PRISM pair on a natural colour
IKONOS (© GeoEye 2005) image. The SRTM DEM was used for the Z control in
this example. Approximate scale 1:20k.
Three image views of an area along Spit Rd, Spit Junction, Sydney showing the detail in the dem at a scale of approximately 1:10 000. Image A: Anaglyph presentation of the
back and forward PRISM images that have been orthorectified using the NSW LANDS DTM. Red and green fringes indicate features that differ in height from the ground. Image
B: Pseudocoloured ALOS 5m DEM. There is an elevation difference within the area of 2m to 95m and the large white high area in the centre is 23m above ground level. The DEM
generation was unable to separate the heights of the three white buildings that can be seen in Image C. Image C: Pan-sharpened natural colour ALOS image.
Pan-sharpened natural colour ALOS AVNIR-2 at 1:25k scale Sydney area with 3m DEM contours from PRISM. © JAXA(2007) Distributed by RESTEC
The launch of India’s first civilian remote sensing satellite IRS-1A in March
1988, marked the successful start to the Indian Space Programme. Several
satellites later, the current Indian satellites IRS-P6 (RESOURCESAT-1) and IRS-P5
(CARTOSAT-1) provide continuity of data and have catapulted the Indian Remote
Sensing Programme into the world of large scale mapping and terrain modeling
applications. Data from these satellite although mainly aimed at the Indian user
community are increasingly being made available worldwide.
RESOURCESAT-1 is the tenth satellite developed by the Indian Space Research
Organisation (ISRO) and was launched on October 17, 2003. The satellite operates
in a circular, sun-synchronous, near polar orbit with an inclination of 98.69deg, at
an altitude of 817Km. The satellite takes 101.35 minutes to complete one revolution
around the earth and completes about 14 orbits per day. The entire earth is covered
by 341 orbits during a 24 day cycle. The Equatorial crossing time in the descending
node is approximately 10:30am. The three sensors onboard the satellite ( LISS
-3, LISS -4 Mono, LISS -4 MX and AWiFS) provide imagery with varying spatial,
spectral, temporal and radiometric resolutions. The distance between adjacent paths
is 117.5Km and each sensor collects data with different swaths.
Spectral Resolution (m)
Swath (km)
Spectral bands (microns)
Quantisation (bits)
23.9 ( MX mode)
70.3 (PAN mode)
0.52 - 0.59
0.62 - 0.68
0.77 - 0.86
0.52 - 0.59
0.62 - 0.68
0.77 - 0.86
1.55 - 1.70
0.52 - 0.59
0.62 - 0.68
0.77 - 0.86
1.55 - 1.70
The LISS-4 sensor is a high resolution multi-spectral camera operating in three
spectral bands of visible green, visible red and near infrared. LISS-4 provides a
ground resolution of 5.8m (at nadir) and can be operated in either of two modes. In
multi-spectral mode (Mx) LISS-4 covers a swath of 23km (selectable out of 70km
total swath) in three bands. In panchromatic mode (Mono) the full swath of 70km
will be covered in any one single band which is selectable by ground command
(nominally in B3 - red band). The camera is tiltable within 26.5deg off nadir and
this allows coverage of any area from several paths and a maximum wait period
of 5 days to view an area at the equator and even less at higher latitudes.
AWIFS Image - Sydney area. Single sensor 370km wide strip.
CARTOSAT-1 was launched on 5th May 2005. The satellite operates in a circular,
sun-synchronous, near polar orbit with an inclination of 97.87deg, at an altitude
of 618Km. The repeat cycle is 126 days and the distance between paths at the
equator is 21.46Km.
This satellite carries two PAN sensors with 2.5m resolution and fore-aft stereo
capability. The platform is continuously steerable to allow both fore and aft cameras
to look at the same ground strip with a time gap of the order of approximately
52 seconds. Both cameras operate in the 0.5-0.85microns spectral band.The
payload is designed to cater to applications in cartography, terrain modeling and
cadastral mapping.The cameras can be operated either in real time mode by direct
transmission to ground station or in Record-Playback mode using a 120Gb On-Board
Solid State Recorder (OBSSR).
The fore camera provides an across track resolution of 2.452m (at nadir). It covers a
swath of 29.42km .The fore camera can be tilted up to +23deg in the across track
direction, thereby providing a revisit period of 5 days. The resolution of the camera
when tilted by 25deg is 2.909m, resulting in a swath of 34.91km.
The aft camera provides an across track resolution of 2.187m (at nadir). It covers a
swath of 26.24km .The aft camera can be tilted up to +23deg in the across track
direction, thereby providing a revisit period of 5 days. The resolution of the camera
when tilted by 25deg is 2.789m, resulting in a swath of 33.47km.
Temora NSW LISS-4 RESOURCESAT-1 1:25k scale.
The LISS-3 sensor is a non-tiltable 4 band (visible green, visible red, near infrared
and mid infrared) camera with a 6000 CCD detector for each band giving a swath
of 141km at a pixel resolution of 23.5m.
The AWiFS camera operates in four spectral bands which are identical to the LISS3 camera . The camera uses two separate electro-optic modules which are tilted
by 11.94° with respect to nadir. Each module includes a 6000 pixels linear array
CCD detector for each spectral band which images a swath of 370km providing a
combined swath of 740km with some overlap at nadir. The resultant pixel size at
nadir is 56m and at the extreme edges 70m. The CCDs used in AWIFS are identical
to those of LISS-3. The output signals from each CCD are amplified and digitised
into 10bit parallel data in the video processing electronics. The large swath of
the camera (740km) compared to the path spacing of 117.5km results in global
coverage every 5 days.
GA Building Canberra CARTOSAT-1 1:15k scale.
All imagery © ANTRIX Corporation, Indian Space Research Organisation (ISRO)
Alice Springs LISS-4 RESOURCESAT-1 28 March 2004 at 1:25k scale. © ANTRIX / ISRO 2006
Of the 5 SPOT Satellites launched since SPOT 1 in
February 1986, SPOTs 2, 4 and 5 continue to provide
worldwide image data. Full details of the SPOT Satellites
are contained on pages 25-27 of our 2005 Brochure and
from the Spot Image website at
Only a summary of the main features of the satellites and data are described here.
“All three SPOT satellites are operating nominally. SPOT 5 exceeded its lifetime in
orbit, continuing to acquire high quality images and is expected to provide commercial
imagery products for the next 5-7 years, as will SPOT 4. Originally designed for a
three-year mission, SPOT 2 is still sending back images of excellent quality after
almost twenty years of active service; it will be decommissioned within the next
12-18 months.” – Rob Lees , Managing Director of SIS, July 2007.
Important Features of the SPOT Satellites
1. The SPOT network comprises two main stations at Toulouse (France) and Kiruna
(Sweden). These stations are capable of receiving the data recorded by the on-board
recorders or directly within their visibility circle of approximately 2500 km radius
centered on the stations. There are also 22 direct receiving stations which can only
receive data within their visibility circle. Each DRS effectively manages its own visibility
zone according to the satellite resources allocated by SPOT IMAGE. Data within the
Australian area is captured by a receiving station located in Adelaide.
2. SPOT 4 and 5 have onboard solid-state recorders so that data anywhere
around the world can be captured and later downloaded to the Toulouse or Kiruna
3. SPOT’s oblique viewing capacity allows it to image any area within a 900 kilometre
swath and can be used to increase the viewing frequency for a given point during a
given cycle. The frequency varies with latitude: at the equator, a given area can be
imaged 7 times during the same 26-day orbital cycle, while at latitude 45 degrees,
a given area can be imaged 11 times during the orbital cycle.
The result of these features is that when combined with SPOT programming,
the SPOT constellation of satellites offer users assured image capture (weather
permitting) for ongoing monitoring programs or for timely capture of single events.
SPOT IMAGE is the only satellite operator to offer on demand image capture apart
from the VHR (sub 1metre) satellite operators .
SPOT IMAGE offers two programming services based on levels of urgency and
specific acquisition requirements.
The Standard Programming Service is best suited for applications that do not require
images to be acquired within specific time windows or at extreme viewing angles.
The price of programmed products includes programming service fees and applies
to images acquired with less than 10% cloud cover.
GEOIMAGE acts as an agent for SPOT IMAGE so can offer you raw data and
processed imagery includingarchive searches for data covering your area of interest
acquire the data and orthorectify to your datum/projection
processing of pan-sharpened and natural colour enhancements
seamless mosaicing of images
supply of digital data on DVD or for FTP download
supply of normal or pseudo-stereo hardcopy prints
SPOT imagery offers an optimum combination of resolution and spectral and spatial
coverage where Very High Resolution data is not required. Pertinent information
when ordering is
Nominal scene size is 60km by 60km and a half-scene (42km by 42km) is available
for SPOT5. Note the east-west dimension may vary up to 80km at maximum offnadir angles.
Levels of Processing
Three main levels of processing are available of which GEOIMAGE would recommend
the purchase of Level 1B to most clients:Level1A. Corrected by normalizing CCD response to compensate for radiometric
variations due to detector sensitivity. No geometric corrections.
Level1B. The same radiometric corrections as level 1A. Geometric corrections
compensate for systematic effects, including panoramic distortion, the Earth’s
rotation and curvature, and variations in the satellite’s orbital altitude.
Level2A. The same radiometric corrections as level 1A and 1B. Geometric corrections
to match a standard map projection (UTM WGS 84), without using ground control
points. For SPOT 5, a global DEM with a post spacing of one kilometre is used; for
SPOT 1 to SPOT 4, the mean rectification elevation is constant across the scene.
Archived Data
Archived data is usually less expensive than newly captured imagery however the
details of what constitutes archived data varies for the different satellites and it is
best to check with GEOIMAGE for the latest pricing.
Resolution and Spectral Modes
By combining imagery from all five SPOT satellites, it is possible to generate data at
4 levels of resolution between 2.5m and 20m in black and white, and colour, across
the same 60km swath. This multi-resolution approach offers users the geospatial
information they need at different scales and includes.
Pixel Size (m)
Data Type
Best Scale
SPOT 1, 2, 3 & 4
SPOT 1, 2, 3 & 4
The Priority Programme Service is particularly suited for applications that are
subject to urgent time constraints or require full coverage of an area under specific
conditions. This option guarantees high-priority image acquisition after analysis of
available satellite capacity and previous commitments. Priority programming fees
are charged on top of the price of programmed products acquired with less than
10% cloud.
High-Resolution Instruments onboard SPOT 1 to 5
SPOT1, 2 and 3
Spectral bands and
2 panchromatic (5m),
combined to generate a
2.5m product
3 multispectral (10m)
1 short-wave infrared
1 panchromatic (10m)
3 multispectral (20m)
1 short-wave infrared
1 panchromatic (10m)
3 multispectral (20m)
Spectral Range
P: 0.48 - 0.71µm
B1: 0.50 - 0.59µm
B2: 0.61 - 0.68µm
B3: 0.78- 0.89µm
B4: 1.58 - 1.75µm
M: 0.61 - 0.68µm
B1: 0.50 - 0.59µm
B2: 0.61 - 0.68µm
B3: 0.78- 0.89µm
B4: 1.58 - 1.75µm
P: 0.50 - 0.73µm
B1: 0.50 - 0.59µm
B2: 0.61 - 0.68µm
B3: 0.78- 0.89µm
Imaging swath
60km x 60km to 80km
60km x 60km to 80km
60km x 60km to 80km
Image dynamics
Angle of incidence
Revisit interval
(dependent on lat.)
1 to 4 days
1 to 4 days
1 to 4 days
SPOT 5 imagery over the Sydney Cricket Ground. 2.5m Pan, 10m
Multispectral and 2.5m Pan-sharpened natural colour. © CNES 2005
5 Pan-sharpened
at 1:25k
at ©
Port Hedland Mangroves 06 May 2002. © DigitalGlobe 2002
On January 29, 2001, an agreement (ORFEO - Optical and Radar Federated Earth
Observation) was signed in Turin between the French and Italian governments
to implement co-operation for setting up an Earth observation capability using
optical and radar sensors. This memorandum of understanding defines a
global co-operation scenario in the civilian and defence fields and included the
development of Pleiades, a multi-missions project for the procurement of high
resolution optical imagery.
The Pleiades project is made up of two “small satellites” (mass of one tonne)
offering a spatial resolution at nadir of 0.7m and a field of view of 20km. Their
great agility will enable a daily access all over the world, which is a critical need
for defence and civil security applications, and a coverage capacity necessary for
the cartographic applications at scales better than those possible from the SPOT
satellite family. Moreover, PLEIADES will have stereoscopic acquisition capacity
to meet the fine cartography needs, notably in urban regions, and to complement
information from aerial photography.
With respect to the constraints of the Franco-Italian agreement, cooperative
agreements have been set up for the PLEIADES optical component with Sweden,
Belgium, Spain and Austria.
The first satellite in the constellation is expected to be launched in 2009 and
the second in 2011.
Korea’s KOMPSAT-2 Earth observation satellite, launched 28 July 2007, has
returned its first images from an orbit 685 kilometres above the planet. The Korean
Aerospace Research Institute (KARI) developed the KOMPSAT-2 programme in
close collaboration with EADS Astrium, in particular to acquire VHR imagery for
South Korea’s needs in mapping, urban planning and hazard management.
In-orbit commissioning and operational qualification are proceeding according to
plan, paving the way for commercial sales of very-high-resolution (VHR) imagery
in a few months’ time.
SPOT IMAGE was appointed the exclusive distributor of KOMPSAT-2 data outside
Korea, the United States and the Middle East on 1st June 2007.
KOMPSAT-2 is a VHR optical imaging satellite (1m resolution in black and white,
4m in colour) capable of acquiring up to 7,500 images with a ground footprint
of 15 by 15km every day equivalent to 1.7 million sq km a day. These features
and imaging capacity, ideal for detecting and identifying ground features, make
KOMPSAT-2 a key asset for mapping at scales of 1:5k to 1:2k.
As prime contractor in charge of Korea’s satellite programmes, KARI is pursuing
an ambitious space plan: following the KITSAT series of microsatellites (1 to 4) and
the KOMPSAT-1 and KOMPSAT-2 Earth observation satellites, it is now pursuing
its programme with the development of KOMPSAT-3 and KOMPSAT-5, as well as
COMS-1, a communications, meteorology and oceanography satellite. KARI has
plans for 10 more satellites in the pipeline for the coming decade.
Mass: 1000 kg
Panchromatic resolution: 0.7m at nadir
Multispectral resolution: 2.8m at nadir
Swath: 20km at nadir
Sun-synchronous, orbit at 694km altitude
Acquisition capability: up to 450 images/day
Solar generator power: 1500W
Instrument TM link rate: 450Mbits/s
On board mass memory: 600Gbits
Lifetime: 5 years
The first remote sensing satellite developed by National Space Organization (NSPO) of
Taiwan, FORMOSAT-2 was successfully launched on May 21, 2004 . The main mission
of FORMOSAT-2 is to conduct remote sensing imaging over Taiwan and on terrestrial
and oceanic regions of other areas of the earth. Sales of image data to clients in Taiwan,
China, Hong Kong and Macao is through a domestic sales division with SPOT IMAGE
distributing imagery to all other clients.
FORMOSAT-2 is unique amongst the remote sensing satellites in that it is in a fixed
daily orbit, i.e. it repeats, the same ground track every day. Therefore its coverage of the
earths surface is a factor of its height and off-nadir viewing angles of ±45deg resulting
in approximately 60% coverage at the equator and successive higher coverage towards
the poles.
The unique features of FORMOSAT-2 including the daily revisits with the same viewing
parameters, the relatively early Equatorial crossing and the natural colour and Near
Infrared imagery makes the data ideally suited for:
> strategic and operational intelligence missions with the ability to identify and characterise
military sites, naval bases, air bases and refugee camps; perform reconnaissance of ships,
aircraft and other assets; and surveillance of strategic or industrial facilities,
> mapping shallow waters,
distinguishing between bare
earth and vegetation, mapping
forests and identifying crops,
> checking the health of
crops on a regular basis using
NDVIs and
> use in the Equatorial
regions where the multiple
revisit capability plus the early
morning overpass before
the convective clouds form,
FORMOSAT-2 composite, 8-12 June 2005
increases the chances of
© NSPO - Distribution SPOT IMAGE
acquiring useful imagery.
Features of FORMOSAT-2
> Earth Remote Sensing Satellite
> Weight: Approximately 760kg including payload and fuel
> Size and Shape: Hexagonal, height 2.4m, outer radius approximately 1.6m
(with solar panels folded)
> Orbit: Sun-synchronous orbit at altitude of 891km, passes over Taiwan twice daily
> Orbital time: Approximately 103 minutes
> Local Equatorial crossing: 9.30am
> Remote sensing resolution: 2m for black and white images and 8m for colour
images - Standard blue, green, red and NIR bands.
> Image Width at Nadir 24km; In track and cross-track viewing ±45deg
> Mission Life: 5 Years
> Launch date: May 21, 2004
Daily FORMOSAT-2 image coverage areas at 30deg and 45deg off-nadir viewing.
The satellite completes 14 orbital revolutions every day.
SPOT 5 Pan-sharpened false colour infrared Sydney at 1:25k scale. © CNES 2005
In previous GEOIMAGE brochures, we have highlighted the important contribution
that ASTER imagery has provided in the area of DSM generation, spectral processing
especially for mineral exploration and in the production of colour imagery at scales down
to 1:50,000. Because of the large worldwide archive of data, the relative cheap data cost
and its lack of competition in the spectral area, ASTER remains an important source of
imagery for a variety of applications. Perhaps the only area where ASTER’s importance
has waned is the production of DSMs and this is discussed below.
In this section, we have provided a brief summary of the ASTER sensors and have provided
an update of the major applications of the ASTER data. More details are available in the
GEOIMAGE 2005 Brochure and by accessing
It has been reported that “The Japanese instrument team is assessing the cost and
schedule to modify existing algorithms to adjust for the degradation and eventual
loss of the SWIR band (predicted Dec. 2007).”
ASTER ( Advanced Spaceborne Thermal Emission and Reflectance Radiometer) is an
imaging instrument onboard Terra - the first Earth Observing System (EOS) satellite.
Terra was launched on December 18, 1999 from the Vandenberg Air Force Base
in California and flies in a sun-synchronous polar orbit, crossing the equator in the
morning at 10:30. ASTER is one of the five state-of-the-art instrument sensor systems
onboard Terra with a unique combination of wide spectral coverage and medium
spatial resolution in the visible, near-infrared through shortwave infrared to the thermal
infrared regions. It was built by a consortium of Japanese government, industry, and
research groups. ASTER data contributes to a wide array of global change-related
applications including vegetation and ecosystem dynamics, hazard monitoring,
geology and soils, land surface climatology, hydrology and land cover change.
ASTER consists of three different subsystems; the Visible and Near Infrared (VNIR) ,
the Shortwave Infrared (SWIR) and the Thermal Infrared (TIR). The VNIR subsystem
consists of two independent telescope assemblies. One is vertical looking and
has three detector arrays collecting data in the visible green, visible red and near
infrared wavelengths, while a backward looking telescope has one detector array in
the same spectral band as the near infrared of the vertical array. These the infrared
(3N and 3B) generating an along-track stereo image pair with a base-to-height
ratio of 0.6 and an intersection angle of 27.7degs which can be used to generate
digital elevation models.
ASTER Sensor Systems
Band No.
Spectral Range
0.52 – 0.60
0.63 – 0.69
0.78 – 0.86
0.78 – 0.86
1.600 – 1.700
2.145 – 2.185
2.185 – 2.225
2.235 – 2.285
2.295 – 2.365
2.360 – 2.430
8.125 – 8.475
8.475 – 8.825
8.925 – 9.275
10.25 – 10.95
10.95 – 11.65
archive searches for data covering your area of interest
acquire the data and orthorectify to your datum/projection
mosaicing of multiple images
spectral processing of the VNIR, SWIR and TIR and supply in ER Mapper and/or
ECW compressed format with headers for display in MapInfo
> supply of digital data on DVD or for FTP download
> supply of normal or pseudo-stereo hardcopy prints
this level and the data has to be processed through a 1A to 1B conversion to be
used spectrally.
Level1B data is radiometrically calibrated and geometrically co-registered for
all bands. This level is used by those who want to process the data into spectral
reflectance values using their own software.
Level2 products are produced from the Level1B data and the main ones which
GEOIMAGE recommends to clients are.
The ASTER Level 2 Surface Emissivity product (AST_05) is an on-demand product and
provides an estimate of the thermal emissivity for each of the five ASTER TIR bands.
The ASTER Level 2 Surface Reflectance Product (AST_07xt) is a new product that
has only been available since 2006 and contains separate files of atmospherically
corrected Visible and Near Infrared data (VNIR) and Short Wave Infrared data (SWIR).
The SWIR data has also been corrected for cross-talk which is a signal scattering
problem that only became apparent after launch. The problem is caused by light
incident on the B4 detector being reflected by the detector’s aluminum-coated parts
and then projected on to the other detectors. Bands 9 and 5 are most affected
because of their closeness to the Band 4 detectors. Further details are available at
1:50k Scale Prints
The 15m resolution of the visible and near infrared bands of the ASTER data lends
itself to hard copy output at 1:50 000 scale. The most common band combinations
used by GEOIMAGE are the B321 and B731 in RGB. The B321 image uses all the
15m resolution bands in a false colour infrared combination with vegetation in red.
The disadvantage of this colour scheme is that there is very high correlation between
B1 and B2 resulting in a very cyan image. The B731 image is usually prepared with
a linear addition of the short wave infrared (B5+B6+B7+B8) in the red channel,
the vegetation band (B3) in the green channel and one of the visible bands (e.g.
B1) in the blue channel. This band combination uses the least correlated bands
and therefore has the highest colour information and is similar to a Landsat B742
colour composite.
Many of our clients also get us to prepare pseudo-stereo pairs of image and interpret
the imagery under a mirror stereoscope. This style of presentation was described
on Pages 50-51 of the 2005 Brochure.
DSM Generation
When ASTER was launched in 1999, the best worldwide DEM product that was
available was the GTOPO30 dataset from the USGS which had a resolution of 30arc
sec or approximately one kilometre. The ASTER along-track stereo pairs (3N-3B)
allowed the production of DSMs with an accuracy of about 10-15m at a grid spacing
of about 30m. The SRTM DSM (see Page 33) commenced to be released in 2004
and covered the land areas of the globe between 60deg north and 55deg south at
a 3arc sec spacing (approximately 90m at the equator). The accuracy of this DEM
has been shown to be generally less than 10m and certainly much better than that
possible from the ASTER imagery, albeit at a coarser grid cell. GEOIMAGE now
considers that the generation of DSMs from ASTER stereo pairs is only warranted
in cases where there is some doubt as to the accuracy of the SRTM data.
ASTER Processing Levels
On Page 35 of the 2005 Brochure, we described in some detail the various levels of
data available from the ERSDAC and EOS sites and this information is also available
from the following sites
The main processing levels are:
Level1A is the least processed data and usually only used for the stereo pair of
bands 3N and 3B to produce DSMs. The ASTER bands are not co-registered at
ASTER B765 Decorrelation Stretched ASTER SWIR at 1:100 000 scale Sar Cheshmeh Deposit, Iran.
ASTER B731 in RGB Stereo-SATMAP over the Sar Cheshmeh Deposit in
IRAN. The image covers an area of 30km by 30km at 1:50 000 scale.
ASTER Geometrical and Spectral Processing
GEOIMAGE has developed a standard methodology to process ASTER imagery for
mineral exploration and this procedure is explained here using an ASTER scene
over the Sar Cheshmeh Deposit in Iran as an example. The procedure involves
1. the selection of the best cloud free data over the area of interest,
2. the purchase of the AST-05 (Emissivity) and AST_07xt (cross-talk corrected
surface reflectance VNIR and SWIR) imagery,
3. the orthorectification of the ASTER imagery,
4. preparation of Lithological and spectral images in ERMapper format,
5. ECW compression of the imagery for display in MapInfo or ArcView,
6. vectorisation of the highest 1% predicted clay mineral groups into TAB and/or
Shape files,
7. preparation of a report describing the processing, and
8. supply of all data on DVD.
After downloading the data from the EOS FTP site, the VNIR and SWIR files are
orthorectified using the Geocover 2000 Landsat panchromatic image for X and Y
control and the SRTM DEM for height control. The final file is a 9 band coregistered
VNIR+SWIR integer ER Mapper file at 15m resolution. The file is usually in WGS84
and the relevant UTM zone (unless the client requires the data in a different
projection) and the locational accuracy of the file is probably of the order of 25m.
The TIR emissivity file is rectified using system parameters to 45m cell size and
its location is fine tuned to the VNIR+SWIR file. The final file is a 5 band integer
ER Mapper file at 45m resolution. The SRTM data used in the orthorectification
process is supplied as well as a coloured ECW file.
Silica Index
Carbonate Index
Mafic index
Quartz Index
Kaolinite/alunite group (B4 + B7)/(B5+B6)
Illite Group
(B5 + B7)/B6
Chlorite Group
(B7 + B9)/B8
Several TIR emissivity ratios are used separately or as a
colour composite.
Spectral Images And Vectors
Minerals of exploration interest that produce recognizable spectral patterns in
ASTER SWIR imagery are the epithermal clay mineral groups - kaolinite, alunitepyrophyllite and illite, and the propylitic minerals. These mineral groups can be
predicted using the Relative Band Depth estimator and the output of the estimator
is presented as both imagery and as vectors. The estimator is first presented as a
greyscale image where the brighter the pixel the more likely the probability that the
pixel is the relevant mineral group and then the highest 1% predicted distribution
is output as a MapInfo TAB and/or ArcView Shape file.
B731 composite image
stretch of Bands 765 in RGB
RBD image
Illite probability image
Alteration probability image on
albedo. Red is highest 1% alunite,
Green is highest 1% kaolinite and
blue is highest 1% illite
Highest 1% probability illite
TIR Indices with silica
index in red, carbonate index in
green and mafic index in blue
Quartz index
Lithological Information
Colour composites using bands from the ASTER VNIR, SWIR and TIR subsystems
can be used to give complementary information on ground cover and lithological
information. The images commonly prepared (bands in RGB order) areB321
Near infrared colour scheme with vegetation in red. Uses the bands
with the best spatial resolution (15m)
This is the band equivalent of a Landsat 742 colour composite and
gives more spectral information than the B321 although at the loss
of spatial resolution in the Red channel.
Iron ratio image. Ferric iron in red, band 3 (vegetation) in green and ferrous
iron in blue.
Decorrelation stretched image of SWIR bands 765.
Decorrelation stretched image of SWIR bands 876.
Decorrelation stretched image of SWIR bands 987.
Larry Rowan’s Relative Band Depth image highlights spectral details
in the SWIR subsystem.
ASTER B321 VNIR at 1:50 000 scale Mongolia.
The Landsat Project is the longest-running enterprise for acquisition of moderate
resolution imagery of the Earth from space. The Landsat 1 satellite was launched
in 1972 and the most recent, Landsat 7, was launched in 1999. The instruments
on the Landsat satellites have acquired millions of images. These images form a
unique resource for applications in agriculture, geology, forestry, regional planning,
education, mapping, and global change research.
The data presented here on the historical aspects of the Landsat Program is a
summary only and more details can be found on the web sites
Six Landsat satellites have now been successfully launched commencing with
Landsat 1 in July 1972. All platforms have operated from a repetitive, circular,
sun-synchronous, near-polar orbit and on each day-side pass, scan a ground swath
185km wide beneath the satellite. The first three satellites carried the Multispectral
Scanner (MSS) as the main imaging instrument with a Return Beam Vidicom (RBV)
as a subsidiary. The paths of these satellites were inclined 99 degrees with an
18 day repeat cycle and an Equatorial crossing of between 8:30 and 9:30am
local time. Landsats 4 and 5 have the Thematic Mapper (TM) as the main sensor
together with an MSS. They were inclined 98 degrees, have a repeat cycle of 16
days and an Equatorial crossing between 9:30 and 9:45am local time. Landsat 6
was unfortunately lost on launch in 1993, however Landsat 5 continued to provide
good data even after the launch of Landsat 7 in April 1999.
Landsat Program Summary
Landsat 1
Landsat 2
Landsat 3
Landsat 4
Landsat 5
Landsat 6
Landsat 7
Launch (End) Sensors Resolution
23 Jul 72
(6 Jan 78)
22 Jan 75
(25 Feb 82)
5 Mar 78
(31 Mar 83)
16 Jul 82
(TM - Aug 93)
1 Mar 84
5 Oct 93
(5 Oct 93)
15 Apr 99
15 PAN
30 MS
15 PAN
30 MS
Direct Downlink
with Recorders
Direct Downlink
with Recorders
Direct Downlink
with Recorders
Direct Downlink
Direct Downlink
TDRSS (Failed)
Direct Downlink
with Recorders
Direct Downlink
Solid State Recorders
Altitude Revisit Equatorial
8.30 am
9.00 am
9.30 am
9.45 am
9.45 am
10.00 am
10.00 am
Current Landsat Sensors
Landsat 5 is 22 years old and no redundancy remains for most of its mission critical
subsystems however it is still producing useable Thematic Mapper data. There have
been some scares in the last few years, notably with the backup Solar Array Drive which
malfunctioned in November 2005, however the problem was fixed by January 2006.
There has also been some deterioration in the bumper mode operation which affects the
locational accuracy of Landsat 5 systematic corrected imagery (ie Path Image and Map
Image Products). This deterioration appears to have become worse for acquisitions after
August/September 2006.
The Enhanced Thematic Mapper (ETM+) sensor on Landsat 7 developed an instrument
malfunction on May 31, 2003 and this has severely limited the usefulness of the current
data. The problem was caused by the failure of the Scan Line Corrector (SLC), which
compensates for the forward motion of the satellite. Subsequent efforts to recover the
SLC were not successful, and the problem appears to be permanent. Without an operating
SLC, the ETM+ line of sight now traces a zig-zag pattern along the satellite ground track.
Landsat 7 ETM+ is still capable of acquiring useful image data with the SLC turned off,
particularly within the central portion of any given scene. Landsat 7 ETM+ therefore
continues to acquire image data in the “SLC-off” mode. The SLC-off impacts are most
pronounced along the edge of the scene and gradually diminish toward the center of
the scene. The middle of the scene (approximately 22 kilometres with a L1G product)
contains very little duplication or data loss, and this region should be very similar in quality
to previous (“SLC-on”) Landsat 7 image data. Several different Level 1G (L1G) product
options have been developed by EROS in order to increase the utility of the L1G SLC-off
product. These options reflect the three different methods by which the original duplicated
data may be replaced during processing.
Level 1G Standard - This product
includes the original data gaps. Duplicated
pixels have been replaced with null values
(zero-fill) and the image will therefore
contain alternating “stripes” of missing
Landsat7 ETM+ B741 Colour Composite 139/29 05 September 2000 at 1:200k scale.
Level 1G Interpolated - Maximum (15-pixel)
interpolation provides a fully populated
image, in which all missing pixels have
been filled with DN values interpolated from
neighbouring scan lines.
Level 1G Gap-filled - This product provides
a fully populated image, in which all of the
missing image pixels in the original SLC-off
image have been replaced with histogrammatched data values derived from one or
more alternative acquisition dates.
ACRES refers to the Gap-filled products as Composite images and now offers two types
of products.
Landsat 7 SLC-Off ACRES Composite Product - one product made from the
combination of several separate scenes (in an attempt to fill the gaps); Orthocorrected
processing only; and
Landsat 7 SLC-Off Customer Composite Package - several separate products over
the one area for customers to composite in their own way; Path Image processing only.
(Recommended for experienced users only.)
To encourage use of Landsat 7 SLC-Off data, both of these options are provided at the
same low price.
These composite products and packages provide more useful Landsat 7 data because:
> Customers can choose SLC-Off scenes acquired in the same season (subsequent
passes) thereby reducing the ambiguity in interpretation due to temporal changes
> Customers can choose more than two SLC-Off scenes (maximum of five) to provide
a greater chance of filling the data gaps
> One additional SLC-On scene can also be included to ensure there are absolutely no
gaps in the final product. However this option is less useful for recent acquisitions
because an SLC-On scene would now be several years old.
Full details of the Australian SLC-off products offered by ACRES can be obtained by
contacting GEOIMAGE or from
Landsat Data Continuity Mission (LCDM)
Landsat 7 was launched in 1999 with a design life of five years. Early plans for
Landsat data continuity after Landsat 7 were based on NASA purchasing data from
a privately owned and commercially operated satellite system however these plans
were cancelled in Sept 2003. Subsequent attempts to place Landsat-type sensors
on National Polar-orbiting Operational Environmental Satellite System (NPOESS)
platforms were also cancelled and in Sept 2005, NASA was directed to acquire a
single Landsat data continuity mission in the form of a free-flyer spacecraft. The
instrument will collect land surface data similar to that of its Landsat predecessors.
The data will be delivered to the U.S. Geological Survey who will be responsible for
mission operations as well as data collection, archiving, processing and distribution.
Current planning is for the satellite to be launched in July 2011 and after a successful
launch it is anticipated that the sensor will be called Landsat 8.
Thematic Mapper Wavelengths
Number Wavelength (um) Applications
0.45 - 0.52
(visible blue)
coastal water mapping, differentiation of soil from vegetation,
has poor penetration through haze
0.52 - 0.60
(visible green)
vegetation vigour assessment
0.63 - 0.69
(visible red)
vegetation discrimination, also has high iron oxide reflectivity
0.76 - 0.90
(near infrared)
determining biomass content and delineation of water bodies
1.55 - 1.75
(middle infrared)
vegetation and soil moisture content, differentiation of cloud
from snow
10.40 - 12.50
(thermal infrared)
vegetation heat stress analysis, soil moisture discrimination,
thermal mapping, has limited use as a large percentage of
thermal radiation in daytime is reflected
2.08 - 2.35
(middle infrared)
discrimination of rock types and hydrothermal clay mapping
0.52 - 0.90
(visible green - near
textural detail
* Pan band only on Landsat 7
With six reflectance spectral bands, it is possible to make a total of 120 different
3 band colour composite images from a TM image. From a practical viewpoint
however, the first three visible bands are very highly correlated as are bands 5
and 7 so the number of significantly different 3 band combinations reduces to
only a few. The main ones being:Bands 147 or 247 in BGR simulated natural colour with the visible blue in blue,
vegetation in green and iron oxides in red. This band combination usually has the
least correlation between bands.
Bands 345 in BGR is usually preferred for vegetation studies.
Bands 123 in BGR is a natural colour image and although the bands are very
highly correlated often shows a remarkable amount of detail.
Bands 457 in BGR incorporates all the infrared bands and is useful in areas
affected by haze.
Of course, the selection of the bands is only the first step and the selection of the
type of histogram modification is just as important. This is especially so in the
case of a bimodal histogram where separate enhancements may be necessary
to give maximum information e.g. in a greenstone terrain with very bright granitic
and very dark greenstone areas. Region of interest contrast stretching may be
required where the image area used in the histogram generation is restricted to
the immediate area of interest e.g. over a particular rocktype, so that the resulting
contrast modification is maximised to show that particular rocktype.
In the case of Landsat 7 data, it is normal to pan-sharpen the 30m multispectral
imagery with the 15m panchromatic band and this results in much sharper imagery
and often highlights roads and tracks better.
Ratios of bands are commonly used because they highlight the spectral
differences between materials, at the same time decreasing the variations in
surface brightness due to topography. For example, the ratio of bands 3/1 is often
used to highlight iron oxide and 5/7 is used to highlight clays. Before ratios are
calculated it is important to remove the effects caused by scattering of light in the
atmosphere from each band. This effect which is highest in band 1 decreases to
almost negligible values in bands 5 and 7. Theoretically the atmospheric effect
would be the reflectance value in the image from a black body reflector on the
earth’s surface. This might be expected to be the minimum value in any band
however there may be noise in the image well below this value. The best method
of estimating the value is to examine the histogram and select the value where
the back slope of the band’s histogram intersects the zero reflectance. Use these
values in the ratio and examine the image produced. If there is still topographic
detail visible in the image slightly modify the subtraction values until it disappears.
When there is minimal slope information visible in the image, you will know that
the atmospheric corrections are correct. Ratios tend to highlight the noise in an
image and this can be minimised using a median filter.
Most clay minerals have an absorption feature in the area covered by Landsat
TM Band 7. Many techniques have been developed to highlight this effect and
� �
Spectral Reflectance Curves for common cover types with
wavelength bands of the main remote sensing satellites.
This diagram shows the spectral curves which cover the visible to mid-infrared part of
the spectrum. The spectral limits of the bands in the major satellite sensors are shown
diagrammatically at the top. The blue, green and red colours within the bandwidths denote
the bands in these sensors which are most commonly used to make a three colour composite
image. Note the absorption in the kaolinite spectra in the area corresponding to TM band 7 (A).
Most “clay” minerals have an absorption in this area although the shape of the spectra varies
and the feature can be used to predict the presence of clay minerals using a TM 5/7 band ratio
or the LSFIT technique. ASTER has five wavelength bands covering this clay absorption area
and allows some discrimination of clay groups.
Ferric iron is usually highlighted by TM band ratio 3/1 which highlights the visible red against
the visible blue. The ASTER sensor does not have a visible blue band and the ratio of ASTER 2/1
is not as effective in discriminating ferric iron. Ferrous iron has an absorption high in the area
corresponding to TM band 4 (B), and this characteristic is highlighted in a TM 7/4 or 5/4 ratio.
thus enable identification of areas of clay alteration indicative of mineralisation.
The LSFIT technique developed by the Australian CSIRO is a linear regression
technique that compares the predicted band 7 with the actual band 7 to identify
areas of anomalously high absorption and hence infer the presence of clays.
The above techniques are generally very important in providing information on
the rock types and possible alteration from TM imagery prior to field visits. It is
however important that the spectral techniques be applied properly and that
common sense is used in their interpretation. For example, techniques that try
to predict clays may also highlight water and snow as well as defining all playas
with thin coatings of clay. For this reason, it is important to interpret such images
in conjunction with a standard colour composite image of the same area.
Landsat 7 ETM+ images Koh-i-Sultan,
Pakistan. Path156 Row40. Date 10 May
2001. 15km by 15m area. Approximate
scale 1:200 000.
Image A. Bands 3 2 1 in RGB natural colour.
Image B. Band 3/1 ratio without atmospheric
correction. Note the image is like a greyscale
albedo representation of Image A.
Image C. Band 3/1 ratio with atmospheric
correction. Note the loss of topographic
Image D. Bands 741 in RGB. Vegetation is
now in green.
Image E. LSFIT of Band 7 with predicted clay
in white.
Image F. Abram’s Ratio. Band 5/7 ratio which
highlights clays and veg in red, 3/1 ratio in
green, and 4/3 ratio which also highlights
vegetation in blue.
Landsat7 ETM+ Alteration Image 139/29 05 September 2000 at 1:200k scale. High predicted clay in red and high ferric iron in green on a greyscale
GEOIMAGE is pleased to undertake searches of Landsat data from the various
worldwide archives and to provide price quotations for the purchase of the data
and processing to the clients specifications. This section is aimed at providing
brief details regarding the quality and the types of products available from the
various sources.
The Center for Earth Resources Observation and Science (EROS) is a data
management, systems development, and research field center for the U.S.
Geological Survey’s (USGS) Geography Discipline. EROS is located near Sioux
Falls, South Dakota and since it was established in the early 1970s has been the
largest repository of worldwide Landsat imagery. GEOIMAGE is affiliated with the
USGS through the Satellite Business Partner Program.
Data searching and ordering is performed through a variety of information
management systems and interfaces, including Glovis and Earth Explorer.
The USGS Global Visualization Viewer (Glovis) is a quick and easy
online search tool for selected satellite data. GloVis provides greater data availability,
allowing the user visual access to browse images from the Landsat 7 ETM+, Landsat
4/5 TM, Landsat 1-5 MSS, MRLC, and Landsat Orthorectified data sets, as well as
ASTER TIR and ASTER VNIR browse images from the DAAC inventory.
USGS Global Visualization Viewer (Glovis)
Earth Explorer ( ) is the USGS’s data
discovery and access tool for querying and ordering satellite images, aerial
photographs and cartographic products.
GEOIMAGE normally purchases the system corrected Level 1G data from EROS.
This data product provides systematic radiometric and geometric corrections using
data collected by the sensor and spacecraft. The scenes are rotated, aligned, and
georeferenced to a user-specified map projection. Geometric accuracy of the
systematically corrected product should be within 250m (1 sigma) for low-relief
areas at sea level. GEOIMAGE takes this product and orthorectifies it using ground
control points and the SRTM DEM and the result is a locational accuracy of less
than 50m at all elevation levels.
Thematic Mapper data from the Landsat 4, 5 and 7 sensors is available for a
minimum area of a standard full scene (180 by 180km) and non-standard framing
is available where the centre of the scene is moved in 10% increments between
adjacent centres. Multi-scene units defined as consecutive scenes (rows) within a
single path acquired and archived on the same date are available to a maximum
of 3 standard scenes.
The Australian Centre for Remote Sensing (ACRES) is Australia´s major public
satellite remote sensing organisation having begun operation as the Australian
Landsat Station in 1979. ACRES is within the Geospatial and Earth Monitoring
Division of Geoscience Australia (GA), which itself falls within the Australian
Government Department of Industry, Tourism and Resources. The Centre routinely
receives data from Landsats 5 and 7, RADARSAT 1, ERS 2, MODIS, NOAA 17
and 18 and ALOS. As well it has an extensive archive of Landsat MSS (extending
back to 1979), TM and ETM+.
The main Landsat processing levels provided by ACRES are 1.Path Image- Satellite path oriented data with systematic radiometric and
geometric corrections applied to the data. The data has been resampled in two
dimensions to fit a specific Earth datum and map grid. These products are only
available in full, double or triple scenes.
NASA GeoCover 2000 B742 Mosaic at 1:200k scale.
Landsat archive searches for data covering your AOI
acquire the data and orthorectify to your datum/projection
processing of ETM+ pan-sharpened bands
processing of iron ratios, Lsfit predictions and stretched band composites in ER
Mapper or ECW compressed format with headers for display in MapInfo
> supply of GeoCover and EarthSat NaturalVue Mosaics
> supply of digital data on DVD or for FTP download
> supply of normal or pseudo-stereo hardcopy prints.
2.Map Oriented Image - Map oriented data with systematic radiometric and
geometric corrections applied to the data. The data is rotated to fit a specific
Earth datum and map grid.
3.Orthocorrected Image - Map oriented data with systematic geometric
corrections refined with the use of GCPs and the GEODATA 9 second DEM. This
product is the most locationally accurate.
Map Oriented and Orthocorrected Image products can be framed anywhere along
the path of the satellite in the form of a variable sized rectangular “window”. The
sides of this rectangle are not parallel to the path of the satellite, but are aligned
in the north-south and east-west map-grid direction. The variable window product
is designed to remove the limitations of Path Image products by providing more
flexibility in choosing imagery to cover an area of interest, either across multiple
scenes or as smaller areas within existing scenes. The pixels are aligned to the
map grid rather than parallel to the path of the satellite (swath), allowing them to
be used with software and other products that require map grids.
To define a variable window product in the ACRES digital catalogue a particular path
and date of imagery is nominated and a rectangular window is specified by the
latitude and longitude of the window centre with the window extent in kilometres
east-west and north-south. All imagery from that path (image swath) falling inside
the specified window is then extracted to produce the Map orientated product.
Window Size
The product is extracted only from the path specified. If the nominated window
extends outside the image
swath, that area will be filled with
blank pixels. The maximum size
window that can be specified
is 310 by 530km. The E-W
extent is larger than the width
of the available imagery due
to the satellite path being tilted
N-E by 9 to 13deg, depending
on the latitude. Note that the
maximum N-S extent is larger
than the path oriented scene
size of 184.8km. This allows for
much more data to be ordered
in a single data set.
Pixel Sizes
The imagery is resampled to
square pixels which:
> are slightly smaller than
the original imagery to avoid
aliasing and loss of information
> are multiples of each other to
allow ready registration
> fit neatly into one hundred
metre grid intervals
Pixel sizes for the Landsat
data are
TM Multi
ETM+ Pan
ETM+ Multi
ETM+ Thermal
Some examples of variable windows are displayed
in the above image.
A - window wholly within path
B - window straddling the path
C - window cutting one side of path
Example window for larger ETM+ or TM Images
The maps below show the locations of ground stations operated by International
Cooperators (ICs) for the direct downlink and distribution of Landsat 7 (L7) and/or
Landsat 5 (L5) image data. The red and green circles show the approximate area over
which each station has the capability for direct reception of Landsat data. Landsat
5 initially had the capability of uplinking data to the TDRSS satellites however this
capability failed so the data is now only available through direct downlink.
> 50m RMS positional accuracy
> 28.5m pixel size for 7 spectral bands - Landsat 4/5
> 28.5m pixel size for 6 multi bands, 14.25m pixel for the pan band and 57m pixel
for the thermal bands - Landsat 7
> Interpolated with Cubic Convolution
> GeoTIFF file format - Load&Go GIS compatible
> ERMapper format (provided by GEOIMAGE)
The stock scenes can be reprojected into any customer selected projection and/or
datum. While the GeoCover scenes are a very cost effective dataset, it should be noted
that for certain applications they may not be ideal. For example, because the data were
preferentially selected when the vegetation growth was at a maximum, some scenes
may not be suitable for applications such as mineral exploration where it is important to
get maximum soil/rock exposure. In such cases, there may be better scenes available
from the EROS archive.
The Geocover scenes have high frequency noise on a grid pattern in both the pan and
multispectral data. This noise is presumably introduced during the resampling and is
most pronounced in the nearest neighbour resampled data which is why GEOIMAGE
always purchase the cubic resampled data. In the example above of the same
ETM+ pan image, A is EROS data, B is cubic resampled GeoCover and C is nearest
neighbour resampled GeoCover.
GEOIMAGE has established relationships with several of these stations to obtain
Landsat data that is not available from EROS, and this is especially important in areas
of high cloud cover such as the Amazonian Basin.
NASA sponsored the creation of an orthorectified and geodetically accurate global land
dataset of Landsat data from three time periods. These datasets incorporating Multispectral Scanner(MSS), Thematic Mapper (TM), and Enhanced Thematic Mapper (ETM+)
data, from the 1970s, circa 1990, and circa 2000, respectively, were produced to support
a variety of scientific studies and educational purposes. This is the first time a geodetically
accurate global compendium of orthorectified multi-epoch digital satellite data, at the 30to 80-m spatial scale spanning 30 years, has been produced for use by the international
scientific and educational communities. These sets of orthorectified Landsat images are
referred to as GeoCover and were compiled by EarthSat through NASA’s Commercial
Remote Sensing Program. Full details of the data selection, orthorectification, accuracy,
access, and other aspects are described in Photogrammetric Engineering & Remote
Sensing Vol. 70, No. 3, March 2004, pp. 313-322.
Approximately 8,500 Landsat TM images were selected from each of the 1990 and
2000 epochs. The acquisition dates of these images were relative to a 1990 and 2000
acquisition baseline, and the images were either cloud-free or contained minimal cloud
cover. In addition, only TM images with a high quality ranking in regards to the possible
presence of errors such as missing scans or saturated bands were selected. One of
the selection criteria used by NASA was that scenes be selected during periods of peak
greenness if possible.
The Landsat data were orthorectified, using geodetic and elevation control data, to correct
for positional accuracy and relief displacement. Large blocks of Landsat data were
adjusted through a patented procedure that uses pixel correlation to acquire tie-points
within the overlap area between adjacent Landsat images. Ground control points were
fixed, and images projected to the Universal Transverse Mercator map projection, using
the World Geodetic System 1984 datum (WGS84). All bands are individually resampled,
using a nearest neighbour or cubic convolution resampling. The result is a final product
with a Root Mean Square (RMS) Error of better than 50m in positional accuracy.
GEOIMAGE is a reseller of the GeoCover data and offers the following products. Please
check with your local office for pricing.
GeoCover-Ortho Single Scenes
The stock scenes are “off-the-shelf” orthorectified full scene TM images for the 1990
epoch and ETM+ images for the 2000 epoch with the following product specifications > Worldwide coverage of the Earth’s landmass
> Data set size based on Landsat WRS (Path/Row) Reference system
> UTM projection, WGS84 datum
GeoCover-Ortho Mosaics
GeoCover-Ortho Landsat mosaics are created by digitally suturing a group of juxtaposed
Landsat scenes into a single seamless digital image. Mosaics are colour (3 band)
products with Bands 2, 4 and 7 in blue,green,red (BGR). The stock mosaics have
been prepared from both the 1990 and 2000 epoch GeoCover stock scenes and are
“off-the-shelf” products with the following Specifications > Worldwide coverage of the Earth’s landmass
> Data set size based on 5 degree (N-S) segments of standard UTM
> UTM projection, WGS84 datum
> 50m RMS positional accuracy
> 28.5m pixel size - 1990 epoch
> 14.25m pixel size (Pan-sharpened) - 2000 epoch
> Contrast adjusted colour composite of TM bands 2,4,7 in BGR
> Cubic Convolution interpolation
> Compressed using MrSID or uncompressed GeoTIFF
> GEOIMAGE has compressed with ERMapper ECW compression.
EarthSat NaturalVue(TM)
This product is based on the 2000 epoch Landsat GeoCover imagery and is available for
the same 6 by 5deg tiles as the GeoCover mosaics. The data has been processed using
a proprietary process developed by MDA Federal to produce a Simulated Natural Colour
image which renders a pleasing green/brown/blue dominated colour pallet while at the
same time preserving the textural and spectral information of the 14.25m Landsat data.
This product therefore differs considerably from the NASA GeoCover B742 mosaic product
described above which has the undesirable side effect of turning cultural features purple
and vegetation a florescent green. The tiles are in Geographical projection WGS84 datum
and the 0.5 arc second pixel size approximates to15m. Each 6 by 5deg tile is delivered
in four 3 by 2.5deg subtiles which are about 1.1Gb each in GeoTiff format.
The NaturalVue product is a trademarked product developed by MDA Federal Inc. and
is available from GEOIMAGE as a licensed distributor.
Comparison of the EarthSat NaturalVue product (left) and the NASA sponsored
GeoCover B742 product over San Francisco Airport.
EarthSat NaturalVue(TM) Mosaic at 1:200k scale.
The Japanese Earth Resources
Satellite (JERS-1) was launched
in February 1992 by the National
Space Development Agency
of Japan (NASDA) It carried a
Synthetic Aperture Radar (SAR)
L-band instrument and an Optical
Sensor (OPS). The OPS had seven
spectral bands from the visible to
short-wave infrared. It was made up
of two sensors - the Visible and Near Infrared Radiometer (VNIR) and the ShortWave
Infrared Radiometer (SWIR). The satellite was in a circular sun-synchronous orbit at
an altitude of 568km and an inclination of 97.67deg. The revisit interval was 44 days
and the descending overpass on the equator was at 10:45am. The satellite was lost
in October 1998.
The L-band SAR on JERS-1 had a wavelength of 23.5cm and operated at HH
polarization. The swath width is approximately 75km and spatial resolution was
approximately 18m in both range and azimuth. The imaging geometry of JERS-1
was slightly shallower than either SEASAT or the ERS satellites, with the incidence
angle at the middle of the swath being 35deg. Thus, JERS-1 images are slightly less
susceptible to geometry and terrain effects. The longer L-band wavelength of JERS-1
allows some penetration of the radar energy through vegetation and other surface
types. JERS-1 Radar scenes are approximately 75km square and are supplied at
12.5m resolution.
Data Availability
The JERS-1 SAR could operate for 20 minutes per orbit and had an onboard tape
recorder which could record either SAR or optical data for 20 minutes. Thus worldwide
data from direct ground reception and from tape recorded data is available from
RESTEC and catalogue searches with quick looks can be carried out on their CROSS
site. Data is also available from individual international ground stations.The price of
JERS-1 SAR data has been reduced and the data provides a good source of historical
information in Equatorial areas where clouds and water vapour are problematic for
optical imagery.
The Global Forest Mapping (GRFM/GBFM) Project
The Global Rain Forest Mapping (GRFM) project was initiated in 1995 by the Japan
Aerospace Exploration Agency, JAXA (then known as NASDA). The project set out to
acquire spatially and temporally contiguous L-band Synthetic Aperture Radar (SAR)
data sets over the tropical belt of the Earth using the Japanese Earth Resources Satellite
(JERS-1), and to generate semi-continental scale, 100m resolution, image mosaics.
Building on the success of the GRFM project, the Global Boreal Forest Mapping
(GBFM) project was subsequently launched in 1997, with the aim to produce 100m
resolution JERS-1 mosaics also over the Earth’s boreal zones.
The GRFM/GBFM projects are led by JAXA’s Earth Observation Research and
applications Center (EORC), and undertaken in close collaboration with the NASA/Jet
Propulsion Laboratory (JPL), the Institute of Environment and Sustainability of the Joint
Research Centre of the European Commission (JRC/IES), the Alaska Satellite Facility
(ASF), Swedish National Space Board (SNSB), the Earth Remote Sensing Data Analysis
Center of Japan (ERSDAC) and the Remote Sensing Technology Center of Japan
(RESTEC) with scientific input from the University of California Santa Barbara (UCSB),
the Brazilian National Institute for Space Research (INPE), the National Institute for
Research of the Amazon (INPA), and the Swedish Land Survey office (LMV/METRIA).
All mosaic products generated within the GRFM and GBFM projects are available free
of charge for research and educational purposes worldwide.
archive searches for RADAR data covering your area of interest
arrange for new acquisitions where possible
acquire the data and orthorectify to your datum/projection
supply of digital data on DVD or for FTP download
supply of hardcopy prints
The Phased Array type L-band
Synthetic Aperture Radar (PALSAR)
on ALOS is an active microwave sensor
using L-band frequency for cloud-free
and day-and-night land observation.
The development of PALSAR was a
joint project between JAXA and the
Japan Resources Observation System
Organization (JAROS) and provides
higher performance than the JERS-1’s
SAR. Apart from the conventional fine resolution mode, PALSAR has the ScanSAR
mode, which will allow the acquisition of a 250 to 350km width of SAR data with
coarser resolution. This is a three to five times wider swath than conventional SAR
PALSAR Characteristics
Chip Bandwidth
Incident angle
Bit Length
Data rate
14MHz, 28MHz
HH or VV
HH or VV
8~60 deg.
8~60 deg.
18~43 deg.
8~30 deg.
100m (multi look)
40~70 km
40~70 km
250~360 km
20~65 km
5 bits
5 bits
5 bits
3 or 5 bits
120Mbps, 240Mbps
The PALSAR Observation Strategy
The PALSAR acquisition strategy features routine observations at four pre-selected
sensor modes. The mode selection represents a compromise solution where
scientific requirements, user requests, programmatic aspects and satellite operational
constraints have been taken into consideration. To assure spatially and temporally
homogeneous data collection over regional scales, acquisitions are planned in units of
whole (46-day) repeat cycles, during which only one of the available default modes is
selected. The PALSAR strategy is furthermore separated into one plan for ascending
(evening) passes, and one for descending (morning) passes.
PALSAR sensor default modes
Sensor Mode
Fine Beam Single pol.
Fine Beam Dual pol.
ScanSAR 5-beam
short burst
PNG JERS-1 SAR mosaic at 1:1 million scale. © JAXA/METI
(Experimental mode) *1
1270 MHz (L-band)
Fine Beam
GRFM JERS-1 SAR mosaic over Papua New Guinea.
1-2 obs/year
1-4 obs/year
2 obs/2 year
(a) Global
(b) Regional
(a) 1 obs/year
(b) 8 obs/year
Details of the PALSAR ascending mode plan (evening, ~22.30) and the
descending acquisition plan (morning, ~10.30) are described in some detail at
while graphics showing which regions are scheduled for ascending and descending
acquisitions during a particular cycle are available at
ERS-1 and ERS-2
The ERS satellites, operated by the European Space Agency (ESA), offered the first
commercially available microwave radar data with the launch of ERS-1 in July 1991
and ERS-2 in April 1995. Both satellites had basically the same instrumentation,
allowing applications to profit from the tandem operation of the satellites. The major
instrument on board these satellites is the Synthetic Aperture Radar (SAR) on the
Active Microwave Instrument (AMI), operating in Image Mode. This instrument can
also be operated in Wave Mode for measuring wave heights and frequency and there
is a separate radar on the AMI, called the Wind Scatterometer, for measuring wind
speed and direction.
Both ERS satellites operated in a sun-synchronous orbit at an inclination of 98° 52’
and an altitude of between 782 and 785km. ERS-1 was switched off in June 1996
and retired in March 2000. During the period August 1995 to June 1996, referred to
as the “tandem phase” ,ERS-2 was flown 1 day behind ERS-1, and in an orbit offset
a few hundred metres from that of ERS-1, to allow similar acquisition conditions for
stereo imaging and interferometry. ERS-2 remains fully operational.
In Image Mode, the SAR obtains strips of high-resolution (10 - 30m) imagery, 100km
in width, to the right of the satellite track.. The SAR is a C-Band radar (5.66cm, 5.3GHz)
with a bandwidth of 15.55MHz, and operates at an incidence angle of 23deg, utilising
VV polarisation. Due to power limitations, the SAR Image Mode can operate for a
maximum of 12 minutes per orbit. Acquisition in eclipse (i.e. during night-time passes)
is restricted to 4 minutes; a single image takes 15 seconds to be acquired
> 7 beam modes/35 beam positions for a wide range of imaging options
> Varying Resolutions (8 - 100m)
> Swath Widths of 50 - 500km
> Incidence Angles from 10 - 59deg
Within each RADARSAT beam mode, a number of incidence angle positions are
available. These are called beam positions. For example, Standard beam mode,
which covers a 100 x 100km area, has seven beam positions. By specifying a beam
position, one of seven 100 x 100 km images within a 500km accessible swath will
be collected.
and Resolutions
Beam Mode
Area Covered
500 x500
100 m
300 x 300
50 m
Extended Low
170 x 170
150 x 150
35 m
30 m
100 x 100
Extended High 75 x 75
50 x 50
25 m
25 m
The ENVISAT satellite was launched by ESA on March 1, 2002. Data from the ten
instruments onboard the satellite support Earth science research and allow monitoring
of the evolution of environmental and climatic changes as well as facilitating the
development of operational and commercial applications. The satellite has a 35 day
repeat cycle, 30 minutes ahead of ERS-2 and both satellites are controlled to overfly
the same ground track, within ±1Km.
The main instruments onboard the satellite are described at
The Advanced Synthetic Aperture Radar (ASAR) operates at C-band (5.331GHz) and
is an advanced version of the ERS-1,2 SAR in terms of coverage, range of incidence
angles, polarisation, and modes of operation. Its applications cover observations of
land and sea characteristics under all weather conditions and is ideal for studies of:
> Land: surface roughness, geomorphology, soil moisture, vegetation (moisture,
geometry, bio-mass)
> Ocean: surface topography (currents, fronts, eddies), wave characteristics
> Cryosphere: sea-ice dynamics and dissemination, ice sheets and shelf dynamics,
snowpack properties
In order to provide the ability to adapt to various observation requirements, ASAR
incorporates the capabilities to utilise different swath positions (incident angles) and
widths as per the following table.
wide swath
Dual Pol
Data rate (Mbps)
Max. operational
time per orbit(mins)
Max. data volume
(7 selectable)
(7 selectable)
Swath width(km)
HH or VV
HH or VV
HH and VV
HH or VV
HH or VV
The RADARSAT-1 satellite was launched on November 4, 1995 and has been providing
imagery for operational monitoring services on a global basis since that time. It is
equipped with a state-of-the-art Synthetic Aperture Radar (SAR) that can be steered
to collect data over a 1,175km wide area using 7 beam modes. This provides users
with superb flexibility in acquiring images with a range of resolutions, incidence angles,
and coverage areas and offers the following key benefits:
> C-band synthetic aperture radar (SAR)
> Cloud-free images of the Earth
> Frequent revisit for monitoring and emergency response
> Programming for emergencies and priorities
> Near-Real Time processing of data
> Direct downlink and onboard recorder storage capacity
> Data calibration for change detection studies
RADARSAT-2 is due for launch in late 2007. This satellite will have the same orbit,
repeat cycle and ground track as RADARSAT-1 and will provide continuity of the
RADARSAT-1 mission. Several improvements will be a 3m Ultra-Fine resolution
mode, more frequent revisits and faster response to user requests, fully-polarimetric
imaging modes and selective polarization (HH,HV,VH,VV), and GPS receivers onboard
for real-time position knowledge within 60m.
A new German radar satellite TerraSAR-X was successfully launched on June 15, 2007
and has provided preliminary data and should be fully operational within six months
of launch. The satellite is a public-private partnership between DLR and Astrium. The
scientific use of the TerraSAR-X data is the responsibility of DLR, as is the mission
planning and operation of the satellite, whilst Infoterra GmbH, a subsidiary of Astrium
specifically established for this purpose, will be responsible for the commercial
exploitation of the data.
The satellite has an active phased array X-band SAR, with single, dual and quad
polarisation, and is in a polar orbit at an altitude of 514km. Data can be acquired in
three different operational image modes:
> SpotLight (1m) the most sophisticated radar imagery available on the market:
Across flight direction, the bandwidth of 300MHz achieves a 1m resolution. In flight
direction, the radar beam can be steered like a spotlight, illuminating a particular
ground scene for the longest possible time period, thus achieving a 1m class
resolution as well.
> StripMap (3m): The ground swath is illuminated with a continuous sequence of pulses
while the antenna beam is fixed in elevation and azimuth. This results in an image
strip with a continuous image quality (in flight direction) at a resolution of 3m.
> ScanSAR (16m): Areas of up to 100,000sq km anywhere on the globe can be
covered in TerraSAR-X’s 16m resolution within only one week. In the ScanSAR
mode, a swath width of approximately 100km will be achieved by scanning four
adjacent ground sub-swaths with quasi-simultaneous beams, each with a different
incidence angle.
ALOS PALSAR Fine resolution imagery at 1:50k scale. Darwin area. © JAXA/METI 2007
GEOIMAGE specialises in processing airborne (and ground) geophysical surveys of all
vintages. We use quality software (including Intrepid, Geosoft, ERMapper, Geomatica,
ArcGIS, and MapInfo), and our staff have extensive experience in handling the good
and the not-so-good digital data. Examples of some of the surveys we have worked
on include re-processing and mosaicing the extensive publicly-released surveys from
Australia, Canada and west and southern Africa.
Geophysical Data Sources
GEOIMAGE can source geophysical data over your area of interest from a variety of
sources, the main ones beingThe Queensland OpenFile Survey Database
GEOIMAGE holds an extensive archive of openfile surveys flown in Queensland
and can supply as is or include in regional grid stitches over areas not covered by
the governments’ detailed surveys. Outlines of the openfile surveys processed are
available as a download from our website.
MAGNet Database
MAGNet, the Multiclient Airborne Geophysics Network, is a database of companyfunded
airborne geophysical surveys. It was jointly established by GEOIMAGE Pty Ltd and Pitt
Research Pty Ltd but is now controlled exclusively by GEOIMAGE.
MAGNet aims to provide:
> A centralised, current and consistent database of airborne geophysical, surveys
(complementing open-file surveys where appropriate) available for re-sale to the
exploration community,
> A mechanism which will relieve contributor companies of much of the day-to-day
distractions of managing, negotiating, and effecting geophysical data swaps and
sales and
> A repository that meets the requirements of the current guidelines for WA Dept
of Minerals and Energy for the lodgment of “multi-client” surveys.
The MAGNet database currently contains airborne geophysical surveys collected
by BHP Billiton, Rio Tinto, Newcrest Mining, De Beers Australia, and many more
companies within Western Australia and the Northern Territory.
For contributor companies, the benefits of belonging to MAGNet are:
> recoup part or all of the costs of their airborne surveying programmes,
> outsource the task of managing their archive of airborne geophysical data and
> leave non-core activities such as geophysical data sales to an agent.
A contract agreement covering conditions of data contribution, confidentiality, sales
and royalties is required to be signed by contributors.
The benefits to the purchasers are:
> improved access to existing data from a centralised, current and consistent
> data at a cost of around 1/5th the cost of new surveys and
> access to reprocessing that will offer substantial improvement to the utility of the
Please check with your nearest GEOIMAGE office for further details including survey
coverages, costs etc or if you would like to contribute data.
Geophysical Archive Data Delivery System (GADDS)
The Geoscience Australia website, is an initiative of the
Australian Chief Government Geologists Committee. It is aimed at providing a
portal to link the information of the following state, territory and federal government
geoscience agencies> Geoscience Australia
> New South Wales Department of Primary Industries
> Queensland Department of Mines and Energy
Three channel radiometrics K,Th,U in RGB 1:400k scale. Mount Oxide area.
> Western Australian Department of Industry and Resources
> Tasmanian Department of Infrastructure, Energy and Resources
> Northern Territory Department of Primary Industry, Fisheries and Mines
> Victorian Department of Primary Industries
> South Australian Department of Primary Industry and Resources
GADDS provides magnetic, radiometric, gravity and digital elevation data from
Australian National, State and Territory Government geophysical data archives as
either located (VECTOR) data in ASCII Columns or Intrepid Database format or gridded
(GRID) data in ER Mapper BIL format and in the Datum/Projection of your choice.
Staff at GEOIMAGE can download data over your area of interest from GADDS and
process to your requirements, whether that be provision of located or gridded data
in a suitable format, line or grid stitching, microlevelling or supply as GIS-ready
Datasets from four new Government geophysical surveys, three airborne magnetic
and radiometric surveys (the Ashburton region and Officer Basin in Western Australia
and the Mount Isa region in Queensland) and one gravity survey in the Mount Isa
region have been released since February 2007. The surveys were all conducted
in 2006 and were managed by Geoscience Australia on behalf of the Geological
Surveys of Queensland and Western Australia.
The GEOIMAGE website is updated regularly with advice on new government surveys
as they become available.
Geophysical Data
After carrying out all the preprocessing
of your geophysical data sets including
gridding, levelling, filtering, and
mosaicing, we can then provide you with
a complete set of integrated datasets.
We can take all your raster datasets,
which may include magnetic intensity,
RTP magnetics, vertical derivatives
of the magnetics, the radiometric
channels, gravity, satellite imagery, etc
and produce composite images that
show the spatial relationships between
the parameters.
These composite images can be output
as ECW compressed images that can
be viewed in MapInfo or ArcView where
they can be further integrated with your
geological and other vector infomation.
Alternatively we can make hardcopy
from the images.
Examples of integrated datasets in the
Georgetown region of Queensland.
A. High U on vertical derivative
B. High U on EarthSat NaturalVue image.
C. Composite radiometrics on 45az
shade on TMI.
Data Reprocessing
Data acquired by airborne geophysical surveys have been routinely stored on digital
media and processed digitally since the late 1970s. However, in comparison to
more recent surveys, these data are commonly seen to be inferior due to:
> Lower specification survey parameters like broader line spacing, increased
terrain clearance, and smaller volume crystals for spectrometer surveys,
> Inaccuracies in the survey location data due to poor navigation and recovery
> Difficulties in the digital processing due to lack of computing power and (by
today’s standards) unsophisticated software and computing techniques.
Gridded representations of these earlier surveys are commonly characterised by:
corrugations and “busts” due to poor tie-line levelling; crossing flight-lines and
terrain clearance variation along lines; spikes due to inadequate diurnal corrections
and poor processing; parallax and heading errors or introduced errors due to
excessive filtering and smoothing of flight-line data by contractors. The lack of
separation between potassium, thorium, and uranium channels in spectrometer
surveys is sometimes also a problem.
In reality though, this earlier data is an invaluable resource. Careful re-processing
can make any survey more interpretable, and allow easier extraction of relevant
information. A re-processed survey with lesser specifications (or better still a
careful mosaic of a number of these surveys) can provide a regional overview
that facilitates interpretation of a structural setting, as well as providing the basis
for better defining the parameters and location of “high-resolution” surveys over
potential target areas, possibly suggesting airborne EM or airborne gravity as
preferred or ancillary acquisitions. The re-processed survey may even serve
in its own right (depending upon survey specifications) as a basis for detailed
interpretation and information extraction.
This graphic depicts an example of geophysical processing and mosaicing of multiple
aeromagnetic surveys. Individual surveys were edited to correct obvious problems due
to elevation errors, reflights, and noise spikes in the original data. Surveys were then
microlevelled to a common base, and grid “stitched” at optimum cell size. This preserves
high frequency information from surveys flown at closer line spacing. Data shown here is
provided by the Northern Territory Geological Survey.
Total magnetic intensity (TMI) grid stitch of detailed state and federal government
airborne surveys and company openfile airborne surveys. Background TMI is the
Geoscience Australia terrestrial magnetics grid at 250-metre resolution. The state and
federal government grids were downloaded from GADDS and the company openfile
grids are held in GEOIMAGE’s archive.
Background image is as per the image on the left and showing the detailed state
government airborne surveys in blue, the detailed federal government airborne surveys
in red and the openfile surveys in yellow used in the TMI grid stitch.
Pseudocoloured shaded total magnetic intensity 1:400k scale. Mount Oxide area.
Digital elevation models (DEMs) are an integral part of any GIS. They are required both
for the description of the three dimensional surface but are also required to orthorectify
imagery which is used both as a backdrop and to provide information for the GIS. The
following is an update of pages 48-49 in the GEOIMAGE 2005 brochure.
High resolution DEMs with less than 1.5m accuracies are best obtained from high
resolution airphotos or from Airborne Laser Scanning (LIDAR) however for accuracies
in the 1m to 5m range DEMs produced from satellite imagery can be a cost effective
solution. Such DEMs are of course Digital Surface Models (DSMs) and include
buildings, vegetation, roads and natural terrain features. DTMs or Digital Terrain
Models are models of the bare earth that have had the buildings, vegetation and
natural terrain features removed.
Of the current satellite sensors capable of in track stereo, IKONOS is currently the
best as DigitalGlobe will no longer capture stereo with QuickBird. The successful
launch of WorldView-1 in September 2007 will however give DigitalGlobe the ability
to obtain 0.5m resolution stereo pairs.
The CARTOSAT Fore and Aft imagery should be capable of producing DEMs of similar
quality to the ALOS PRISM (See Page 17).
The SPOT5 payload includes the HRS imaging instrument developed by Astrium for
DEM generation. This instrument uses cameras looking 20deg fore and aft to image
stereo pairs over a surface area of 600km along track and 120km across track
centered on the satellite track. The spatial resolution of the instrument is 10m and
sampling cross-track is 10m and along track is 5m. SPOT IMAGE offers two DEM
products from the HRS data which are referred to as SPOT DEM and Reference3D.
Full descriptions of these products is available at
167_224_807_.php. The raw stereo imagery is not available for purchase.
The agile pointing capability of IKONOS with its ability to collect imagery up to
51deg off nadir in any direction, makes it ideal to collect same pass stereo pairs.
In fact, the height of the satellite and therefore the time it has over a target area
often allows it to capture adjacent stereo pairs on the same overpass. GeoEye will
capture stereo pairs between +/- 82deg latitude. In-track stereo pair acquisitions
are almost invariably new acquisitions and areas of interest must meet the normal
IKONOS ordering criteria i.e minimum area of 100 sq km and minimum 5km wide
in any dimension. GEOIMAGE usually orders the Reference Stereo products and our
normal experience has been that the imagery has been collected within 2 weeks of
the order except in one case where the order was cancelled because of continuous
cloud cover (northern Peru). GeoEye provides the stereo imagery pairs with a rational
polynomial coefficient (RPC) camera model file. The RPC file provides camera model
data to popular software packages for photogrammetric extraction of 3D feature
coordinates, DEMs and orthorectified imagery. Each stereo pair contains an image
collected at a low elevation angle (above 60 degrees) as well as an image collected
at a higher elevation angle (above 72deg) with 30 - 45deg convergence (0.54 to
0.83 base-to-height ratio). If the client also requires pan-sharpened imagery, we
order the most-vertical imagery as standard multispectral data.
GEOIMAGE usually gets the raw IKONOS imagery supplied at 0.8m resolution
and uses PCI OrthoEngine to create IKONOS DSMs at 1.6 or 3.2 m cell size and
with 8-10 accurate ground control points have been able to generate DEMs with
accuracies of 1-2m. NOTE: IKONOS stereo imagery is not capable of producing a
surface model where the requirement is for 1m contours to be used for pre-feasibility
volume calculations.
DEMs from ALOS PRISM triplets have been discussed on Page 16.
IKONOS image in India with 2m contours from an IKONOS stereo pair. Scale 1:10k.
Height range 178 to 255m. © GeoEye 2007.
Approximate DEM accuracies that might be expected from stereo image generation of the VHR to
medium resolution sensors compared to the documented accuracies of SPOT HRS and SRTM DEM.
Satellite height
Resolution (m)
DEM Accuracy
resolution (m) Absolute (m)
0.54 to 0.83
+24, 0, -24
+5, -26
(cross-track stereo)
20, -20
0, -27.7
SPOT HRG – SPOT IMAGE will not take orders for stereo image capture from SPOT 5.
DEM accuracies are dependent on many factors and the figures above are for high quality imagery in relatively low slope areas.
DEMs are usually produced at 2-4 times the image pixel size.
3m coloured DEM contours from ALOS PRISM on pan-sharpened IKONOS imagery 1:20k © GeoEye 2004
The actual accuracy of a DSM generated from VHR satellite imagery
will depend upon many factors. Chief among these are> The accuracy and spread of the client supplied control (in X, Y
and Z). Points should be spread evenly over the overlap region
and there should be a good spread of values in elevation with
points close to the minimum and maximum. The accuracy of
points should be of the order of 1m.
> The accuracy with which we can identify the locations of the
control points in the stereo imagery.
> Recognition of the fact that this is a surface model (hence
DSM) and will reflect all surface features, including trees,
buildings etc. These are not edited from the final product
> Where steep slopes are in shadow, or not visible to the satellite,
no surface model can be created.
> Environment factors affecting the imagery (cloud, shadow,
haze etc) will reduce the continuity of the surface
> Areas of the earth’s surface which are uniform, (e.g. water
bodies, salt lakes, beaches, some alluvial wash areas) reduce
the effectiveness of the auto-correlation process used to
define the surface. This may result in either holes in the DSM
or spikes which require editing out.
Primary producers have been quick to realise the benefits of using high resolution
satellite imagery in their farm management. They have recognised that to maximise
business profits, they need to monitor crop and pasture conditions to increase the land’s
production potential. In the GEOIMAGE 2005 Brochure we showed examples of the
use of satellite imagery for farm applications on pages 20, 21 and 29. Many of these
examples were from Controlled Traffic Farming Solutions (CTF Solutions) who have been
at the forefront of the use of imagery. See and especially their
section on the use of “High Resolution Imagery for Forensic Agronomy”.
From the high resolution satellite imagery, CTF Solutions have detected many issues
relating to crop production including:
> Nutritional disorders (particularly nitrogen)
> Paddock history effects (fence lines, headlands, etc)
> Grain harvester windrow effects
> Missed fertiliser striping
> Random wheel tracks and soil compaction from previous operations
> Water logging and poor drainage
> Sowing problems
> Varietal responses
> Pest damage
> Manure responses
> Missed fertiliser (DAP and urea)
> Planter blockages
> Responses to free range pigs
> Weed patches
> Crop diseases (Rhizoctonia, Nematodes), and also
> Consistent areas that are growing well
The early identification of problems affecting cropping is obviously important in their
management so that the long term effects can be minimised.
What type of Imagery is best for my application?
This brochure describes the common satellites that are used for farming applications
and the table below summarises their characteristics. Remember the finer the
resolution of the imagery the more expensive it usually is so you need to determine
the size of the features you want to be able to identify.
Pixel Size(m)
Data Type
Best Scale
SPOT 1, 2, 3 & 4
4. imagery that will input straight into your software with coordinate information built
into the image format,
5. files are only compressed slightly for ease of use, which doesn’t impact on
image quality.
ALOS imagery purchased for a group of ten farmers in a catchment group, Bombala,
New South Wales. Pan-sharpened natural colour split with pseudocoloured NDVI to
show the health and amount of vegetation. © JAXA 2007. Distributed by RESTEC.
Scale 1:50k.
NC represents a natural colour composite and NIR signifies the presence of the near
Infrared channel to produce an NDVI for measuring the health of the vegetation.
How do I get a Quote?
The price of an image depends on the size of the property and availability of archive
imagery, so in order to provide a quote, GEOIMAGE need some information about the
location of your property such as Lot and Plan number shown on your Rates Notice,
or GPS co-ordinates of the property. Once we have this information, we can provide a
list of options of imagery and dates available, and discuss with you the best imagery
for your requirements. See also Page 10 for details on ordering.
Can I use the imagery straight away?
GEOIMAGE will deliver the imagery to you on CD in a format suitable for your farm
mapping software. Various options are available such as ECW, JPEG, Geotiff and
bitmap, and we can provide advice on which formats you require. Imagery comes in
co-ordinated format which means you just input it straight into your software, and it
will know the location of your property, so you’re ready to measure, map and plan.
The locational accuracy of the imagery will vary but will generally be of the order of
about 20 to 30m in flat terrain with a higher error caused by height offset in rugged
terrain. If you require higher accuracy down to a few metres, this can be done by
orthorectification of the imagery using a digital terrain model and using differential
GPS points supplied by yourself.
Why not just use Google Earth?
Google Earth is great as a free viewing tool however it is not possible to use it in an
operational sense because the imagery does not go into your farm management
software. With GEOIMAGE supplied imagery, we can offer you
1. a more extensive archive of imagery from the various satellites,
2. we can manipulate colours specifically for your property to provide the best contrast
3. NDVI images to highlight the health and amount of vegetation,
Part of a 100sq km capture of a wheat farm in the Esperance Area of Western
Australia. Natural colour 0.6m resolution QuickBird. Imagery captured within 10days
of placing the order. © DigitalGlobe 2007. Scale 1:25k.
IKONOSfarm by GeoEye is a discounted package for 3 or more captures in one
year. Minimum area is reduced from 100sq km to 50sq km, effectively halving
the cost of new capture for properties under 5000ha. This is a great tool for
regular monitoring of your crops.
For further information on imagery for farms, contact Shona Chisholm in the
Brisbane office. ([email protected]) Shona has specialised in providing
satellite imagery for farm mapping for over 5 years, and can advise on all types
of satellite imagery for properties or farm applications, as well as different types
of software available. She has provided imagery to farmers all over Australia, as
well as overseas including countries such as Sudan, PNG and Vanuatu.
Natural colour 0.6m resolution QuickBird image Esperance Region. © DigitalGlobe 2007. Scale 1:10k.
We at GEOIMAGE are often asked by clients in the Mineral and Petroleum
Exploration industries - What Satellite image should I be using in this area? And
we always counter with a list of questions which might include> What scale do you want to use the imagery at?
> Do you want to process the data spatially (use it to define tracks and other
topographic detail) or spectrally (use it to define geology and alteration)?
> How large an area do you want to cover?
> What is the smallest feature that you want to be able to define?
> What other digital datasets do you have to integrate with the imagery?
In the next few pages, we hope to provide you with enough information for you to
decide what type of Satellite data you need for your application using examples
of the relevant imagery over the Kalgoorlie region. This area was chosen only
because GEOIMAGE either had or could obtain the imagery over the area from
our suppliers. Obviously similar data is available or can be programmed over your
tenement or area of interest.
Geoscience Australia © 1 million scale map showing the outlines of the satellite image
datasets referred to in the next few pages. Orange line is the northern limit of a Geocover 2000
sheet (S-51-25_G2000). Blue is Landsat scene 109/81, green is an ASTER scene, red is an
ALOS AVNIR-2 scene, purple is an ALOS PRISM scene and pink is a QuickBird scene.
APPLICATION: Broad scale structural studies and regional exploration
particularly in grass roots areas.
METHOD: Interpretation of digital data on screen or hard copy
prints both flat and pseudo-stereo. The pseudo-stereo is important
in structural studies to give a third dimension and also in areas
where it emphasises topographic relationships e.g.
in the search for laterites.
OTHER DATA: Other datasets which may be
incorporated include SRTM DEM, regional gravity
and magnetics, and geological vectors.
IMAGERY: Landsat Geocover Mosaics, Earthsat
NaturalVue or Landsat mosaics prepared by
Geocover 2000 sheet (S-51-30_G2000) covering
UTM zone 51 between 120 and 126deg East and 30
to 35deg South. The white graphic is the outline of the
Landsat scene 109/81.
A. SRTM image.
B. Gravity image.
C. Airborne Magnetics Image.
Images B and C ©
Commonwealth of Australia
(Geoscience Australia) 2004
The Shuttle Radar Topography Mission (SRTM) obtained elevation data on a near-global scale
to generate the most complete high-resolution digital topographic database of Earth. SRTM
consisted of a specially modified radar system that flew onboard the Space Shuttle Endeavour
during an 11-day mission in February of 2000. SRTM is an international project spearheaded
by the National Geospatial-Intelligence Agency (NGA) and the National Aeronautics and Space
Administration (NASA). See Page 55 of GEOIMAGE’s 2005 Brochure or http://www2.jpl.nasa.
gov/srtm/. The data is available at 30m cell size over the USA and 90m over the rest of the world
between the latitudes of 60deg North and 56deg South. GEOIMAGE has reprocessed the data
and produced the following products.
The data over Australia is available in either lat/long or UTM with the GDA94 datum and covering
the area bounded by 9deg South to 44deg South and 112deg East to 154deg East. The lat/long
data is provided subdivided into the different states. Files sizes vary from 1338Mb for Western
Australia to 466Mb for the Northern Territory and fit on one DVD. A coloured ECW file of the data
which also includes the islands to the north of Australia is also supplied.
For further information please visit our website
or contact one of the GEOIMAGE Offices.
Regional pseudocoloured magnetics Yilgarn area 1:2500K © Geoscience Australia.
Worldwide data is available in the same tiling system as the GeoCover mosaics i.e. 6deg by 5deg
UTM tiles in WGS84 or by special request in lat/long projection.
APPLICATION: Geological and structural interpretations, regolith studies and limited
spectral studies.
METHOD: Interpretation of digital data on screen or hard copy prints both flat and pseudostereo (especially for structural studies).
OTHER DATA: Other datasets which may be incorporated include SRTM DEM, airborne
magnetics and radiometrics, geological and tenement vectors, mineral deposit
IMAGERY: Landsat 5, Landsat 7 ETM+, Geocover Landsat Mosaics, and Earthsat
Landsat is the image of choice at these scales and there are numerous advantages
including a good selection of imagery for most areas in the world, relatively low cost and
ready availability. Some care needs to be exercised in carrying out searches and ordering
data mainly relating to the amount of vegetation cover and staying clear of fire scars and
GEOIMAGE is happy to look after these aspects for you.
Landsat spectral work should not be dismissed. The presence of the blue band allows the
calculation of a ferric iron ratio (b3/b1) which is much superior to the ASTER ferric iron ratio
of b2/b1. Remember however that the ferric iron ratio must be calculated on reflectance data
rather than on raw data as the lower bands are most affected by atmospheric absorption. The
absorption of light in b7 due to clay minerals is of course important in mineral exploration
and predicted clays can be estimated by either a b5/b7 ratio or the LSFIT technique. The
Abram’s and Crippen’s Ratios are ratio composites which combine three different ratios in
a colour image to achieve maximum discrimination.
Imagery at these scales only shows a minimum of infrastructure and AUSLIG vectors are
available if you wish to locate features.
Landsat 7 ETM+ imagery Bands 741 in RGB pan-sharpened at 1:100k scale with
geological boundaries from the Geoscience Australia 1:250k geological mapping
series in white.
APPLICATION: Tenement scale structural and geological interpretations and
spectral studies
METHOD: Interpretation of digital data on screen or hard copy prints both flat
and pseudo-stereo. The pseudo-stereo is important in structural studies to give a
third dimension and also in areas where it emphasises topographic relationships
especially in the search for laterites.
OTHER DATA: Other datasets which may be incorporated include SRTM DEM, regional
gravity, airborne magnetics and radiometrics, geochemistry and geological vectors.
Landsat data that has been pan-sharpened will just enlarge to 50k and the
commonest band combinations are B321 (natural colour) and B741.
The ASTER VNIR which has a pixel size of 15m is very sharp at 50k however it is
often best to incorporate a SWIR band (30m pixel) for better discrimination of surface
materials even at the expense of some sharpness (See example at right).
The main advantage of ASTER is for discrimination of clay-groups using the SWIR
bands for studies of alteration associated with mineralisation or regolith studies
to determine geochemical sampling locations (See Pages 20-21).
SPOT imagery including SPOT5 multispectral (10m) and SPOT4 pan-sharpened
multispectral (10m) enlarge well to 50k however they are limited by the absence
of a visible blue band (see Page 34).
ALOS AVNIR-2 at 10m resolution and covering a 70km by 70km footprint enlarges
well to 50k scale and the presence of a visible blue band allows the preparation
of natural colour composites makes it the ideal image at this scale.
There is a definite limitation as to how much infrastructure can be seen at 50k.
This is usually confined to major and minor tracks and some grid lines and areas
that have been cleared of vegetation for drilling etc.
ASTER imagery Bands 321 in RGB giving a false colour infrared composite to take
advantage of the VNIR 15m resolution. Scale 1:50k. The inset image (at 100k scale)
is a Band 731 composite (equivalent to a Landsat band 742) which has better
discrimination of surface materials but because it incorporates one of the 30m SWIR
bands is not as sharp at 50k scale.
APPLICATION: Structural studies and semi-detailed exploration commonly in brown fields
work in mineralised provinces and around existing mines.
METHOD: Interpretation of digital data on screen or hard copy prints both flat and pseudostereo. The pseudo-stereo is important in structural studies to give a third dimension and
also in areas where it emphasises topographic relationships especially in the search for
OTHER DATA: Other datasets which may be incorporated include DEMs (from ALOS,
IKONOS, airphotos, airborne geophysical surveys or SRTM), gravity and magnetics, and
geological vectors.
IMAGERY: The main satellites used at these scales are ALOS and SPOT5.
Both these satellites have a 2.5m panchromatic resolution however in the case of SPOT5, this
is a contrived resolution produced from two 5m offset sensors and in the case of ALOS PRISM,
the data has stripping and JPEG compression noise. In both cases, the 2.5m black/white
and pan-sharpened colour imagery is best output at a maximum scale of 1:15k.
The SPOT imagery does not have a visible blue band and is therefore less useful than the
ALOS AVNIR-2 for geological applications. The blue band is important in differentiating soil
types, clays and ferric oxides (vis red/vis blue ratio). SPOT can be programmed to obtain
data if there is no existing imagery over your area of interest or if you need imagery that
has been captured after an event of interest. In contrast, ALOS data is only available from
the existing archive although the programmed captures until 2009 are available on the
website or in the JAXA search engine.
One of the surprising features of SPOT 5 is the quantity of data available in archive. This
has the advantage that in areas of the world where you would least expect to find low cloud
imagery, such as the Philippines and Equatorial Africa, these is a surprising amount of useable
imagery. Because this imagery often has same date 2.5m pan and 10m multispectral data,
it is useable at all scales from 15k to 50k and you don’t have to worry about having to find
same date pan and multispectral as is the case with ALOS.
The locational accuracy of SPOT 5 data (using ephemeris data without ground control) at
better than 50 is superior to ALOS data however the worldwide availability of the GeoCover
Landsat imagery with a documented accuracy of better than 50m and the SRTM DEM allows
orthorectification of SPOT and ALOS to locational accuracies better than 50m. If higher
accuracies than this are required then client supplied ground control is necessary.
The ALOS PRISM DEM generation which costs about A$5000 for a 5-10m cell size DEM
with an accuracy of about 5m for a 35km by 35m area is a major improvement on the SRTM
DEM. The archive of available PRISM data suitable for DEM generation is increasing rapidly
and DEMs with very high relative accuracies can be produced anywhere in the world where
PRISM triplets are available and the SRTM DEM is available for the Z control.
Pan-sharpened natural colour ALOS
imagery at its best scale of 1:12.5k
showing lines of percussion drill holes.
Kalgoorlie area. © JAXA/Distributed
Pan-sharpened natural colour
QuickBird imagery at 1:12k for the
same area. Insert at 1:5k scale.
© DigitalGlobe 2005
Pan-sharpened natural colour AVNIR-2 ALOS imagery at 1:25k scale. © JAXA 2007
APPLICATION: The VHR satellite imagery can be used for most applications that
aerial photos were previously used for. This includes> detailed structural and geological interpretations,
> fracture mapping
> definition of previous exploration including gridding, drill locations, etc
> limited spectral work for ferric iron
> defining existing access or planning new access routes,
> outlining ownership boundaries
> environmental monitoring and change detection
> planning seismic lines
> visualisations for better understanding of topographic relationships or for
management presentations
> planning new infrastructure developments including pipeline planning
METHOD: Interpretation of digital data on screen or hard copy prints both flat and
pseudo-stereo. At these scales, a higher resolution DEM than the SRTM is required
for pseudo-stereo prints and this is usually derived from an IKONOS or ALOS DEM.
OTHER DATA: Other datasets which may be incorporated include SRTM DEM, ALOS or
IKONOS derived DEM, detailed airborne geophysics, ground geophysics, geochemistry
and geological vectors.
IMAGERY: Currently the options are QuickBird or IKONOS.
Both IKONOS and QuickBird imagery are purchased on a per sq km basis. Areas of
interest (AOIs) can be defined by a polygonal Shape file or a bounding coordinate
rectangle. The swath width of an IKONOS capture is ~11km and for a QuickBird
capture is ~16km. Hence more than one date swath may be required to cover the
AOI, however this does not affect the purchase price of the imagery, although it may
affect the processing costs. For IKONOS, minimum areas of purchase are 49sq km
for archived data and 100sq km for fresh capture. For QuickBird, minimum areas of
purchase are 25sq km for archived data and 64sq km for fresh capture. For either
imagery, the minimum width anywhere within the polygon or bounding rectangle
must be 5km.
Types of Imagery
All operators collect high resolution panchromatic and lower resolution 4 band multispectral imagery. These images can be ordered separately, as a bundled product or as
a pan-sharpened multi-spectral image at the higher resolution of the pan image. The
pan-sharpened colour product is generally sufficient for most applications. IKONOS
pan-sharpened imagery comes standard as 4 bands (vis blue, vis green, vis red and
near infrared) while QuickBird can be purchased as three or four bands. The 3 band
option lowers the cost and usually the only reason to purchase the near infrared band
is if you have an environmental application or you want to use it in conjunction with the
visible green to make a visually more pleasing enhanced natural colour image.
Location Accuracy
Both DigitalGlobe and GeoEye quote accuracies for the supplied data of about 20m
excluding topographic displacement and off-nadir viewing angle. Both satellites can
collect imagery at up to >40deg from the vertical and use this capability to increase the number of
possible captures of an area. When you consider that for imagery collected at 30deg from the vertical,
QuickBird imagery is offset by 50m for every 100m of elevation difference, orthorectification with
as accurate a DEM as possible is a minimum requirement. To get accuracies down to a few metres,
surveyed ground control points and an accurate DEM are required.
Elevation Models required for the orthorectification process can be derived from stereo imagery from
SPOT or from ALOS. If a better resolution DEM is not available, the SRTM data with its 90m resolution
and with height accuracies of about 5-8m is the best alternative.
Ground control points which are identifiable on the imagery are essential to the accurate location of
the imagery. If available, several points with accuracies approaching the resolution of the data (and
an accurate DEM) can result in a final accuracy of less than 2 pixels. Even with only one ground
control point and an accurate DEM, accuracies of less than 10m should be obtainable using either
the Rational Polynomial Coefficients (RPC) that are standard with QuickBird data or with the satellite
model in PCI OrthoEngine. The accuracy of the final product when using a less accurate DEM such
as the SRTM will depend on the angle of capture of the imagery and the slope of the terrain.
GEOIMAGE has an information document that offers advice to clients for collection of ground control
points or for the laying down of targets before the satellite collection. Based on the type of ground
control and DEM that the client can provide, GEOIMAGE can recommend the most suitable level of
imagery to purchase.
Cloud Cover
Cloud cover (and cloud shadow) will always be a problem especially in Equatorial areas. All the
operators have a similar stipulation for new collection with less than 20% cloud being seen as a
successful capture. GeoEye customers can designate a 2km square area that will be cloud free. In
the case of in-track stereo imagery for DEM production, cloud and shadow are a significant problem.
Because of the different look angles for each part of the stereo pair, the projection of the cloud and
shadow on the ground will be significantly different in each of the two images. Even with 20% overall
cloud, the resulting DEM will have significantly greater than 20% unresolved elevation model.
Archive Imagery vs New Capture
GeoEye offer significant cost savings for imagery that has been in archive greater than 1 month
(Australia and Oceania) or 6 months (rest of the world) and DigitalGlobe offers a price reduction on
all archival QuickBird imagery. However, the main advantage of purchasing imagery from archive is
that the extent and distribution of cloud is known prior to purchase of the imagery. Turnaround time
from order to delivery for archival imagery is also significantly quicker.
The satellite operators quote various time delays for the capture of new imagery and this will depend
on a number of factors:
> season of capture and therefore the cloud cover
> the size of the area - if the required area has to be captured in several strips, then IKONOS is more
likely to be able to capture adjacent strips in a single overpass
> the existing programming orders in the same area or on the same satellite trajectory.
DigitalGlobe allows priority ordering of new capture at a price premium.
Pan-sharpened natural colour QuickBird imagery at 1:15k scale with 2m coloured
DEM contours from an ALOS PRISM derived DSM. © DigitalGlobe 2005
Pan-sharpened natural colour QuickBird imagery at 1:5 000 scale Kalgoorlie area. © DigitalGlobe 2005
1:2 500 Scale
QuickBird Brisbane
Enhanced natural colour
© DigitalGlobe 2003
Pan-sharpened enhanced natural colour QuickBird imagery at 1:5k scale Port Hedland Mangroves 06 May 2002. C DigitalGlobe 2002
1:5 000 Scale
QuickBird Brisbane
Enhanced natural colour
© DigitalGlobe 2003
1:10 000 Scale
QuickBird Brisbane
Enhanced natural colour
© DigitalGlobe 2003
1:15 000 Scale
SPOT5 Pan-sharpened
ALOS Pan-sharpened
multispectral imagery
Enhanced natural colour
© CNES 2005
AVNIR-2 © JAXA 2006
1:25 000 Scale
SPOT5 Simulated
ALOS Pan-sharpened
Pan-sharpened natural colour
Enhanced natural colour
imagery © CNES 2005
AVNIR-2 © JAXA 2006
1:50 000 Scale
Landsat 7 ETM+ Bands
Landsat 742 band
741 in RGB
look alike
Pan-sharpened image
1:50 000 Scale
SPOT5 Pan-sharpened
Simulated natural colour
Enhanced natural colour
© CNES 2005
© JAXA 2006
PO Box 789, Indooroopilly QLD 4068
13/180 Moggill Road
Taringa QLD 4068 Australia
TEL: +61-7-3871 0088
FAX: +61-7-3871 0042
Email: [email protected]
PO Box 208, Crows Nest,
SydneyNSW 1585 Australia
TEL: +61-2-9460 0929
FAX: +61-2-9460 0929
Email: [email protected]
PO Box 8013, Subiaco East WA 6008
27A Townshend Road
Subiaco WA 6008 Australia
TEL: +61-8-9381 7099
FAX: +61-8-9381 7399
Email: [email protected]
There are many reasons to consider talking to GEOIMAGE when you are thinking about spatial
information. Besides our friendly and efficient service, below are some of the main reasons why our
loyal client base extends across Australia and throughout Asia, Africa, North and South America.
Independent – GEOIMAGE is independent of all satellite operators and therefore not restricted to
providing only one or two types of imagery. GEOIMAGE can source or acquire the most appropriate
imagery and supply an unbiased, cost effective solution to your needs.
Flexible – Whatever your application is, GEOIMAGE not only provides the best satellite imagery for
your needs, we can also process the imagery to make sure that it seamlessly integrates into your GIS
or other applications. So all you have to do is simply load onto your system and off you go!
Reliable – We are an Australian-owned company that has been providing and processing spatial
information for over 19 years. We have expanded over the years and now have three offices servicing
our clients from many differing backgrounds. GEOIMAGE keeps up to date with the latest that technology
has to offer, and will continue to evolve with the ever changing needs of the spatial industry.
Knowledgeable – Our staff have experience in many different applications from geology and
geophysics, natural resources, government applications, mapping, GIS & the environment, mathematics
and IT. By combining our technical skill and experience, we are able to deliver an innovative solution
which will fulfil your objectives.
Global – GEOIMAGE has access to imagery from around the globe, so if you have a mining application
in China, a utilities application in Zimbabwe or require imagery over Antarctica, we can provide the
spatial imagery that you need. Don’t let your location restrict you in your choice.
Home of Penfolds Grange, Vineyards of the Barossa Valley.
Enhanced Natural Colour QuickBird Image 1:5 000 scale.
28 February 2005. © DigitalGlobe 2005.