International Archives of Photogrammetry and Remote Sensing. Vol. XXXII Part 7C2, UNISPACE III, Vienna. 1999
24
I5PR5
UNISPACE m - ISPRS Workshop on
‘"Resource Mapping from Space"’
9:00 am -12:00 pm, 22 July 1999, VIC Room B
Vienna, Austria
ISPRS
7. Present Capabilities of Microwave Satellite Systems
Optical satellite systems have sofar shown difficulties with
marginal resolution and with cloud coverage. For this reason the
European Space Agency' ESA has launched the ERS radar
satellite series followed by Japan with JERS and Canada with
Radarsat.
Radar signals follow a very different geometry from optical
devices. The existing satellite systems have also not been
designed for land, but for ocean coverage, giving differences in
the amount of foreshortening or shadows to be encountered.
Moreover, radar backscattering does not behave like optical
reflection. Thus the criteria for planimetric accuracy and for
detectability of objects from radar images do not easily
correspond to pixel size resolution of optical systems.
Depending on the grey level differentiation desired radar
processing can be executed (e.g. for ERS-1) to 12.5 m or 25
pixels related to the ground.
Thus radar images are more suitable for image fusion with
optical images rather than for mapping alone.
Radar, when used in multitemporal mode or multipolarization
mode, on the other hand, compares to optical multispectral
differentiation capabilities, when classifying area objects. Yet
satellite radars have a very distinct advantage, they generate
coherent radiation. Coherent radar signals influence radar
geometry since the images are reconstructed using phase and
Doppler informations. They can only properly be reconstructed
if a D.E.M. is used in the reconstruction.
On the other hand phase information may be utilized for D.E.M.
generation using interferometry principles. While radar
interferometry has been carried out using repeated images from
quite different orbits (as for JERS-1 and Radarsat), as long as
there is a base between the two imaging stations, the conditions
for interferometry were greatly improved during the ERS-
l/ERS-2 tandem mission which directed the 2 ERS satellites so
that corresponding radar coverages over the same area were
obtained one day apart with a small base between orbits.
Radar interferometry requires not only that the base between
orbits is known, but that the sensor position and the sensor
attitude is accurately determined. In the present satellite systems
such possibilities have not been available with sufficient
accuracy.
In the Hannover area a digital elevation model generated from
radar interferometry has been fitted by a 7 parameter
transformation to the few identifiable common control in the
radar images. In most areas the comparison with an existing and
accurate conventional D.E.M. resulted in average discrepancies
of less than 10 m for 90 % of the points, but in hilly areas
discrepancies of over 100 m were encountered.
This confirms that radar interferometry may be useful in
differential changes of image portions, which are due to
dynamic changes of the terrain, but it sofar fails in reliable
elevation determinations.
An improvement of this situation has been suggested by the
U.S. German SIR-C Space Shuttle Mission in 1999, which not
only will use three radar frequencies (X-band. C-band, L-band)
to resolve atmospheric ambiguities, but also will have tw'o types
of receiving antennas mounted on a 30 m long beam, the
positions and attitudes of which are continuously monitored by
differential GPS. In this way it is hoped to obtain a system
capable of rapid D.E.M. mapping for large areas to ± 10 m.
8. Digital Elevation Models
There are currently at least six alternative methods considered
for future generation of digital elevation models, competing in
accuracy and price:
* the digitization of existing line maps at 5 to 20 m accuracy,
depending on the original map quality, at prices of 1 $/knr
■ the existing US Military global coverage at 20 m accuracy,
obtainable for some portions of the globe, at prices of 1
$/km 2
■ aerial photogrammetric mapping at 5 m accuracy' for 40
$/km 2
■ the future use of US commercial optical satellites,
following the Stereo-MOMS principles, but at higher
resolution with 5 m accuracy for 50 $/km 2 , with the
advantage to map areas, when aerial photography is not
available
■ optical stereo sensors following the Stereo-MOMS or the
SPOT/IRS-1C principle with 10 m accuracy at 5 $/knr
■ interferometric SAR of the SIR-C type with 10 m accuracy
at 5 $/knr.
9. Use of Satellite Remote Sensing Systems
The use of existing satellite images for monitoring the
environment in the largest sense is manifold. The choice of
imagery is always a compromise between availability and
spatial, spectral and temporal resolution considering
repeatability, swath, and pixel size.
All remote sensing conferences in Asia and in other parts of the
world show that weather and meteorological dangers (storms,
cyclones) can be monitored by global satellites such as Insat.
NOAA satellites permit to measure sea surface temperature,
pigment and chlorophyll concentration of ocean and coastal
areas. But they are also able to monitor the state of vegetation
on the basis of the normalized digital vegetation index to follow
patterns of drought or floods, the health of tropical forests or the
devastation by forest fires.
Resource satellites are useful in monitoring crop patterns in
detecting vegetation diseases, to determine erosion risks or to
follow' uncontrolled growth of industrial activity and urban
settlements and their pollution effects on the environment.
There would be no hope to gather the amount of this type of
information without satellite remote sensing systems.
These interpreted results together with socio-economic data
constitute the needed thematic information which needs
reference to the base mapping system in form of a geographic
information system.
Noteworthy are the recent activities of the Committee on Earth
Observations by Satellites CEOS, an organization of the space
agencies formed on the initiative of the 67 group. CEOS
propagates the creation of an information locator system which
helps the user to find pertinent information via the Internet. This
CEOS-ILS is to contain types, locations, and times of satellite
sensor images taken, if possible with reduced content
