|

|
ERS-2
and RADARSAT Browser implemented using ACT's ProVIEW
and ProVIEW WEB.
|
|
The
central objectives of the NATO's Rapid Response (RR) is the
dissemination of time-critical data and information to war-fighters
and war-fighter analysts. For RR96-98 ACT demonstrated an
interactive network-based image processing system that provides
high speed access and processing to large volumes of remotely
stored data.
In
support of RR-96, a total of 40 ERS-2 and one RADARSAT SAR
images were acquired during the period of August 12 through
5 October, 1996. The raw SAR data stream was first down-linked
in real-time to the receiving station in West Freugh in UK.
The data was then routed via high speed links to DRA Farnborough
remote sensing facilities for image formation. Then, high
resolution (150m) SAR images were sent via the Internet to
Naval Research Laboratory (NRL) in Washington, D.C. for post-processing
and image interpretation. Finally, the annotated feature maps
were sent via the Internet from NRL to the SACLANTCEN World
Wide Web (WWW) homepage.
|

|
Sample
of Future Data Fusion Products. The image above
shows the modulation of SAR imagery by known Bathymetry
in the area of observation.
|
|
The approach used during the exercise
by the image analyst was
to view the low resolution (1.5 km) imagery, select regions
and features of interest, and post the final products on the
home page for the operator. While this approach served the
demonstration purpose of the RR-96 to disseminate data
over the WWW, it severely restricted the end-user from interacting
with the data or from choosing his/her own regions of interest.
Thus it became clear that a new approach for information extraction
was needed in order to realize the full benefits of high speed
networking and to transform RR from a experimental concept
to operational status.
During
RR-97 ACT used only the ERS-2/SAR imagery which, because of
its high spatial resolution, is by far the most computationally
demanding sensor, requiring massive amounts of storage, network
throughput and processing power. A separate GIS module for
multi-layered data integration and sensor fusion that uses
a "hypercube" architecture was developed. ACT is
presently working on RR97-98 under which additional sensor
modalities will be incorporated.
|