|
From
the Publisher
By Roland Mangold
Synthetic Aperture Radar - The Mystery Sensor
Aerial film photography is the genesis of remote sensing, and
multispectral aerial and satellite imaging are not new technologies
either, having been extensively used by corporate and government
organizations for over 30 years. However, Synthetic Aperture
Radar (SAR) is quite new to all but the most ardent remote sensing
users, and subsequently it is an information source looking
for its applications.
Granted, Canada's RADARSAT and
the European ERS-1 satellites have been providing terrabytes
of SAR data for the last three or four years on a commercial
basis. Yet there are very few users, and frankly, the operators
of the satellites and the value-added reseller community are
still grappling with the capabilities, applications and markets
for SAR.
Having said that, there are half
a dozen companies out there making up for lost time. One of
the companies spearheading the efforts to resolve the issue
of SAR capabilities, and this should come as a surprise to no
one, is Lockheed Martin Astronautics of Denver. As a 20 percent
shareholder, and exclusive distributor to the U.S. government
of RADARSAT, they have been conducting benchmark studies to
determine just how accurate digital elevation models (DEMs)
can be created from RADARSAT data, and what other applications
can benefit from SAR.
Vexcel Corporation of Boulder
Colo., has participated in the Lockheed Martin studies which
have resulted in demonstrating that 30 meter DEMs can be reliably
produced using stereo images from RADARSAT data. Even greater
accuracies are possible using interferometric processes. Don't
ask me to explain interferometry, suffice it to say that it
is some type of magic. But, the studies have shown that DEMs
can be created with this method providing accuracies between
10-20 meters. However, this level DEM is not repeatable in all
regions or from all platforms.
The Canadians do not keep as
tight a control on the orbit of RADARSAT as the Europeans do
on ERS-1, and this lack of control makes it more difficult to
obtain the desired results from RADARSAT data for interferometric
processing. Also, it seems that tropical, or heavily vegetated
regions cause problems using interferometry for the creation
of DEMs.
Also pursuing the holy grail
of radar, Intermap Technologies Ltd. of Calgary, Alberta, has
been a pioneer in operational SAR remote sensing for over 20
years. Intermap's studies and experience has shown them they
can produce DEMs for virtually any region, in the 20 - 30 meter
range, depending upon terrain. The nearest thing to a consensus
that I was able to come up with is that one can safely expect
30 meter DEMs for anywhere in the world from RADARSAT data.
However, everyone I spoke to about this topic has some type
of caveat: That greater accuracies are achievable given certain
conditions such as terrain, vegetation and processing method.
In my search for a consensus,
I was impressed by a number of revelations about radar that
are not as well publicized as they could be. For example, Intermap
owns and flies the STAR-3i system on a LearJet36. STAR-3i uses
an interferometric SAR which produces DEMs and order of magnitude
better than the current standard provided by the USGS. Intermap's
STAR-3i provides DEM accuracies of 2 meters, image resolution
of 2.5 meters, map scales of up to 1:5,000, and no ground control
points required.
I have seen some of this imagery
an it is spectacular. But, since it is an airborne system, and
on a LearJet no less, the service is expensive. However, Intermap
is considering flying areas on "spec," and providing the data
on an "off-the-shelf" basis. I think a great economic opportunity
lies there, but it is costly and risky to develop.
Another radar revelation was
recently presented to me by Joel Porter, director, Advanced
Remote Sensing Systems, Lockheed Martin Astronautics. Mr. Porter
is heading up Lockheeds Martin's RADARSAT initiatives, and recently
showed me some processed RADARSAT data from the fine beam mode.
It was amazing. At least 7-meter resolution has opened eyes
to applications for RADARSAT that many people have not considered.
In fine beam mode, RADARSAT data is a viable tool for urban
planning, and larger scale mapping. He also showed me some RADARSAT
imagery of the recent flooding in Ohio and Kentucky-there is
no doubt that SAR is an invaluable source of information.
In talking to these firms, I
am very bullish about the potential capabilities and markets
for SAR. However, it is virtually impossible to get a consensus
on the capabilities of SAR. This leads to uncertainty on the
part of the user community, and it opens the potential for "snake
oil" marketing on the part of the vendors.
I have heard numerous complaints
from VARs claiming that the other guys are selling "snake oil"-that
they cannot support the claims they are making. Well, I hate
to break it to you but, SAR will go down the same path as GIS:
he who does the most, and better, marketing will win. There
are countless GIS vendors bemoaning the fact that their GIS
is better that the leaders-but no one is buying their product.
My warning to all those who are providing value-added SAR products
and services: You better throw your hat in the ring, and let
the world know what you can do because if you don't, the "snake
oil" salesman will.
Cheers!
 Roland Mangold
Back
|
