| High Resolution Data Presents New Possibilities
With the lanches of several commercial very high resolution satellites on the horizon, space-based imagery is being seen as a new GIS mapping tool.
By Steven D. Hamilton It was not all that long ago that digital orthophotography was "the new kid in town" in terms of image-based mapping. Generally introduced to the commercial GIS market in 1988, digital orthophotography has developed into an almost default data layer for most new mapping projects as of this date. Considering the time that has elapsed from introduction to almost universal acceptance, this is an impressive evolution path for a data product that was initially met with cries of "6 megabyte files! Where am I going to store all of that data? How will I ever use it?"
1996 is coming to a close, but as the year ends, a new mapping and GIS data source is literally coming over the horizon in the form of very high-resolution, low-Earth-orbit commercial imaging satellites. Aimed at providing resolutions and coverage previously reserved for classified military reconnaissance satellites, several companies are actively developing satellite imaging and launch capabilities. Two U.S. companies, EarthWatch Inc. and Space Imaging Inc., are poised to launch the first commercial very high resolution satellites for gathering data suitable for general use in GIS, mapping and land use management.
Several events have occurred in the U.S. which allow the commercial pursuit of high-resolution space-based imagery by U.S. companies. Interestingly enough, most of the impetus for the pursuit of commercial satellite operations was a follow-on to the "Star Wars" Strategic Defense Initiative (SDI), which proved the concept of smaller, less costly lightweight satellites in low-Earth-orbit (350-500 km orbital altitudes). Expanding on that concept, use of "off-the-shelf" components for sensor, storage and communications systems provided an avenue for drastically reducing the cost of assembly, launch and operation of "lightsats" over larger, heavier, more costly traditional satellite platforms.
Previously, limits had been placed on the resolutions attainable by U.S. program satellites to protect military interests. The launching of the Indian Resource Satellites 1C at 5-meter resolution effectively stepped beyond those restrictions and opened the door for expanded U.S. development. A U.S. initiative in 1992 allowed the development of commercial remote sensing capabilities by U.S. companies, with no resolutions or limits specified. WorldView Imaging Corp. applied for a 3-meter license for the EarlyBird series of satellites and was granted the first commercial high-resolution satellite imagery license January of 1993. Lockheed-Martin subsequently applied for a 1-meter license, and following Presidential Directive 23, allowing commercial development of very high resolution satellite capability, was issued a 1-meter license in March of 1994.
Ball Aerospace, later to join WorldView and become EarthWatch, applied for and was granted a 1-meter license in the fall of 1994. During that same period, Lockheed-Martin and E-Systems, now Space Imaging, requested a modification of the original license from 1-meter to .8 meter resolution. EarthWatch followed suit, and was granted an amended license for .82 meter capability for the QuickBird satellite series.
Other satellites planned for launch and operation include Orbital Science / OrbImage Corporations' OrbView, the Indian Resource Satellite 1D, P5 and P6, the Israeli Aircraft Industries/Core Software Technology EROS systems, the Japanese ALOS 2.5 meter system, the French Spot-5 and Helios-2 systems, and the Russian SPIN-2 photographic-media based system, which is currently on-line. The SPIN-2 is an interesting system, in that the commercial data is originally captured at 25 centimeter resolution via a KVR -1000 camera system on photographic emulsion. The film is recovered from the satellite via military means, then scanned and decimated by resampling to a final resolution of 2 meters panchromatic for public consumption.
As of this date, EarthWatch is planning a launch window in late December 1996 for the 3-meter EarlyBird satellite. The EarthWatch sub-meter QuickBird system, as well as the Space Imaging Commercial Remote Sensing Satellite (CRSS) sub-meter system are currently anticipating late 1997/early 1998 launch dates, with launch dates for the other systems falling within the 1998 to 2002 time span. These dates may be subject to change depending on the commercial and operational successes and failures of the previously launched high-resolution systems.
So what do all of these satellites circling overhead mean to the mapping, GIS and land management communities? Will space-based imagery provide an intrinsically higher value in coverage, timeliness, accuracy or spectral capability, and what limitations are inherent with this type of data? Moreover, what costs will be associated with a given scene, unit area, or custom product compared to more traditional image sources, such as aerial photography?
One of the biggest impacts from this constellation of Earth-observation systems will be global coverage of easily accessed and affordable data at multiple resolutions, in both panchromatic and multispectral wavelengths. The systems are designed to cover large areas with compressed revisit times, and to generate high resolution spatial data at greatly reduced cost per unit area compared to conventional aerial photo systems.
Additionally, image fusion will provide color infrared capability by joining coarser-resolution multispectral data (15m EarlyBird/4m QuickBird) with finer resolution panchromatic data (3m EB/1m QB), much the same as remapping Thematic Mapper 30-meter multispectral wavelengths with 10-meter SPOT panchromatic or 5-meter IRS 1C data to arrive at a high spatial resolution with color for enhanced visual utility, interpretability and vegetation classification.
From a technical point of view, satellite imagery offers highly repeatable spatial and temporal revisit capability, depending on the orbital inclination, characteristics and altitude. Up until this point, the limiting factor for GIS applications has been ground pixel resolutions of 5, 10, 20 and 30 meters with IRS, SPOT and Landsat sensors. An exception to this is the Russian SPIN-2 data at 2 meter output resolution, with a somewhat incomplete distribution and order/answer capability, and no imagery available over Eastern Europe or the entirety of the former Soviet Union.
In terms of photogrammetry and GIS applications, such imagery has had limited use and application for anything other than wide-area overview and land use / land cover analysis. For the most part, this is due to the gross resolution of the current generation of sensors in orbit compared to the ground resolutions of most air-photo based GIS projects, which generally fall in the 1/2 to 4-foot pixel resolution range.
The comparatively finer resolutions of the upcoming series of satellites offers access to spatial image data that can be used in a broader range of applications than previously possible. Such applications include refined forest management and species typing, petroleum exploration, utility routing, rural cadastral mapping and photogrammetric mapping support in the 1"= 500', 800' and 1,000' scale range with 3-meter data.
Sub-meter data has distinct applications in mapping support in the 1"=200' and 400' scale ranges, including utility right-of-way planning, encroachment and RF energy studies, landbase quality control and updates, vehicle location, dispatch and routing operations, surface feature extraction, mineral exploration, land use / site planning, urban-area cadastral mapping and updates, watershed modeling and telecommunications surface modeling, using panchromatic or pan-sharpened multispectral data for color or near-IR visual presentation.
Other topical applications for both 3-meter and sub-meter imagery include environmental monitoring, forest fire damage assessment, flooding events and assessment, precision farming crop monitoring, and logging activity permit oversight, all of which benefit from the regular revisit capability of spaceborne platforms, as well as the stereo capture capability of the current SPOT satellites and planned very high resolution EarthWatch, Space Imaging, OrbImage and other announced systems.
There are, of course, limitations to this type of data, similar to early 2- and 4-foot resolution digital orthophoto projects. Just as one could not expect to successfully map water valves from a 2-foot resolution digital orthophoto dataset, manhole covers will not be visible in 3-meter data. Buildings and roadways will be clearly discernible, and will provide good backdrop images for a wide variety of GIS applications. With 1-meter data, pre-marked manhole covers, buildings, roadways and vegetation will be well defined, but water valves and features of that nature will not be captured in the imagery. For projects requiring that level of very fine surface feature display, 1/4, 1/3 or 1/2 foot digital orthos will still be required, allowing for a good "mix and match" of both photographic and space-based imagery and resolutions to meet project requirements, budgets and coverage area.
In essence, we are on the threshold of having commercial systems in space at an altitude of 600 kilometers, or 375 miles above the Earth's surface, being able to image a pre-marked manhole cover, and repeat that coverage on a daily basis with a two-satellite system. This is the same as launching an aircraft every day, and photographing a manhole cover in Grand Junction, Colo., from the Colorado-Kansas border. Not bad, considering that this was once the exclusive domain of spy satellites in terms of coverage and resolution.
We are not at the point of being able to image donuts on a windowsill yet, but just as improvements to computer systems, storage devices and image scanners provided increased resolutions and image utility over time in digital orthophotography, so too will satellite imaging systems improve to provide even greater resolutions and image utility. This is definitely a good start. About the Author:
Steve Hamilton is a photogrammetric consultant at EarthWatch Inc. in Longmont, Colo. He has been involved with raster image data sets and their application in a number of different GIS applications since 1988. He may be reached at 303-702-5571 (phone); 303-682-3848 (fax); or e-mail: [email protected] Back |
