| AIRBORNE
Penetrating Through the Clouds
An introduction to Synthitic Aperture Radar
By Barton D. Huxtable, Ph.D. Introduction
Synthetic Aperture Radar (SAR) is a method of microwave remote sensing where the motion of the radar is used to improve the image resolution in the direction of the moving radar antenna. SAR imaging instruments support a wide range of commercial applications due to their ability to achieve very fine resolution from great distances while covering large areas of the Earth.
SAR instruments can penetrate through clouds, haze, smoke, and vegetation. The active nature of SAR sensors means they can operate equally well in all lighting conditions, not requiring the smoothing normally necessary for optical imaging due to sun position or sun glint off reflective surfaces. History of SAR
The beginnings of current SAR instruments date back to the 1950s. The first SAR was placed in space in 1978. In spite of this long heritage, SAR data has not made the same popular appearance in the commercial market as high-resolution electro-optical image sensors carried on satellite such as Landsat and SPOT. Several factors have contributed to this slow adoption. One being the computationally intensive transformation required to form images from the raw radar data. Unlike electro-optical raw data, SAR raw data is spatially and temporally incoherent, looking like late night TV snow. Only recent advances in computer processing have made SAR image formation a "desktop" process.
The genesis of the synthetic aperture concept is relatively recent, and is generally attributed to the work of Carl Wiley of the Goodyear Aircraft Corporation. In the early 1950s, Wiley observed that a one-to-one correspondence exists between the along-track coordinate of a reflecting object (being linearly traversed by a radar beam) and the instantaneous Doppler shift of the signal reflected to the radar by that object. He concluded that a frequency analysis of the reflected signals could enable finer along-track width of the physical beam itself. This "Doppler beam-sharpening" concept was not only exploited by Goodyear, but by a group at the University of Illinois. An experimental demonstration of the beam-sharpening concept was carried out by the Illinois group in 1953 through the use of airborne coherent X-band pulsed radar, "boxcar" circuitry, a tape recorder, and a frequency analyzer.
One major problem that was recognized quite early, was the implementation of a practical data processor that could accept wide-band radar signals and carry out the necessary Doppler-frequency processing to form radar images. During 1953, the University of Michigan performed a study called Project Wolverine, which served as a point of departure for expanded efforts for fine-resolution, terrain-imaging radar. Industrial developments at Goodyear, Hughes, and Westinghouse followed. SAR Applications
The two most promising applications requiring SAR data are the generation of digital elevation models (DEMs) by processing interferometric SAR (IFSAR) data, and the ability to identify surface characteristics almost without regard to weather, foliage, smoke, snow, or other obscurants that render visible satellite images of less or no use. SAR usually receives a strong signal from linear features, especially those that are naturally corner-cube shaped. Light aircraft tail sections intrinsically have this shape, and almost always survive a crash. This feature allows, for example, the identification of the tail section of an aircraft during the night or day, in a storm, or even a sunny afternoon.
Recent technology developments have precipitated a SAR renaissance in space. An increasing number of SAR sensors in space and the expanding availability of computers capable of processing the data and its derivatives are allowing access to SAR data by more end users. The launch of ERS-1 ushered in a decade of SAR systems in space. Combined with the expected SAR systems in the near future, the current constellation of SAR instruments provide the seeds for a rich data source with each system adding unique characteristics in terms of radar design and mission data collection. SAR Image Formation
Understanding when, and where, an image is formed in a synthetic aperture radar system is the pivotal "ah ha" necessary to remove a large part of the mystery surrounding synthetic aperture radar sensors, raw data, and processed data. As alluded, SAR technology requires various components and functions to produce a coherent image. In all cases, a target is required that displays some correlation between its physical characteristics and its effect upon the radar returns being collected by the sensor.
The concept of an optical target reflecting or absorbing specific colors of light is a very familiar concept. Similarly, the wavelengths of energy used by synthetic aperture radar sensors interact with the target in a characteristic way such that the correlation among the radar energy collected and some physical characteristic of the target is recognizable.
A primary difference between optical sensors and radar-based sensors is that radar-based sensors provide their own energy source. Optical sensors utilize the vast amount of energy from the Sun that is reflected off the targets. In the radar wavelengths of interest there is almost no energy available from natural sources such as the Sun. This necessitates the use of an artificial source of radar energy to "illuminate" the target, similar to using a flashlight in a dark room.
Conventional optical systems are composed of a target, reflected energy from the target moving toward the optical system, focusing lens, recording device, and a display device. This system provides as its output a recognizable image. These systems record focused images (on film or digital focal planes) as opposed to synthetic aperture systems that record the energy not in the image plane, but in the Fourier Transform plane, which is in front of where a lens might be placed. Collection of the raw data in front of the image plane yields raw data that is unrecognizable without significant processing to focus it. This data is additionally scrambled due to the collection process that records information about adjacent pixels on the ground in non-adjacent positions on the recording media.
Synthetic radar processing computationally reorders and focuses the raw radar data such that an image is formed. This process achieves numerically for radar systems what is accomplished with a physical lens for optical systems. Once both systems output data of similar order and focus, they both yield images that are recognizable.
It is at this departure point where substantive application processing can be done. Both types of data can at this point be subject to algorithms that may extract, summarize, and display important characteristics about the target. Optical data is routinely used to classify and measure land use, condition of vegetation, identity of surface or vegetation type, and so on. Similar analysis exploiting the unique characteristics of radar energy and its interaction with the target can reveal information such as soil moisture, surface texture, elevation, land use, and so on. An additional benefit of radar-based systems is that the energy used by them penetrates fog, haze, smoke, and vegetation, unlike optical systems. Since they are active systems they operate in any lighting condition, not needing corrections for Sun angle or glint.
User Systems, Incorporated, in Gambrills, Maryland, has conducted studies, from which a new paradigm is being set forth for processing all radar imagery to a state, referred to as fully focused, single look, complex imagery. Established in 1982, User Systems has developed PROSAR as the tool to ingest raw radar data and output fully focused, single look, complex imagery. It also is in a state that contains all the information inherent in the data, making it an ideal jumping off point for performing a variety of application specific analyses. Typically, if the application specific analysis is combined with the process that reorders and focuses the data, this makes it impossible to conduct further analysis on the imagery without performing the most computationally difficult task relative to radar imagery every time a new analysis is desired. This would be somewhat analogous to incorporating specific applications algorithms into an optical system and only recording the data after the application processing. Conclusion
With the advances in both SAR technology, computer hardware, and analysis programs, both non real-time and near real-time applications for SAR technology are expanding. Multidimensional SAR's incorporating simultaneous frequency and full polarized observations will open the door to sub-pixel target detection and classification that will expand the range of commercial and military applications. Search and rescue is an outstanding single example of the utility of this new technology. NASA's Goddard Space Flight Center has successfully demonstrated the utility of low frequency SAR image processing, image and map geo-registration and referencing, and automatic target detection and ranking made possible through the use of SAR sensors. Other applications include drug interdiction plus geophysical products such as soil moisture and canopy mapping. Evaluation of moving targets and elevation mapping is also possible via multiaperture data obtained using techniques called interferometry. As radar and signal processing technologies advance, many new and wonderful applications will surely emerge.
In addition to their central business of developing signal processing systems for airborne and space-based radar, USI is currently developing the Synthetic Aperture Radar image processing systems for the Defense Nuclear Agency's Open Skies Treaty Aircraft, NASA/Goddard Space Flight Center's Search and Rescue Program, and several commercial customers. With nearly twenty years of direct experience providing radar signal processing systems and software, User Systems was recently selected for award under NASA's Earth Observations Commercial Application Program to produce marine winds products for commercial weather broadcasts using space-based radar scatterometry data. This work is presently being completed. About the Author:
Barton D. Huxtable, Ph.D., directs User System's core business of producing user-friendly, portable, high performance SAR processing systems. He may be reached at 301-912-3022 (telephone), or [email protected] (e-mail). Back |
