Airborne Radar Interferometric Repeat-Pass Processing
- Tuesday, 01 March 2011
Improvements have resolved some of the earlier difficulties.
Earth science research often requires crustal deformation measurements at a variety of time scales, from seconds to decades. Although satellites have been used for repeat-track interferometric (RTI) synthetic-aperture-radar (SAR) mapping for close to 20 years, RTI is much more difficult to implement from an airborne platform owing to the irregular trajectory of the aircraft compared with microwave imaging radar wavelengths. Two basic requirements for robust airborne repeat-pass radar interferometry include the ability to fly the platform to a desired trajectory within a narrow tube and the ability to have the radar beam pointed in a desired direction to a fraction of a beam width. Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) is equipped with a precision auto pilot developed by NASA Dryden that allows the platform, a Gulfstream III, to nominally fly within a 5 m diameter tube and with an electronically scanned antenna to position the radar beam to a fraction of a beam width based on INU (inertial navigation unit) attitude angle measurements.
UAVSAR is also equipped with a set of GPS receivers that coupled with INU measurements are used to determine the antenna position to a high degree of accuracy on a pulse-to-pulse basis. The relative position error within a flight track is measured to a small fraction of a wavelength as is required for image formation; however, the absolute accuracy of the position measurements is in the 2–10 cm range limited by the accuracy of post flight processed differential GPS data. In order to make repeat-pass radar interferometric deformation maps suitable for geophysical interpretation, the relative position between the platform positions at the time a point is imaged for a pair of repeat-pass observations, i.e. the interferometric baseline, needs to be known to the millimeter level. Bridging the gap from the 2–10 cm position accuracy of the metrology system to the desired millimeter relative position accuracy uses information contained within the SAR imagery. Image-based residual motion recovery algorithms using radar imagery have been developed for airborne (and spaceborne) platforms previously; however, these algorithms have not been employed for systems using electronically scanned arrays.
The UAVSAR repeat-pass processing software called JPRP, has been specifically designed to permit the generation of radar repeat-pass interferograms of surface deformation suitable for geophysical interpretation. This software automatically processes and co-registers data from multiple flight lines. JPRP has been modified from previous codes to do motion compensation and image formation for airborne systems employing electronically scanned antennas. Since UAVSAR employs an onboard Block Floating Point Quantization (BFPQ) scheme whereby 12 bit recorded radar echoes can be compressed to M bits where M ranges from 2–10 and is a radar commandable parameter, the JPRP includes appropriate BFPQ decoding algorithms. JPRP also includes code and algorithms for computing the repeat-pass interferometric baselines for airborne data using a priori GPS and INU data and for estimating refined residual baselines from the radar imagery needed for resolving dynamic residual baselines at the subcentimeter level. These algorithms have been adapted to work for systems employing electronically scanned antennas. The program includes advanced repeat-pass motion compensation algorithms that include subaperture, terrain-dependent motion compensation for range/Doppler, or wave domain processing. Also, a new algorithm that is computationally efficient was developed for topographic fringe removal and geolocation.