Terrain topography measurement using multipass polarimetric synthetic aperture radar data

A method has been investigated for the measurement of topography using airborne fully polarimetric synthetic aperture radar (SAR) data. Terrain slopes in both the range and azimuthal directions have been estimated using multipass flight geometries. Using these slope values, the Poisson equation was then solved to create a Digital Elevation Model (DEM) of the terrain topography. The method measures polarimetric orientation angles which are then converted into terrain slopes in the azimuthal direction. The conversion of these orientation angles into terrain slopes requires additional knowledge of the radar look angle and the range direction terrain slopes. The solution for slopes is, therefore, a problem coupled between the range and azimuthal directions. For specialized multipass flight geometries these orthogonal terrain slopes are solved for, and maps of terrain slopes are produced. In particular, the processing of two‐pass orthogonal and two‐pass antiparallel (headings Θ and Θ+ 180°) NASA ‐ Jet Propulsion Laboratory airborne SAR data sets has been carried out for an area in central California. When orthogonal slopes are derived using either of these data sets, a digital elevation model may be generated. The L band, polarimetric SAR (POLSAR) DEM created by this solution is compared to a coregistered C band, interferometric SAR (IFSAR) DEM. Similar comparisons are made for terrain slopes in the azimuthal ‐ range directions which are generated by the POLSAR and IFSAR elevation data. The polarimetric SAR, operating from an aircraft, or satellite, in a strip‐mapping mode, is capable of measuring terrain topography for large areas provided that phase‐preserving fully polarimetric data are taken. Polarimetric SAR data are also widely used for studies of crop classification, surface roughness, biomass density, and soil moisture content. All of these studies are adversely affected by scattering changes attributable to topography. The technique investigated here potentially offers a means to correct for these effects by making simultaneous, coregistered estimates of local orthogonal terrain slopes.