A Performance Analysis of Different Sar Tomographic Processing Techniques for Urban Areas

A single SAR image provides 2D reflectivity along slant range and azimuth coordinates. The reflectivity in elevation is not explicitly imaged, as the backscattered echos from multiple targets at the same range distance are superposed into the same range-azimuth resolution cell despite their different elevation coordinates. Layover distortions occur frequently, particularly in urban areas. SAR tomography extends this conventional 2D imaging scenario to 3D, alleviating the layover problem. It forms a synthetic aperture along the elevation direction using a stack of multiple SAR images acquired from different look angles. Various tomographic inversion techniques have been proposed to extract the elevation profiles. The performance of these techniques is influenced by various practical issues, such as a limited number of baselines, lack of uniformity of baselines, decorrelation and other phase noise, etc. In this paper, we present a comparative performance analysis among some of these inversion techniques, viz. beamforming, singular-value-decomposition approaches and the CAPON spectral estimation.