Constellation analysis for SAR tomography with SVD and parametric reconstruction of vegetated areas

Abstract

In SAR tomography, the available information in the height direction is limited by the number of different tracks that can be acquired in practice. To counterbalance this limitation, electromagnetic scattering models, properly exploiting the available a priori information, should be used, the few possible acquisitions should be accurately selected, and processing algorithms, guaranteeing the reliability of the results, should be employed. With this work an approach is presented for reconstructing the vertical reflectivity distribution of vegetated areas facing the above three points. In particular, an eigenvalue optimization procedure is exploited to design an optimal constellation to be flown, the peculiar features of ground and canopy are accounted for and an effective global optimization algorithm is used to enhance the reliability. The approach is tested with both simulated and real data acquired by the E-SAR system of DLR over the Dornstetten forest test-site.