Polarimetric EM scattering and information retrieval in SAR remote sensing

The advance of polarimetric SAR imagery technology. such as SIR-C and RADARSAT-2 has promoted extensive study and applications for information retrieval from polarimetric scattering measurements. This paper briefly reports our research on three issues: (1) solutions of the Mueller matrix, the eigenvalues of the coherency matrix and information entropy to directly relate with measurements of the co-polarized and cross-polarized scattering in SAR imagery; (2) an inversion approach for digital elevation mapping (DEM) by using a single pass instead of two-pass or interferometric SAR (INSAR), of polarimetric SAR data; (3) an algorithm of two-threshold EM (expectation maximisation) and MRF (Markov random field) for automatic analysis of the context change detection in the urban area from multi-year SAR observations.