Investigation on compact polarimetry in SAR interferometry using CETC38 airborne data

In this paper we investigate the classification performance of the compact polarimetric interferometric SAR (C-PolInSAR). The stressed compact modes are π/4 mode and CTLR mode, due to DCP mode equivalent to CTLR mode in theory. First, we provide a state-of-art of the C-PolInSAR modes, and present the different reconstruction algorithms aiming at recovering the full PolInSAR information from the observed C-PolInSAR dataset. Then, an assessment on the classification performance of the C-PolInSAR modes are carried out based on the Wishart ML classifier. Our emphasis is comparisons between the classification results of the C-PolInSAR modes and the corresponding reconstruction modes, versus the dual PolInSAR (D-PolInSAR) and the full PolInSAR (F-PolInSAR) systems. We test on the first PolInSAR dataset in China: CETC38 airborne data. Experimental results show that the C-PolInSAR has a promising application in the terrain classification, and it is an efficient alternative to the F-PolInSAR and the D-PolInSAR.