Model-based polarimetric SAR calibration method using forest and surface scattering targets

This paper proposes a new polarimetric Synthetic Aperture Radar (PolSAR) calibration method that applies an incoherent decomposition model to the uncalibrated covariance data measured for the forest and surface, and determines the polarimetric distortion matrix (PDM). The Freeman-Durden model [1] is used to express the polarization dependent signal reflection from and penetration through the forest. Non-linear equations built for uncalibrated PolSAR data are solved iteratively. This method is applicable to the lower frequency SAR that associates with the polarization dependent signal penetration through forest canopies. Using the time series Phased Array Type L-band SAR (PALSAR) data acquired for the Amazon rainforest for around 3 years, we confirm that the proposed method succeeds the PDM estimation and that the calibrated data preserve the polarimetric performance on HH-VV orthogonality, low cross-talks, and ideal polarimetric signature for the corner reflector.