A stable model-based three-component decomposition approach for polarimetric SAR data

A stable model-based three-component decomposition approach for polarimetric SAR data is proposed in this paper. Two problems in the standard Freeman decomposition: (1) the instability of the decomposition; (2) the emergence of negative powers, have been stated. The proposed approach uses the regularization method to solve the problems, by minimizing a continuous objective function consisting two terms: (1) distance between the measured coherency matrix and the reconstructed one by the decomposition method; (2) the regularization term which implicates prior information. The AIRSAR polarimetric data acquired over San Francisco are decomposited by the proposed method. The results show that the stability is improved by the proposed decomposition approach and the negative powers are also eliminated by introducing constraints to the solution domain of decomposition.