Nonlocal Filtering for Polarmetric SAR Data based on Scattering Model

摘要

Speckle complicates the polarimetric SAR data interpretation problem and reduces the accuracy of image segmentation and classification. Based on the scattering model, a nonlocal polarimetric SAR data filtering method is proposed. The method selects homogeneous pixels by likelihood-ratio test in unit of image patches, optimizes the weights of homogeneous pixels by the polarimetric scattering similarity, and finally performs the weighted average filtering. Because the statistical information and scattering similarity of the polarimetric SAR data are applied simultaneously with nonlocal filtering, the proposed algorithm is improved in the performance of both the fine-structure preservation and speckle reduction. Finally, the effectiveness of the proposed algorithm is validated using the airborne polarimetric SAR data acquired by the N-SAR system.