Segmentation of PolSAR Data Based on Mean-Shift and Spectral Graph Partitioning and Its Evaluation

A new segmentation approach of polarimetric synthetic aperture radar(PolSAR)data is pro-posed based on mean shift and spectral graph partitioning.First,Mean-shift algorithm is used to gen-erate the over-segmentation of PolSAR image.In order to extract edge information,we apply a set ofdetectors based on the Wishart distribution with a hypothesis testing method that has fully consideredthe polarization information in PolSAR images.Then,a similarity matrix is constructed based on theover-segmentation results and image edge information.The graph partitioning process is performed u-sing the normalized cut criterion.With this method,we improve the segmentation efficiency of spec-tral graph partitioning based on the over-segmentation results generated by Mean-shift.The quality ofsegmentation results is also improved as a result of the global optimization of spectral graph partitio-ning algorithm.We applied this method on Radarsat-2full polarization images and evaluated the seg-mentation results.The experiement showed that this scheme can realize PolSAR segmentation effec-tively,speed up the original algorithm,and also demonstrates a better result than eCognition’s Multi-resolution segmentation approach.