A Multi-kernel Joint Sparse Graph for SAR Image Segmentation

Recently, more and more attention has been drawn on the study of sparse graph-based classification with respect to pattern recognition and computer vision. Sparse self-representation method features good category distinguishing performance, noise robustness, and data-adaptiveness. In this paper, a multi-kernel joint sparse graph (MKJS-graph) is proposed to segment synthetic aperture radar (SAR) images. At first, an SAR image is over-segmented to many superpixels. Then, a new multi-kernel sparse representation (MKSR) model is used to express sparsely multiple features of the superpixels in high-dimensional projection space, which can reflect the global similarity of superpixels. Moreover, the local neighborhood spatial correlation of superpixels is combined with the global similarity of that to improve the segmentation performance by formulating the adjacent matrix for MKJS-graph. Integration of the global and local structures of the superpixels provides the MKJS-graph with favorable category distinguishing ability on segmenting SAR images polluted by speckle noise. The simulated, Ku-band, and
$X$
-band SAR images are tested through a series of experiments, and the results indicate that the proposed method is more competitive than other state-of-the-art algorithms in SAR image segmentation.