A Novel SAR Image Segmentation Method Using Run-Length Grouping

Synthetic aperture radar (SAR) image segmentation is the basis of SAR image analysis and understanding. A segmentation method based on the neighbor and near spatial relationship is proposed to reduce the speckle noises influence and improve the segmentation region regularity. The SAR image is coarsely segmented by the scatter intensity of neighbor pixels. For the unregularity of region in coarse segmentation, a region regularity method based on Run-length grouping (RLG) is proposed. The region regularity method adjusts the short run-length segmentation type with the spatial relationship of near pixels. The segmentation region can be more regularity by connecting the broken region, smoothing the tortuous edge and removing patch.