A novel SAR change detection based on radon transform and super-pixel segmentation

Traditional pixel-based change detection methods are undertaken using the pixel as the research unit. These methods may give high false alarm rate and broken areas, because they don´t use semantic information. In order to solve this problem, we present a novel change detection method which is based on radon transform and super-pixel segmentation. First, radon transform is used to achieve a stable difference map, and the Otsu algorithm is used to gain the initial change mask. Second, in order to reduce false alarm rate and preserve edges between changed class and unchanged class, the simple linear iterative clustering(SLIC) super-pixel segmentation is introduced into SAR image segmentation. Then, the spatial-based change detection scheme is utilized for the fusion of initial change mask and segmentation result. Two radarsat-2 images of Suzhou, China, acquired on April 9,2009 and June 15, 2010 are used for our experiment. Experiment results show that the proposed method is effective for SAR image change detection in terms of edge preservation and change detection rate.