Novel Change Detection in SAR Imagery Using Local Connectivity

Most change-detection techniques in synthetic aperture radar (SAR) imagery are based on the analysis of the difference image with a pixel-level decision approach. However, the pixel-level decision approach would cause a noisy change-detection map, with holes in connected regions and jagged boundaries. In this letter, we propose a novel change-detection method to deal with the problem of the pixel-level decision approach by considering local connectivity. We first get an initial change-detection result with an improved Gustafson–Kessel clustering algorithm using local spatial information and then refine the initial result through region-of-interest extraction and consideration of local connectivity of changed areas. Experimental results on real SAR image data sets demonstrate that the proposed method outperforms the related ones for change detection.