Auto-Weighted Structured Graph-Based Regression Method for Heterogeneous Change Detection

Change detection using heterogeneous remote sensing images is an increasingly interesting and very challenging topic. To make the comparable, some graph-based methods have been proposed, which first construct a graph for image to capture structure information then use obtain structural changes between images. Nonetheless, previous change approaches are insufficient in representing exploiting structure. address these issues, this paper, we propose auto-weighted structured (AWSG)-based regression method detection, mainly consists of two processes: learning AWSG perform detect changes. In process, self-conducted weighting strategy employed more robust, local global combined informative. transform one domain other by learned AWSG, where high-order neighbor hidden exploited better image. Experimental results comparisons on four real datasets with seven state-of-the-art demonstrate effectiveness proposed approach.