Multiobjective sparse unmixing based hyperspectral change detection

Xiangming Jiang, Tianqi Gao, Maoguo Gong, Fenlong Jiang, Mingyang Zhang, Jieyi Liu

Abstract: Highlights•An unconstrained sparse unmixing based change detection model was deduced and an evolutionary multiobjective optimization method was adopted to solve it.•An ensemble decision strategy was designed for the Pareto optimal solutions obtained by multiobjective optimization method to generate an aggressive pixel-level change map.•As compared to mainstream change detection methods, the proposed method can detect pixel-level as well as subpixel-level changes which are more practical in real applications.