Object-based change detection on multiscale fusion for VHR remote sensing images

This paper presents a novel Object-based context sensitive technique for unsupervised change detection in very high spatial resolution(VHR) remote sensing images. The proposed technique models the scene at different segment levels defining multiscale-level image objects. Multiscale-level image object change features is helpful for improving the discriminability between the changed class and unchanged class. Firstly according to the best classification principle as "homogeneity in class, heterogeneity between class", A set of optimal scales are determined. Then a multiscale level change vector analysis to each pixel of the considered images helps improve the accuracy and the degree of automation, which is implemented on multiscale features fusion. The technique properly analyzes the multiscale-level image objects' context information of the considered spatial position. The adaptive nature of optimal multiscale image objects and their multilevel representation allow one a proper modeling of complex scene in the investigated region. Experimental results confirm the effectiveness of the proposed approach.