Artificial Intelligence and Security: 6th International Conference, ICAIS 2020, Hohhot, China, July 17–20, 2020, Proceedings, Part II

Kernel fuzzy c-means clustering with spatial constraints (KFCM S) is one of the most convenient and effective algorithms for change detection in synthetic aperture radar (SAR) images. However, this algorithm exists problems of weak noise-immunity and detailpreserving on account of the failure to use spatial neighborhood information. In order to overcome above problems, this paper proposed an algorithm using bilateral filtering and large scale median filtering instead of the original spatial constraints. In particular, the approach uses different calculation methods of constraint terms at different locations of image. The bilateral filtering value is used as spatial neighborhood information at the boundary region for preserving the boundary information while the large scale median filtering value is used at the non boundary region for facilitating noise removal. In this paper, 3 remote sensing datasets are used to verify the proposed approach, and the results show that the proposed approach improves the accuracy of remote sensing image change detection.