Multi-temporal and multi-view based remote sensing image registration for ground surface objects change monitoring

Remote sensing image registration is a key technology in the field of remote sensing image processing. However, the change of time and the change of view angle will cause the non-rigid distortion and low degree of overlap between the images, and further increase the difficulty of image registration. In order to solve the above problems, a registration algorithm of remote sensing image based on image geometric structural and gray information is presented. By comparing the multi-temporal and multi-angle remote sensing images with four similar algorithms (SIFT, CPD, RSOC, GLMDTPS), our algorithm is superior to other algorithms in the comparison experiment, and shows the optimal registration performance.