Remote sensing images matching of decomposed affine model via pixel migration

An approach of the edge feature matching of remote sensing images matching based on parameters decomposition of the affine transformation model is presented. Firstly, the adaptability of the affine transformation model used in the remote sensing images matching is analyzed, and parameters of the affine transformation model are decomposed. Secondly, the sum of squared gradient (SSG) similarity is constructed based on the edge feature in images, and the parameters of the special transformation model between the matching remote sensing images are solved with the Genetic Algorithms (GA) which is used to obtain the global optimum solution of the similarity. Finally, the performance of the approach is analyzed in theory and validated with remote sensing images matching experiments.