A multi-source remote sensing image registration algorithm based on local adaptive similarity analysis and improved cloud particle swarm model

An image registration method based on feature similarity analysis combined with improved adaptive cloud particle model is proposed. Firstly, the SURF algorithm is used to extract the feature points of the image. Innovatively using the spatial gray distance to match the feature vectors. Then we disturb characteristic coordinates to increase micro displacement, finally we introduce the improved particle swarm optimization algorithm with quantum behavior. The whole spatial transformation parameters are classified and the iterative evolution of different weights is carried out, the velocity evolution formula of cloud particle swarm is optimized by means of local and global transformation parameters. It shows that the proposed method can effectively avoid the local optimization of the transform parameters by the multi resolution remote sensing image registration experiments. Compared with the original method, the new method has a higher matching accuracy. In addition, the improved adaptive particle swarm optimization algorithm can effectively eliminate the false registration.