The Research Of Remote Sensing Image Registration Method Based On Orthogonal Learning Differential Evolution Algorithm

Image registration is a critical technology in image processing. The aim of image registration is to match the images which are obtained on different conditions, at different time or from different sensors. The technology of image registration has been widely applied in many areas, such as computer vision, computer graphics, medical image analysis and mechanics of materials.Duo to the complexity of remote sensing image and the images taken from different sensors are always exist distortion. It is difficult to register the remote sensing images. Nowadays, with the deeply research of image registration, there is a growing requirement for the accuracy of image registration. After the research of the existence of image registration techniques and the analysis of the experiments, this paper proposed two efficient methods for remote sensing image registration.The first method combines the feature-based and area-based method. This method has overcome the shortage of feature-based method and area-based method. At first, we get the most similar area in two images. And the corner features in the most similar area can be abstracted by using the Harris corner detector. Then we can get the best transform parameters after registering the most similar area. And the two images can be matched with the best transform parameters. This method is an improved image registration method based on mutual information and Harris corner detector. The experiments show that our method is efficient for registering the remote sensing images.The second method is an Area-based method. The area-based method usually requires optimizing the similarity metric between the reference image and the target image. The orthogonal learning(OL) strategy is efficient when searching in complex problem spaces. Differential evolution(DE) is a heuristic algorithm. It has shown to be efficient in solving the remote sensing image registration problem. So we propose an orthogonal learning differential evolution algorithm for remote sensing image registration. The OLDE method uses the OL strategy to guide the DE algorithm to select promising direction towards the global optimum. Experiments show that the OLDE method is more robust and efficient than genetic algorithm(GA), particle swarm optimization(PSO) and differential evolution(DE) methods for registering remote sensing images.