The Software System Design Of SAR Image Registration

Image registration is the process of matching and overlaying images which is taken at different times and different viewpoints, by different sensors, or under different conditions. It is a very important issue in the field of image processing problems, and it's the foundation of many image processing technologies. Through an analysis of the existing registration methods, we can divide the registration methods into two categories: the manual methods and the automated methods. In this paper, we put forward three methods of SAR image registration considering the different needs and specific problems of SAR image registration and combined with the existing registration algorithm, part of the algorithm is improved, through experiments demonstrate the feasibility of the algorithm and they will be applied to SAR image registration system designs. The completed works is as follows:1.For SAR image and the optical image registration problem, considering the registration algorithm existing which are obvious shortcomings, they can not solve all the problems between SAR and optical image. Based on the need for practical solutions to the problem, this paper introduces the manual registration, by artificially selecting control points, establishing spatial mapping model for image stitching, we achieve the optical and SAR image registration.2. Based on different needs for SAR image registration, we proposed different registration method. On the issue of SAR image registration's speed, we introduce image registration based on Fourier transform which using FFT to achieve faster,and we can achieve higher registration speed. For a larger deformation of SAR image registration, the paper obtained inspiration from a natural image registration, introduced SIFT algorithm and improved the algorithm to adapt to the registration of SAR images. This paper improves the traditional SIFT feature matching process, by KD-tree + RANSAC strategy to effectively avoid the influence of SAR image speckle noise, so the algorithm can adapt to registration SAR images. In this paper, a lot of simulation experiments have been done to verify the validity and accuracy of the method.3. Based on SAR image application requirements for registration, we complete the design of SAR image registration software system, the system comprising the SAR image reading module, the registration algorithm module and the result stored modules. Using the C ++ language and MFC framework under VS2012 development environment, complete the development of SAR image registration software system. The system is for the SAR image registration,it can support registration function between SAR and optical, between SAR and SAR, also can output registration required model parameters, and also shows the results of registration intermediate in order to meet the user registration visual basic needs.