Multi-source Image Fusion Research

Image fusion is a significant and promising research branch of information combination, which is the process of combining multiple sources images into a single one with advantages of different sensors. The actual fusion process can take place at different levels of information representation. A common categorization is to distinguish between pixel, feature and decision level. Complementary and redundant information from different input images can be used to get a good description of the scene. As the basis of image fusion, pixel-level image fusion is considered to influence the fusion result dramatically.This dissertation laid emphasis on the following points:1. In consideration of the current problem, seeking a general framework of image fusion, a new fusion framework has been proposed on the basis of multi-resolution decomposition. Major current fusion schemes in the literature have been covered by it. Meanwhile, the framework of image fusion can be taken as a guide for relevant research in the future.2. Through the fundamental research, multi-wavelets theory has been applied to the process of image fusion. According to the need of multi-focus image fusion, a new fusion approach on the basis of multi-wavelets decomposition has been proposed, which combines the appropriate fusion scheme and methods. It proves to have a better and stable effect than traditional MR fusion approaches.