Research Of Multispectral SAR Image Fusion

Multi-spectral images provide info with redundancy, supplementarity and cooperativity. The aim of fusion is: to get more complete, reliable, precise estimation and adjudgement from fusion image than from single image. Based on application, image spatial registration, multi-spectral image fusion enhancement, information restoring and optimizing fusion and the evaluation technology were studied in this paper.On studying image registration, a new layered multistage approach of image registration from low precision to high precision is presented. It solved the shortcoming of traditional registering methods, uniting registering velocity and precision. The approach is applied to multi-spectral SAR image registration. On studying multi-spectral SAR image fusion, approaches based on wavelet transform were improved and a new approach based on edge enhancement is presented. Firstly, the sub-images from wavelet transforming were fused. Then, the edge of high-spectral SAR image was fused into the first result. After twice fusing, the method improves blurred edge. An information restoring and optimizing fusion approach is advanced, too. It deals with each feature of wavelet packet transforming. The approach forms an image of more abundant details and more proper for eyesight. The result can achieve ' stereo' data. On studying image classification fusion, improved nerve net(NN) is used to classify the original SAR image and fuse the pre-classified SAR image. On studying integrated evaluating technique of fusion approach, performance evaluation of image fusion is presented. The quantitative and qualitative evaluating criterion is also summarized in this paper. Linked with the specific application, a evaluation method based on blurred synthesis evaluation was used to demonstrate the performance of fusion approaches.Beneficial results were gained after experimented with actual multi-spectral SAR image. The results proved that using above methods can obtain useful results. And these methods solved different requirement of multi-spectral SAR image fusion.