Research On Wavelet Transform Based Image Registration And Fusion

Image fusion is the technique for integrating complementary and redundant information of multi-source images of the same scene to obtain a composite image for observation or subsequent processing. It has been widely used in many fields such as military application, computer vision, remote sensing, robot engineering, medical imaging, etc. The main contributions of this dissertation are summarized as follows:1. Since image registration is the prerequisite of image fusion, the image registration problems are first studied in this dissertation. A new method to solve the registration of serial images is implemented. The source images are decomposed by wavelet transform. The low frequency bands of the source images at each decomposition level are used to estimate the registration parameters. We improved the joint histogram based registration algorithm by using wavelet transform. Experiment results show that the proposed algorithm works well in registering serial images and can increase the speed of registration.2. A contour-based multisensor image registration algorithm is implemented. Salient contours are firstly extracted from the reference image and the input image respectively. Contour matching is performed after contour extraction. Matched open contour pairs are transferred into matched closed contour pairs by linking the two endpoints of open contour together with a line section. Control point pairs are extracted from matched closed contour pairs. The registration parameters are calculated according to the control point pairs.3. A discrete wavelet frame transform based multifocus image fusion algorithm is proposed. The source images are first decomposed by the discrete wavelet frame transform. Then the clarity measure of each pixel is calculated according to the decomposition coefficients. Fused image is obtained by combining all the clearest pixels in the source images. Experiments show that the proposed algorithm works well in gray multifocus image fusion.4. A discrete wavelet frame transform based color multifocus image fusion algorithm is proposed. The intensity component of each multifocus image is decomposed by discrete wavelet frame transform. According to the discrete wavelet frame transform coefficients, the clarity measure of each pixel in each multifocus image is calculated. The fusion process is performed by selecting the clearest pixels in the multifocus images. Experiments show that the proposed algorithm works well in color multifocus image fusion.