Research On Image Mosaic Based On SIFT Feature Matching

Image mosaic technology has been widely applied in many domains, because it has solved the contradiction between the wide field of vision and high-resolution. For example, the analysis of medical images, the panoramic video surveillance system, integrated large-scale aerial or satellite photos, construct the virtual reality scene and so on. So image mosaic technology becomes a hot spot research question in image processing field.This thesis introduces the research background of the image mosaic technology, current status and applications. Then summarizes several main methods and analyzes their performance and problems for registration and fusion which are the key steps in the image mosaic. This thesis focuses on the image mosaic based on SIFT feature matching, mainly to completes the following aspects:(1) Learning and summarizing the existing image mosaic technologies based on feature points. Focuses on the idea and the implementation process, analyzes the advantages and disadvantages.(2) In order to solve the problem that SIFT features matching algorithm exist many errors in the structure and texture similar image matching, which would lead to non-ideal effect of image mosaic, The thesis improves the traditional method. The improved method use color information and spatial information to filter the matching point-pairs, so as to achieves better accuracy matching points. Comparing the results of the improved method and the traditional method, the experimental results show that the proposed algorithm improved the robustness and the effect of matching.(3) In order to solve the problem of image mosaic with exposure differences. Based on the characteristic of Lab color space, we transform the image from RGB color space to Lab color space, deal with the L channel for adjusting the brightness in order to solve the exposure differences, and then on the basis of it, we propose an idea to solve the movement ghosting. Comparing with the muti-resolution splines and poisson fusion which are commonly used to eliminate the exposure differences, this thesis method is more simple and faster.