Research And Application On Automatic Image Mosaic

Image mosaic technology is an important research field of image processing, and has become a research focus on the computer vision and computer graphics, and has important utility value in the fields of remote sensing image processing, medical image analysis, and so on.This thesis analyzes and researches major aspects of image mosaic, depending on the steps of image mosaic based on feature point. Its specific work includes:1. This thesis presents the process of implementing image mosaic technique, summarizes its research significance, research status at home and abroad, main application areas and major existing methods comprehensively. Paper also introduces how to get the images for mosaic, and the common image transformation models.2. On the basis of phase correlation method, paper presents an algorithm which realizes automatic sorting image sequence from the chaotic image sequence according to the scenes content. The algorithm avoids manual intervention during the mosaic process.3. On the basis of ideas of image break and neighboring point eliminating, an adaptive Harris corner detection algorithm is presented. Comparing with the traditional algorithm, the new algorithm avoids setting a threshold, makes the corners are well-proportioned distributing, and avoids too many corners stay together effectively. The algorithm resolved the issue of mistaken image matching, what is resulted from the unreasonable and uneven distribution of the corners.4. On the basis of analyzing the performance of the algorithm and combining the idea of gradient of disparity constraint, a RANSAC purify matched corners based on gradient of disparity constraint is proposed. Comparing with the traditional algorithm, the new algorithm effectively improved the efficiency of more than 40 percent, and not reduced the accuracy of the algorithm at the same time. This algorithm is used to purify the matching points after get the match points by coarse matching algorithm, and it resolves the issue of imprecise image transformation matrix which gets from coarse matching corners.5. Using the image fusion technology based on the Overlapping Areas Linear Transition, achieves the requirement of seamless mosaic.In conclusion, this research work involves all major aspects of the image mosaic process, and the experimental results of each researched algorithm are intending and good.