Research Of The Image Mosaic Method Based On Harris Corner Detection

Digital image mosaic technique is a hot research, has been widely used in computer vision, computer graphics, medical image processing, remote sensing, aerospace field, military field and etc.Based on the existing image mosaic algorithm, researchers continue to try to improve the algorithm structure, make up the deficiency of the original algorithm to improve the effectiveness of the image mosaic process.In this paper, based on the basic theory and gradually understand the image mosaic method, and to use spliced Matlab 6.5 programming for image mosaic with the way the experiment. Try to find an improved algorithm can improve effectiveness and practicality. Then, using experiments to verify, analyze, test results prove that the algorithm is effective and practical.Summary of the chapters are as follows:Chapter 1, Introduction.This chapter detalis the image mosaic technology. Lists the status of domestic and foreign research in related fields , on a brief summary of research results. Clarify the purpose of this paper topics and significance, and finally to set out the major work done in this paper and the overall framework of the thesis structure.Chapter2, image regiestration and image comosition.This chapter in accordance with the image pro-preprocessing , image registration and image comosition. On a clear definition of each process, the system described the classification of image mosaic technology. Among them, a brief introduction of the image pre-processing of the work process. Focuses on image registration as the core part of image mosaic, in accordance with the principle of the different existing algorithms, algorithm can be devided into two categories: feature-based image registration trchniques and region-based image registration trchniques, the working principle and characteristics of existing algorithms are introduced in detail and cited algorithms problems.Compare the various advantages and disadvantages of different algorithms. For image comosition, the process of the existing algorithms are relatively well, the paper on the common principle of all kinds of fusion algorithms are introduced briefly.Chapter 3, corner detection algorithm.This chapter introduces the corner detection algorithm works, corner detection accuracy of measurements, the existing classical algorithm working principle and characteristics, and summarizes the advantages and disadvantages of various algorithms. Which focuses on Harris corner detection algorithm working principle and characteristics of the classical algorithm summarized the shortcomings in the application.In the classical theory of algorithms, based of the algorithm of CRF(Corner Response Function), select a different parameter k, will result in bias. Select a new CRF, to avoid the k value is selected on the corner detection process which would improve the image mosaic practicality.Chapter 4, image mosaic experiment.For the preliminary study on image mosaic, using experimental methods to acalidate. Existing simulation software, including Matlab and VC + +, this paper used Matlab6.5. Designing experiment to demonstrate the improved corner detection algorithm is effectiveness and practical . Finally , analyzed the expetimental results.Chapter 5, summary and outlook.Summarize the process of this research work and the problems encountered, to the need to improve the deficiency, and the image mosaic process research and development prospects.After more than a year of image mosaic techniques, basically involve the existing algorithms, for this techniques has a deep understanding, but for this area, the work of this paper is only a small part. Practical applications, there are many issues to researcher to further explore the ways to solve. As time is limited, this thesis does not continue in-depth research, I will continue to focus on the latest research in this field, spare no pains to learn related technologies, become a expert in this field, strive for early settlement of existing problems.