Research Of Electronic Video Stabilization Algorithms Based On Feature Points Matching

Currently, the technology of EIS(Electronic Image Stabilization, EIS) is an important research area of video processing and computer vision. Because of its advantages of high image stabilizing precision, compact size, light weight, low power consumption and reasonable price, it will have extensive application prospect in military affairs and civil cameras. The key technology of the EIS includes motion estimation and motion compensation, and the motion estimation between adjacent frames is the foundation of EIS. Thus, a lot of researchs are carried out and many algorithms of motion estimation are proposed. Based on the analysis of the advantages and disadvantages of the typical algorithms, a method based on improved feature points matching for video stabilization is presented, aiming at solving the clustered issues and depth of field problem of the feature points. The main research works in this thesis are as follows:(1) The research progress at home and abroad and the common methods of EIS are reviewed and summarized. In this thesis, the basic principles and system structure of the electronic video stabilization are described, and the key technology and critical issues of the realization are discussed. Several typical algorithms of EIS and their advantages and disadvantages are introduced and analyzed. It emphatically points out that the algorithm based on feature points matching is considered as the important method, because of its perfect performance to deal with the complex dither between video frames. And this kind of algorithm is adopted in the thesis.(2) How to improve the efficiency and accuracy of the motion estimation is the research emphasis. In order to solving the clustered issues and depth of field problem of the feature points, a high-accuracy and high-efficiency electronic video stabilization method based on improved feature points mactching is presented. In the proposed method, feature points are firstly detected from the frames via the fast Harris operator, and then the minimum distance constraint mechanism is used to solve the problem of feature points crowding in a local region. Feature points are then utilized to establish and guide feature blocks matching, and an improved random sample consensus (RANSAC) algorithm is used to screen for matches of high quality from local motion disturbance. Finally, in order to solve the problem brought by different depth of field, a significant priority method is applied to the global motion parameters caculation to further improve the precision of video stabilization. Experimental results show that the proposed method can not only achieve fast speed but also obtain more accurate global motion parameters, making a good balance.(3) When the movement of the camera include random jitter and intentional motion, we must distinguish them firstly. According to the basic principle of Kalman filter and the mathematical model proposed in the thesis, a Kalman filter used to distinguish camera scan from dithering by smoothing the global motion vector are designed. Experimental results show that the method can not only alleviate the vibration but also can preserve the initiative scan of the camera.Finally, several video stabilization experiments have been carried out with the video image sequence taken by hand-held camera. In addition, the stabilization effect is evaluated. The simulation and experimental rusults demonstrated that this method of EIS proposed in the thesis is feasible.