Dissertation
Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

The Occlusion Remove Research in Video Intelligent Traffic System

Author GuoLian
Tutor LiuFuQiang;SongChunLin
School Tongji University
Course Signal and Information Processing
Keywords Demarcate Kalman forecast Plaque scoring Camera collaboration
CLC TP391.41
Type Master's thesis
Year 2007
Downloads 126
Quotes 3
Download Dissertation

With the ever-accelerating process of urbanization , traffic congestion has become a serious problem for the constraints of urban development and improved living standards . Under the existing conditions , the use of the ITS system to enhance traffic control and management is the best solution . Are many ways to get traffic information , including video ITS system the most development potential . Target of frequent occurrence in the transport process block seriously affect the accuracy of the video surveillance system , to solve this problem , this paper first systematic analysis of the type of shelter and causes , and design appropriate solutions for different control environment and occlusion , the main research are as follows : Firstly, the camera calibration algorithm conducted in-depth research , analysis of the principles of perspective transformation to derive the parameters of a camera has nothing to do with the two - dimensional camera calibration method , to overcome the original algorithm on the camera parameters over-reliance on the shortcomings , and to test the accuracy and robustness of this new calibration method . Secondly, in the tracking system using a single camera , the traditional tracking algorithm based on the plaque , the respective characteristics based on the color histogram mean-shift tracking algorithm and Kalman prediction algorithm . For plaque tracking operation speed, but by the influence of occlusion ; Mean - shift combined with Kalman forecast able to go to cover the problem of slow computing speed , the proposed plaque scoring system and the value of life parameters based target tracking and occlusion different . This program is to ensure the real-time nature of the system , but also can eliminate the impact of the occlusion . Finally , for multi-camera tracking system , based on the traditional scheme of the three-dimensional reconstruction can not meet the real-time monitoring of a logic-based analysis of two-dimensional multi - project monitoring program , access to different viewpoints of the same target location information to be a comprehensive analysis and judgment, Quweicunzhen to remove the cover , so that the system computing speed is greatly increased , to meet the requirements of real-time monitoring . The end of the article a brief discussion of work problems and further research directions .

Related Dissertations
More Dissertations