Researh On Motion Detection And Object Tracking Methods Of Image Sequence

With the world's sustaining care on security, Video Surveillance System is being a more and more important role in every layer of society as the safe guarding method. Three developed periods of Video Surveillance System are composed of analog video recorder, partially digital video recorder and fully digital video recorder. But they are all depending on human's judgment. The Intelligent Video Surveillance can discover the abnormity by processing and analyzing the image sequence, send the alarm and useful information without human action. Many countries make the Intelligent Video Surveillance as the primary developing direction of the safe guarding technologies. Compare to that, our country is on the underway stage. In order to improve the level of Intelligent Video Surveillance technology, the great subjects of Ministry of Public Security afford great fund support for developing Intelligent Video Surveillance technology.Indoor scene video surveillance is taken as the study object of the dissertation. The dissertation is mainly composed of four parts: image sequence capture, motion detection, object tracking, the design and implement of system software. Firstly, the image sequence capture technology on Windows Operating System is introduced, and then a comparative analysis between VFW and DirectShow is given. Secondly, motion detection method is improved by using the adaptive background update and adaptive thresholding segment methods after analyzing of traditional methods and can be adapt to the slow alterant scene. To meet the farther motion detection need, motion detection is done at the given region. Obtaining the motion object's position and orientation by motion segment and measuring the global orientation of motion history images is used to judge whether the motion objects enter the guarding region. Thirdly, the continuously adaptive Mean Shift based on hue is used for robust object tracking by processing color probability distribution images in dynamic change, which are coming from the image sequence. At last, the software design and implement have been introduced and the experimental results and analysis are given.