2010 2nd International Conference on Industrial and Information Systems
Ii B Ackground Change,Shu Huazhonga,Minab Huang,Chen Leic
Abstract:A stllt/y ojlll OPing o!J/ecl detectioll lHIsed Oil collllJ ining IHIclipollll ll pl'Ofi/e tlifference 1IIg 0ritll lll Abstrac�Reeen t yean, moviDg object deteeti on of video freqnency sequence image is more and more important unseeing In Intellig ent monitory, video compressi on, automatic navigation, man-maehlne Interaetlon, virtnal reality and so on. In this paper , we use a method of moviDg object deteetl on based on combining background diff erence algorithm with edge-texture and object match to get the result. In the method, at fInt updated image background, then detect image edges, extract the objects edges and matches the target, In the end locate and identify the target. From experimental analysis, the method can reduce the impact of noise In some extent and make the final test results more aeeura te. In recent years, the moving object detection of the video sequence image is used more and more widely, in many areas such as in intelligent control, video compression, automatic navigation, human-computer interaction, virtual reality applications, and making it an important research field of computer vision content. It only get more true, complete prospect to be accurate, reliable and efficient tracking. Currently, moving object detection methods include: optical flow method [2], frame difference method [1] and the background difference [3,7]. Optical flow method has the advantage which can detect the alone movement object, but most optical flow method is complex and time-consuming, without the special hardware is difficult to achieve real-time detection. Therefore, it use the difference method and background subtraction in occasions for high real-time demanding. Frame difference method is less sensitive to change in scene lighting, the shadow effect by the target is small, and very suitable for dynamic environment, but it can not completely separate object, which is not conducive to further analysis and identification. Compared with the frame difference, background difference method is generally able to provide a more complete feature data and get more accurate target image, but with the lapse of time , the background im�ges will be found many false movement target point, which affect effection of target detection., for sensibility about scene change because of external conditions and the light. For these reasons, we propose a method of moving object detection based on combining background profile difference algorithm. A Background model establishment Accurate target detection need that the background should have accurate, real-time updates. Background updating method is commonly used [1]: multi-frame …
Engineering,Computer Science