The Research Of Moving Objects Detection And Parameter Estimation

With the flying economic development, the living standard has been enhanced greatly. There is growing possession of automobiles in each country, so modern management methods are needed to administer traffic, which begets the research on Intelligent Transportation System (ITS).ITS mainly computer vision and image processing technology. It can detect, identify and track moving vehicles by analyzing video dynamic sequence, and make with motion estimation for moving vehicles. Thereby, it offer a more advanced and feasible scheme to deal with traffic accidents.The paper firstly introduces some domestic and overseas' general development situations of ITS, some basic theories of image processing and the application of image processing in ITS. Then aimed at the problems that how to detect, identify and track moving vehicles of ITS based on video, a series of exploratory researches done in the paper are as follows:(1)It simply introduces all kinds of ways of moving objects detection, detailedly dissertates the problem that how to attain and update background models in complex condition, and proposes a new method to acquire background. It also introduces the algorithm based background subtraction and the algorithm based image difference, the other algorithm based image symmetrical difference is also presented.(2) Aimed at the problems of moving objects segmentation, this paper introduces some used segmentation methods, and effectively segments moving vehicles from detection image. Then targeted at the segmentation image, to pick up the moving vehicles' characteristic, and process vehicles recognition and vehicles model classification according as characteristic parameters.(3) Based moving objects detection and segmentation, the detection algorithm is applied in the vehicle tracking. Considering moving distance and area change of one vehicles in the two consecutive frames is small, the paper presented an algorithm tracking moving vehicles. Two parameters, centroid and tracking window area of moving vehicles, are used to match in this method.The paper provides many ways to resolve the difficult problems of video vehicles detection, identity and tracking. Experiment shows that, they have important theoretic meaning and application value as universal methods in the research area of computer video, which can be easily generalized to the applications with different situations, and have a wide application prospect as well.