Fast Classification Of Vehicles Based-on The Moving Region

With the development of our economy, motor vehicles and traffic accidents are more and more, and the vehicle detection, location, classification, become more and more important. Information Transportation System (ITS) has a broad application future. The Fast Classification of Vehicles is a hotspot in ITS and image manipulation.By analyzing the usual algorithms for classification of vehicles, a new fast classification approach for vehicles based on the moving region is presented in this paper, also, this paper studies the existing digital image processing technology and algorithms, and applies them to car model recognition system to suppress all kinds of noise, then reviewes the main moving object detection algorithms and their advantages and disadvantages. Further more this paper developes a new algorithm for the detection of high-speed moving objects based on picture blocks and HSV color space of image differences techniques. Five kinds of cars'relative moments are extracted from the car area surrounded by car outline, ratio between car length and car height, ratio between car top length and car bottom length.A high performance is achieved to identify the car at different positions in the picture because of the utilization of relative moments features. The car model identification method this paper presented, based on constructing membership functions of fuzzy theory. Mainly proposes a method of pattern recognition on vehicle by extracting characteristic data. This method can recognize normal car, jeep, truck and bus, then give the program black diagram of realizing the process of recognition. This method has the characters of better recognition efficiency and faster recognition.,which is easy to realize, and has capacious future.