Research On And Implementation Of Traffic Flow Statistics And Identification Method Based On Motion Analysis

Traffic flow statistics and vehicle identification are the basic step in intelligent Transportation System, or ITS. In fact, there are quite a few conventional methods used to detect traffic flow, such as electromagnetic induction coil, ultrasonic detector and microwave detector, but, as the video vehicle detection method based on image processing can detect a larger area and have a more flexible system, it has been more widely used over the recent years. In this paper, we have studied the method of vehicle detection, traffic statistical method and simple vehicle size and color recognition method based on the moving target detection. And a platform of experiments is built using VS and OpenCV. The details are as following:(1)Making an analysis of the methods relating to the moving target detection and obtaining the vehicle foreground area in the traffic video by way of the frame difference method and background subtraction method; comparing and analyzing the results of the moving targets obtained via the two detection methods by doing experiments.(2) Proposing the virtual loop method governed by the dual threshold method in view of the defects existing in the virtual loop traffic flow statistics governed by the single threshold method; obtaining the real-time number of vehicles based on the vehicle foreground image picked up via the moving target detection.(3) Classifying vehicle size and identifying colors; distinguishing vehicle with big or small according to their area and the height; the color of the vehicle is recognized based on the feature of HSV color histogram.(5) Implementing an experimental prototype system based on the above-mentioned research.