Design Of Traffic Detection And Overspeed Vehicle Capture System

This paper discusses the vision-based traffic detection system, which plays an important part in the Intelligent Transportation System. By extracting real-time traffic parameters and making traffic surveillance, the vision-based traffic detection system can optimize the traffic control effectively. With the help of the traffic image processing, we can obtain various parameters under complex traffic conditions, which can't be realized with traditional detection methods. Application of this technology will have great influence on traffic management and alleviation of traffic congestion.First, the disadvantages of traditional detetors are presented respectively. Then, it is pointed out that the vision-based traffic detection system will be the most efficiency detection system and several vision-based traffic detectors are introduced. Aimed at the bad performances on reliability and rapidity of the image processing systems based on common computers, an embedded vehicle detection system based on DSP and FPGA is presented and all the details about the hardware design are discussed later according to their functions, including video input processor module, FPGA module, DSP module and power module. The principles we should adhere to in the PCB design process is summed up also. Then, the module configuration works of DSP and logic design works for image storing in FPGA are illustrated. By extracting background image with Surendra algorithm and thresholding the image with Otsu algorithm, we obtain the vehicle speed with two CCD camera using background subtraction. Further more, the obtaining of some other traffic parameters is discussed. Some problems during the debug work and the settlement to them are also explained and an example is illustrated to show how to get the speed of vehicles. At last, the main work of the paper is summarized and suggestions are given for future study.