| In road traffic management system, the basic task is for a variety of vehicles on roads for testing in order to achieve the traffic information collection, analysis for future transportation system to pave the way. In the early,we often induction coils and other hardware laying on the road to detect vehicle flow. Today, we employ the CCD video to replace the earlier detection methods.The main work is in the low-speed vehicle, to analyze and solve the two main problem that affect vehicle detection and counting: the vehicle’s own shadow, blocking each other or the vehicle is of the background object occlusion between vehicles. CCD camera on the subject to get the traffic moving vehicle in the video for the study object, using background subtraction method, the virtual detection coil and other algorithms to build a real-time traffic flow detection system based on DSP. By process control, the camera to obtain traffic information in real-time processing, access to traffic information. Systems with TMS320DM643 high performance DSP core, capture video signals through the CCD camera, the captured video signal through TVP5150 video decoder chip, then transfer into the TMS320DM642 video image processing, after sending encoded video encoder chip SAA7105, finally, the number of vehicles to calculate and display the target vehicle.In the CCS software development platform, we capture video of the vehicle, display and counting through the Program debugging, and the use of C language designs image preprocessing, shadow detection, vehicle detection and counting algorithm program.Introjects kinds of vehicle detection and counting algorithm in the paper,finally we successfully implements these algorithms in DM643. In a static scene we can quickly and accurately detect the vehicle and counted. vehicle flow detection system based on CCD image acquisition and processing has easy installation and maintenance, real-time detection of high efficiency, wide detection area characteristics and so on. |
PDF Full Text Request