Research On Real-time Video Capture System Based On The DM642

Intelligent Transportation System has become a feasible way to solve urban traffic congestion based on its real-time traffic information. With the rapid development of multimedia technology, video information attracts the people’s attention and video detection method has become more and more important. This dissertation researches on the application of the video detection technology in the intelligent transportation system.This paper provides a DSP-based real-time video acquisition system, in which a camera is installed to the traffic light bracket to collect the traffic video information from four directions in real-time. The traffic video information can be used to restore the traffic. Moreover, traffic information is obtained by image processing of video information.The presented system is designed based on TMS320DMB642supported by T1Company. Firstly，it makes a comparison among several main information acquisition technologies, and select video acquisition method. Then, analyses functional requirements to achieve the overall design scheme of the system. It comprises DM642core module, the codec module, storage module, external communication module and power reset module and so on, and the design diagram is made according to the characters of the main devices of those modules. Finally, a DSP/BIOS-based driver is designed according to class/mini drive model, and it provides the interface up to the application layer and down makes the operation to the underlying hardware. This paper shows the design principles of class driver and micro driver and also some functions’program codes.The presented system can achieve the performances of video capture and display after debugging on CCS2.2, and basically it meets the intended function of the system through the real-time analysis and CPU resource consumption. The advantage of the presented system is that it has a wide detection area and convenient installation and maintenance.