A Research Based On DSP Of Lane Deviate Identification Method About Fatigue Driving Warning System

The Chinese road traffic safety situation is very severe, and ensure the road traffic safety is a completely difficult task recently. There are so much Influence factors of road traffic safety. By means of science and technology to improve road traffic safety has become a government, engineering and other social circles's consensus besides to improving the management and perfecting relevant laws, regulations, strengthen the consciousness of law. Image acquisition and processing parts that based on visual system of the car lane departure system is very important on resolve the safety problem of vehicle.After refer to knowledge of different fields of cross discipline, such as transport security, traffic psychology, auto theory, we establish a fatigue driving behavior model. That laid a clear direction and improved the reliability of the method for the subject of research. Then we consult the applying of Parallel computer, DSP, and FPG A in the image Real-time processing. Analysising the advantage of DSP in the image Real-time processing. Researching the flexibility and favourable collect nature on the designing. Depplying study the design glow and development environment of DSP. The data of dispose image is very big, we determine using TMS320DM6446 as the core of the system, which procreated by TI. The system has four main department—image gether part, image storage part, image dispose part and real-time disply part. Image gether part use CCD, video input interface use TVP5150PBS, and video output use TMS320DM6446's 4 ways 10 emic DAC, which for the CVBS and VGA output. TMS320DM6446 collocates TVP5150PBS.This paper uses CCD camera to gather roads information, after image is decoded by SEED—DVS6446 development platform's video processing subsystem, we use this system for the median filter. Then, the image edge detection is effected and identify the lanes at the edge of the line.At last, basing on the dribering behavioral model image dispose arithmetic and the nature of DSP, we raise a arithmetic which base on BIOS. As the processing of continous theory demonstration and reality testing, we obtain a image edge detection arithmetic for this task which basing on BIOS. And provides a basic method for the whole task.