Design And Implementation Of Intelligent Vehicle Vision System Based On Embeded Linux

Now,the researches on visual navigation system have been developed in many countries. The lane-detection system is an important component of many visual navigation systems. They have long been receiving people's attention, because the precision of lane-detection closely relates to the driving safety of intelligent vehicles.Aiming at the safety problem of vehicle driving, this thesis researches a vision system of intelligent vehicle based on ARM9 and Embedded Linux operation system. The overall framework of system is designed in this paper, the hardware has adopted S3C2440A microprocessor as the hardware which has the USB port connected with front end camera, so that it can achieve monitor image acquisition. This thesis adopts Linux operation system for the embedded processor architecture as software platform which is open source and scalable. Software design is completed which includes Linux kernel and file system transplant, camera driver development, as well as the realization of monitoring application software. The paper completes the design of video surveillance software which is based on v4l and motion detection. Using this system can realize the acquisition and storage of motion video and still images, the USB interface reserved can export image conveniently.In this thesis,the road positioning algorithms based on image are researched to process the images. In order to get a greylevel image from a colour one, we introduce a method called getting from multicolor channel to stronger the white line on the road image. Then given the features of road line, a difference cyclostyle is defined to extract the edge. At the same time, in order to improve the real-time performance of roads and anti-jamming capability, regional growth way is introduced, through it we can chose a proper seed to get a regional road image. The detection method of combining the regional growth way and edge extraction is stated to identify the road edge. Experiments show that the algorithm can detectthe margin of road to some extent.