The Research On Detection Of Speedway Lane Based On Machine Vision

At present, the most studies of lane detection algorithm based on machine vision concentrate on the straight road detection of non-structure path or structure path, but the most accident occer on the highway curve road. So this article puts forward a lane detection algorithm based on the picture restoring, it is suitable to both the highway straight road and the highway curve road.Because the clarity is different between pictures obtained by CCD camera during the day drives and night drives, this paper design day lane detection pattern and night lane detection pattern. In order to enhance the system's accuracy, picture pretreatment which can eliminate picture's irrelevant information and restore the useful real information must be taken before processing picture, thus can strengthen the related information and simplify the data greatly.According to the fact that on heighway the traffic lane's profile is more ruler and the fact that the path region and the non- path region has obvious edge image, this research advance that before traffic lane distinguishion and detection the cutting apart of the goal and background in the pretreated picture must be taken, by doing this we have got the clearer lane border.Road border in the picture after cutting apart is a rectangle sealing area with certain width and certain trend, according to this important characteristic this papers has designed a lane detection method based on the picture restoring, and has obtained the road picture which only contained the traffic lane line. It indicates that does a good job of the adequate preparation in description of the line for the next step.In order to remedy the greater deficiency and error of the most lane detection systems in which the traffic lane line is fitted directly when do not know the form of the lane line, at last this papers put forward the algorithm that judge the trend of the lane first, and then chooses the optimum method of fitting and fits to the point on the lane , in the end working out the equation of the lane line.