Research On Algorithms Of Vision-based Lane Detection On The Highway

The rapid development of society and economy has brought great changes to people's travel tools, and people's demand for cars is also growing. With the popularity of cars, traffic accidents happen frequently, and that is the reason why the Intelligent Transportation System comes out. The research on the road detection and identification in key technologies of Intelligent Transportation Systems plays a vital role in ensuring the safety of cars and fulfilling the unmanned vehicles.Lane Departure Warning technology is one of the research directions on the vehicle safety driver assistant systems, lane detection and recognition is the key part of it. This thesis takes deeply study of monocular vision-based lane-line detection and the algorithms of road trend toward judgment.This thesis takes the highway lane-ray images as the research objects. First, the thesis preprocesses this image based on the analysis of common pretreatment, and then analyzes the common method of detecting lane, it uses the principle of Hough transform to detect and identify the lane marking lines. The thesis bring forth the least square method to fit the slope k of the equation of the lane, through the analysis of images from the time vehicles are on the road, to judge the algorithm of path direction under the connection of slope. Based on the analysis of common model in the current lane-departure warning system, the thesis establishes the model of camera-based imaging model for lane-departure warning.Using this algorithm, the author analyzes the lane images with three kinds of typical lane, such as the straight, left turn, right turn. The results show that the algorithms can fulfill a good detection of lane and road toward judgments.