Research On Lane Line Detection Method Based On Structured Road

With the increasing number of vehicles,traffic safety is becoming more and more serious.In order to reduce the occurrence of traffic accidents,it is imperative to develop a set of perfect auxiliary driving system.It can help the driver to drive safely,so that the driver can have time to brake and slow down before the possible accident to avoid or mitigate the harm caused by the accident.Lane detection is the premise of safe driving assistance system,so lane detection research is very valuable and significant.This paper will conduct an in-depth exploration of lane detection and recognition of structured roads.The main research contents are as follows:1.When extracting lane line information,separate the two colors of lane lines,and finally combine the extracted results to get the complete lane line information.The specific extraction method is to extract the yellow lane lines by converting the image into HSL image and detecting the In Range function of OpenCv library;The white lane line is obtained by image processing of color threshold using the L channel of HLS.The extracted yellow and white lane line images were combined to get the final image.The experiment shows that this method has a good effect.2.Canny edge detection is the commonly used method to extract lane lines by using edge features.The extraction effect of this method is better than most other methods.So this paper is the use of bilateral filtering Canny edge detection algorithm to extract the lane line.3.The lane lines were fitted by sliding window and quadratic polynomial,and the position of the sliding window was determined by the histogram of the lane line image extracted based on color threshold.The lane line images extracted by edge detection were detected and fitted out by sliding window detection.Finally,Kalman filter algorithm is used to track the detected lanes and carry out deviation detection.The experimental results show that the proposed algorithm has a good effect on structured road lane detection and recognition.