A multi-lane line detection method

Abstract

The invention discloses a method for detecting multi-lane lines. Firstly, color images of consecutive frames are acquired, grayscale processing is performed on the current frame image, and the grayscale image is subjected to inverse perspective transformation according to camera parameters and the set grid map size. Obtain the grid map of the road part; then according to the prior information of the control points, use the Thiessen polygon to divide the grid map into regions, and perform image binarization on the divided regions; then group the regions according to the prior information of the lane lines, And record the pixel point coordinates of non-zero pixel values ​​respectively. Then use the improved RANSAC algorithm to do curve fitting on the grouped point set to obtain the lane line equation, and finally combine the lane line equation and the particle filter algorithm to predict the current image control points, calculate the final lane line equation and make a perspective on the lane line equation Transform to obtain the lane line equation in the original image. The method proposed by the invention has high precision and good robustness for lane line detection, and can simultaneously complete multi-lane line detection.

Description

technical field [0001] The invention belongs to the field of computer vision and automatic driving, and in particular relates to a multi-lane line detection method. Background technique [0002] In the field of automatic driving, the detection of lane lines is an important link, and it is also a hot issue in the field of automatic driving. However, most of the current lane line detection methods based on computer vision cannot robustly detect lane lines. The main reasons are as follows: illumination, shadows on the road, other obstacles on the road (vehicles, pedestrians, etc.), Other information such as traffic markings will interfere with the detection of lane lines on the road. [0003] At present, there are many multi-lane line detection algorithms based on computer vision. Many journals and conferences related to autonomous driving at home and abroad have listed the lane detection algorithm as a key research field. Scholars at home and abroad have also made a lot of b...

AI Technical Summary

Problems solved by technology

[0006] However, none of the above-mentioned documents can robustly detect multi-lane lines. The main problems are as follows: first, the entire image is processed uniformly in the image preprocessing stage, which greatly affects the interference information such as illumination and shadow; second Second, if the RANSAC algorithm is used to directly fit the point set to the lane line equation, the strip feature of the lane line in the lane line detection task cannot be fully utilized, so that the RANSAC algorithm cannot obtain better results under limited calculations.

Images