Lane line detection method

Abstract

The invention discloses a lane line detection method, and belongs to the technical field of image processing. The lane line detection method comprises the steps that first a pre-processed image is horizontally divided into K edge image blocks, the height scale of the edge image block at the bottom end and the entire image is [1/4,1/3], and two or more pairs of straight lines which are closest to a vanishing line are determined under the determined vanishing line to be used as candidate line pairs of the current edge image block; then the weights of the candidate line pairs are calculated, the straight line with the maximum weight of the edge image blocks of the candidate line pairs is taken as the only lane line segment pair of the current edge image block based on the weights of the candidate line pairs; at last, the lane line of a current frame image is output based on the end points of the lane line segment pairs of all the edge image blocks. The method is used for detecting lane lines, the method is not sensitive to initialization parameters, the robustness of detection is high, and a good detecting effect can be achieved under the bad conditions of dark light, lane line loss, shadows and the like.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to the detection of lane lines. Background technique [0002] Every year, many people around the world are killed in car accidents due to vehicle deviation. In order to avoid such things from happening, people have proposed an intelligent assisted driving system, which can warn the driver to a certain extent and reduce such accidents. the effect that occurs. Lane line detection is an important part of the intelligent driving assistance system. A series of algorithms with good results have been proposed in this field, including template-based, texture-based, and region-based algorithms. [0003] Due to real-time constraints, even some well-known algorithms have high detection and tracking rates, but because they require a lot of processing time, it is difficult to apply them in practice, such as Yue Wang using B-Snake for lane line detection algorithm. Some al...

AI Technical Summary

Problems solved by technology

[0003] Due to real-time constraints, even some well-known algorithms have high detection and tracking rates, but because they require a lot of processing time, it is difficult to apply them in practice, such as Yue Wang using B-Snake for lane line detection algorithm

Some algorithms can achieve better detection results in the case of weak light and missing lane lines, but have a high error rate for curve detection, such as Massimo Bertozzi's algorithm for lane line detection using a stereo vision system

[0004] Some other algorithms, such as Canny / Hough Estimation of Vanishing Points (CHEVP), locate the lane line by determining the position of the vanishing point on the lane detection picture, but the robustness of determining the lane line through the vanishing point in this algorithm is very low. Weak, because the algorithm assumes that the car lines in each area intersect at the same vanishing point, but due to noise and other interference, the real lane lines in each area do not intersect at the same vanishing point, so As for its weak robustness, in addition, the algorithm cannot deal with the absence of lane lines, and is sensitive to initialization parameters, and its robustness is weak