基于视觉的车道线检测技术研究

Research on Vision-based Lane Line Detection Technology

环境感知技术是汽车自动驾驶的基础,车道线检测技术是环境感知的关键部分。本文提出一种基于视觉的车道线检测技术。首先,运用"张正友标定法"对单目相机参数标定;然后将图像转换成鸟瞰图,并使用LaneRidge检测器对车道线进行分离;其次,使用随机抽样一致算法RANSAC拟合车道线,并对虚实车道线进行分类;最后,在MatLab平台上对本文提出的车道线检测算法进行验证。实验结果表明,本文提出的车道线检测算法能够准确拟合车道线,并对虚实车道线进行分类。本研究能够确保车道线检测的实时性和准确性,从而为智能交通系统的应用打下基础,提高智能汽车行驶的安全性。

Environmental perception technology is the basis of automatic driving,and lane detection technology is the key part of environmental perception.Therefore,this paper proposes a research on vision-based lane detection technology.Firstly,"Zhang Zhengyou calibration method"is used to calibrate the parameters of monocular camera;then the image is converted into an aerial view,and lane ridge detector is used to separate lane lines.Random sampling consensus algorithm(RANSAC)is used to fit lane lines and classify virtual and real lane lines.Finally,the lane detection algorithm proposed in this paper is verified on MATLAB platform.The experimental results show that the lane detection algorithm can accurately fit the lane line and classify the virtual and real lane lines.This study can ensure the real-time and accuracy of lane detection so as to lay the foundation for the application of intelligent transportation system and improve the safety of intelligent vehicles.