一种鲁棒性车道线识别算法

摘要

In order to meet the requirements of the real-time and robustness of lane mark identification algorithm, a novel lane mark identification algorithm was proposed in this paper. Turning the color image into grayscale and filtering out noise by median filter were introduced firstly. Then extracting the lane mark edge by steerable filter and contrasting it with the classical algorithm was introduced. Lastly, the identification algorithm based on Hough Transform was put forward and the validity of which was verified by experiments by using several videos which were collected from highway and common road. The results indicate that the proposed method can recognize lane mark quickly and accurately both on the conditions of highway and more complicated common road.%为了更好地满足车道线识别算法的实时性和鲁棒性要求,提出了一种新的、有效的车道线识别算法。首先,将图像灰度化后,采用中值滤波去除图像采集过程中引入的噪声。其次,应用方向可调滤波器进行边缘提取,并与经典的边缘提取算法进行对比分析。最后,应用霍夫（Hough）变换对车道线进行识别,并分别对在普通公路和高速公路上实地采集的视频图像进行试验测试。结果表明,无论是在高速公路工况还是在路面结构较复杂的普通公路工况,识别算法都能快速、准确地将其识别。