一种基于Edline线特征的车道线识别算法

为了提升车道线检测的准确性和实时性,改良车道偏离预警系统的性能,提出了一种新的车道线识别算法。首先应用投影法对采集到的图像设立感兴趣区域,以此来减少图像中存在的干扰信息;其次应用一种改进后的自适应高斯滤波算法对采集所得图像进行平滑处理,减少图像中不必要的细节;最后采用边缘绘制算法进行边缘检测,在此基础上,提出一种线段检测算法——Edline算法提取边缘线,对检测到的直线段进行筛选和聚类。利用引用计数法对车道线进行跟踪和预测。结果表明,新算法的平均处理时间为17.1 ms,准确率为96.19%,将其应用在车道偏离预警系统中可以有效地提高预警效率,提升预警的准确性和响应速度。研究结果丰富了车道线识别理论,可为车道偏离预警系统的应用及基础研究提供参考。

A lane line identification algorithm based on Edline line features

In order to improve the accuracy and real-time of lane detection and the performance of lane departure warning system, a new lane recognition algorithm was proposed in this paper. Firstly, the projection method was used to set up the region of interest for the collected images, so as to reduce the interference information in them. Then an improved adaptive Gauss filtering algorithm was applied to smooth the collected images and reduce unnecessary details. Finally, edge drawing algorithm was used for edge detection. On this basis, a line segment detection algorithm was proposed: Edline, which extracted the edge line, then filtered and clustered the detected line segments. The lane line was tracked and predicted by using the reference counting method. Experimental results show that the accuracy of the new proposed algorithm reaches 96.19% and the average processing time is 17.1 ms. The efficiency, accuracy and speed of lane recognition have been greatly improved by applying this algorithm to lane departure warning system. The research results enrich the theory of lane line recognition, and provide reference for the application and basic research of lane departure warning system.