In order to solve the problems in the traditional Hough transform algorithm for UGV structured road lane marking identification, such as the inaccurate and incomplete extraction of the middle of the lane marking and the loss of useful lane information, a Hough transform lane marking identification algorithm based on nearby region of interest(ROI) is proposed. Delimiting RIO nearby the lane marks was extracted by the traditional Hough transform from the part of binary preprocessing image, finding out the relation of the auxiliary degree between the pixel in the ROI and the lane marking. The maximum of the accumulative unit is calculated through this auxiliary degree, with the parameters of lane marks obtaired. The new algorithm was tested by using the real-time captured images. Experimental result shows that 80. 42 ms and 92. 22 ms on average are requined respectively by the algorithm to process one image in the case of straight road and bending road. The algorithm is of good real time and the algorithm can extract structured road lane marking effectively.%针对传统Hough变换算法在无人车结构化道路车道标线识别中出现的中间间断车道标线提取不精准、不完整、有用车道标线信息丢失等问题,提出基于感兴趣附近区域的Hough变换车道标线识别算法.运用传统Hough变换对预处理得到的二值图像进行图像部分分块的车道标线提取,在已提取的车道标线周围划定感兴趣附近区域,找出感兴趣附近区域内的像素与车道标线的辅助度关系,通过辅助度计算累计单元的最大值,得到车道标线的参数.对无人车实时采集的图片进行实验,实验结果表明,该算法在直道、弯道情况下平均处理一帧图片耗时分别为80.42 ms、92.22 ms,具有很好的实时性,且该算法可以有效提取结构化道路的车道标线.
展开▼
