一种自适应的车道线检测算法

An Adaptive Lane Detection Algorithm

车道线检测是一种在基于视觉的驾驶员辅助系统中起关键作用的技术手段,可用于车辆导航,侧向控制,防撞或车道偏离警告等车辆系统中。提出一种自适应的车道线检测方法。首先,从原始图像中提取感兴趣区域(Region of Interest,ROI)图像并将其转换为灰度图像;然后,自适应的改变车道宽度识别出可能的车道区域;最后,根据车道的结构和方位特性匹配车道,再利用Hough变换检测直线,并结合前一帧与当前帧的车道信息预测车道段,以避免计算错误。实验数据表明,该方法在准确性和鲁棒性方面表现突出。

Lane line detection is a technology that plays a key role in vision-based driver assistance systems and can be used in vehicle systems such as vehicle navigation,lateral control,collision avoidance or lane departure warning. This paper presents an adaptive lane detection method. First,The region of interest( ROI) image is extracted from the original image and converted into a grayscale image. Then,adaptive lane width changes to identify possible lane areas. Finally,the lane is matched according to the structure and azimuth characteristics of the lane,and the straight line is detected using the Hough transform. In combination with the lane information of the previous frame and the current frame,the lane segment is predicted to avoid calculation errors. Experimental results show that the proposed method has the characteristics of high accuracy and good robustness.