一种用于激光雷达识别车道标线算法

发布时间：2018-03-05 08:40

本文选题：智能车　切入点：激光雷达　出处：《科学技术与工程》2017年16期 　论文类型：期刊论文

【摘要】：利用点云数据空间分布特征和回波强度信息,结合局部均值变点统计方法,提出了一种用于激光雷达数据帧的车道标线识别算法。该算法首先基于车载激光雷达采集的道路周围环境点云数据中高程信息进行滤波,确定可行驶区域。然后利用局部均值变点统计对可行驶区域点云数据中的回波强度值进行标记提取,即车道标线点云数据粗提取。最后基于EM(最大期望)方法聚类,从而完整、准确地识别车道标线。实验结果表明,该算法不仅能够准确定位可行驶区域,进而可以实现车道标线的自动提取;而且有效抑制了道路周围环境对车道标线识别的干扰,验证了算法的有效性。
[Abstract]:Using the spatial distribution characteristics of point cloud data and echo intensity information, and combining with the statistical method of local mean change point, This paper presents a lane marking recognition algorithm for lidar data frames. Firstly, the algorithm is based on the elevation information of the point cloud data of the road surrounding the road collected by vehicular lidar. Then using the local mean change point statistics to extract the echo intensity value of the point cloud data in the driving area, that is, the rough extraction of lane marking point cloud data. Finally, based on the EM-based method, the maximum expectation method is used to cluster the data. The experimental results show that the algorithm not only can locate the driving area accurately, but also can automatically extract the lane marking, and effectively restrain the interference of the road surrounding environment to the lane marking recognition. The validity of the algorithm is verified.
【作者单位】：山东理工大学交通与车辆工程学院智能交通研究所;清华大学汽车安全与节能国家重点实验室;
【基金】：国家自然科学基金(61074140,61573009,51508315) 汽车安全与节能国家重点实验室开放基金(KF16232) 山东省自然科学基金(ZR2014FM027) 山东省社会科学规划研究项目(14CGLJ27) 山东省高等学校科技计划(J15LB07)资助
【分类号】：TN958.98

【相似文献】