基于HDL-64E激光雷达道路边界实时检测算法  

An approach of real-time road boundary detection based on HDL-64E lidar

作　　者：王俊 孔斌[1,2] 王灿 杨静 WANG Jun;KONG Bin;WANG Can;YANG Jing(Institute of Intelligent Machines,Chinese Academy of Sciences,Hefei 230031,China;School of Information Science and Technology,University of Science and Technology of China,Hefei 230026,China)

机构地区：[1]中国科学院合肥智能机械研究所,安徽合肥230031 [2]中国科学技术大学信息科学技术学院,安徽合肥230026

出　　处：《合肥工业大学学报（自然科学版）》2018年第8期1029-1034,共6页Journal of Hefei University of Technology：Natural Science

基　　金：国家自然科学基金重大研究计划集成资助项目(91320301);国家自然科学基金资助项目(61304122);中国科学院战略性先导科技专项资助项目(XDA0804109)

摘　　要：为了使智能车辆在多种道路环境中能够快速有效地提取前方道路,文章提出一种基于HDL-64E激光雷达的道路边界检测算法。该算法首先对激光雷达数据进行空间邻域分析,获取平滑度特征图像;然后利用自适应方向边界搜索算法获取候选道路边界激光雷达数据;为了解决激光雷达数据中存在的干扰及不连续问题,对候选道路边界激光雷达数据进行聚类分析及曲线拟合。实验结果表明,在高速、城区以及乡村道路环境下,该算法能够实时、准确地提取道路边界信息,满足智能车辆道路环境建模及路径规划的需要。In order to enable intelligent vehicle to extract the front road in a variety of road environment effectively,an approach of road boundary detection based on HDL-64E lidar is presented.Firstly,the image of smoothness characteristics change is gotten by analyzing the spatial neighborhood of lidar data.Then the candidate lidar data of road boundary is gotten by using adaptive directional circular search.Moreover,to solve the interference and discontinuous problem of lidar data,the algorithms of cluster analysis and curve fitting are presented to process the candidate lidar data of road boundary.Experimental results demonstrate that the approach can accurately extract road boundary information in urban and rural road and highway environment,and meet the needs of road environment modeling and intelligent vehicle path planning.

关 键 词：HDL-64E激光雷达 平滑度特征 自适应方向边界搜索 聚类分析 曲线拟合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]