基于车载4线激光雷达的前方道路可行驶区域检测

发布时间：2018-01-21 15:29

  本文关键词： 线激光雷达 DBSCAN算法 加权欧式距离 可行驶区域　出处：《汽车技术》2016年02期 　论文类型：期刊论文

【摘要】：通过分析大量激光雷达数据,提出了改进的DBSCAN算法对雷达数据进行聚类,以使智能汽车获得前方道路和障碍物信息。根据不同的密度参数,多次调用该算法完成多密度聚类。结合提出的道路边沿、路面和障碍物等信息提取方法,在结构化或半结构化的城市道路中对提出的方法进行实车试验,结果表明该算法可实时准确的提取智能汽车的可行驶区域信息。
[Abstract]:Through the analysis of a large number of laser radar data, an improved DBSCAN algorithm is proposed to cluster the radar data, in order to make the smart car road ahead and obstacle information. According to the density of different parameters, calls the completion of the algorithm of multi density clustering. In combination with the proposed edge extraction method, pavement and obstacle information, vehicle test the proposed method is in a structured or semi-structured road in the city. The results show that the algorithm can real-time and accurate extraction of the intelligent vehicle driving region information.

【作者单位】： 北京工业大学北京市交通工程重点实验室;
【基金】：北京市属高等学校人才强教计划资助项目(038000543115025)
【分类号】：U463.6
【正文快照】： Preceding Drivable Area Detection Based on Four-layer Laser Radar(Beijing Key Laboratory of Traffic Engineering,Beijing University of Technology,Beijing 100124)1前言智能汽车利用可通行区域信息来保障其在道路区域内安全行驶,防止侧翻和碰撞等危险状况发生。车辆，

本文编号：1451850