同济大学学报（自然科学版）2017，Vol.45Issue(7)：1069-1074,1090,7.DOI:10.11908/j.issn.0253-374x.2017.07.019
基于断面激光扫描的移动车辆交通流参数提取
Extraction of Mobile Vehicle Traffic Flow Parameters From Sectional Laser Scanning Data
摘要
Abstract
The relationships between vehicles,such as position,velocity and distance,are the main cantonments of micro-traffic flow parameters.These parameters are important to unmanned driving,intelligent traffic,etc.A novel method was proposed for micro-traffic flow extraction from mobile laser scanning data.Based on the mobile sectional laser scanning data,a threshold was selected to segment and classify the original point cloud into different vehicles.Then,the quadratic polynomial weighting method was used to extract the feature point from vehicle's point cloud.The distance and velocity parameters were then computed from adjacent vehicles or adjacent sections.Finally,an experiment was conducted in Shanghai Yah'an elevated road to verify the traffic flow parameter extraction method from mobile laser scanning data.The results show that such parameters could be easily and accurately calculated.The average distance error of directly front or behind car is about 0.058 m and its average velocity error is about 1.62 km · h-1.The average distance error of sideward car is merely 0.100 m,and its average velocity error is about 1.29 km · h-1.关键词
车辆识别/激光扫描/点云/交通流参数/特征点提取Key words
vehicle identification/laser scanning/point cloud/traffic flow parameters/feature point extraction分类
天文与地球科学引用本文复制引用
吴杭彬,刘豆,刘启远,孙剑,姚连璧,汪志飞,刘春,吴声援..基于断面激光扫描的移动车辆交通流参数提取[J].同济大学学报（自然科学版）,2017,45(7):1069-1074,1090,7.基金项目
国家自然科学基金(41671451) （41671451）
国家重点研发计划(2016YFB0502104,2016YFB050210,2016YFB1200602-02) （2016YFB0502104,2016YFB050210,2016YFB1200602-02）
中央高校基本科研业务费专项资金 （）
