中文摘要：

提出一种基于车载激光移动测量数据的微观交通流参数提取方法．将移动断面激光扫描数据根据一定的阈值进行分割分类，获得路域范围的移动车辆．根据二次多项式加权拟合法，获得移动车辆的特征点信息．再基于特征点的时空关系，计算车辆之间的距离、位置、速度等微观交通流参数．以上海延安高架路为例进行试验，对车辆之间的交通流参数提取进行实际验证．结果表明，该方法可以有效地提取车辆交通流参数，前、后方车辆的距离平均误差仅为0．06m，速度平均误差为1．62km·h-1；侧方车辆的距离平均误差仅为0．10m，速度平均误差为1．29km·h-1．

英文摘要：

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 Yan＇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.