摘要
为解决汽车动态称重多车道车辆检测与分车问题,提出了一种基于二维激光雷达的多车道车辆检测与分车算法。基于所建立的平面直角坐标系,将二维激光雷达对车道扫描得到的极坐标值转换为直角坐标系下的坐标点,对所拾取数据点进行中值滤波处理后,采用车辆边缘信息点提取算法对所有点数据进行分类截取,用凝聚层次聚类算法和最小二乘法曲线拟合,提取出车辆的左右边缘信息。以车辆左右边缘为多车道分车标志特征,确定各车辆的位置信息,从而实现对多车道的横向分车。以车道有无车辆的特征标志位来实现车辆的前后分车。试验结果表明,该方法能实现对多车道车辆的检测和车辆分车,对今后的车辆检测和分车具有一定的参考。
In order to solve the problem of multi-lane vehicle detection and separation in vehicle dynamic weighing,a multi-lane vehicle detection and separation algorithm based on two-dimensional LiDAR is proposed.Based on the established planar rectangular coordinate system,the polar coordinate values obtained through two-dimensional laser radar scanning the lane are converted into coordinate points in the rectangular coordinate system.After performing median filtering on the picked data points,the vehicle edge information point extraction algorithm is used to classify and intercept all point data,and the agglomerative hierarchical clustering algorithm and least squares curve fitting are used to extract the vehicle information of the left and right edges.The information of left and right edges of the vehicle are used as the characteristics of the multi-lane car division signs to determine the position information of each vehicle,so as to realize the horizontal division of the multi-lane vehicles.The front and rear separation of vehicles is realized based on the characteristic signs of whether there are vehicles in the lane.The test results show that the method can realize multi-lane vehicle detection and vehicle separation,and to some extent,the results of study have certain reference significance for future vehicle detection and separation.
作者
陈建
于殿泓
惠鑫鑫
辛慧
马歆
CHEN Jian;YU Dianhong;HUI Xinxin;XIN Hui;MA Xin(College of Mechanical and Precision Instrument Engineering,Xi’an University of Technology,Xi’an Shaanxi 710048,China;Xi’an Sanshe Electronic Information Technology Co.,Ltd.,Xi’an Shaanxi 710000 China;Xi’an Daoheng Traffic Equipment Technology Co.,Ltd,Xi’an Shaanxi 710000 China)
出处
《传感技术学报》
CAS
CSCD
北大核心
2022年第11期1513-1520,共8页
Chinese Journal of Sensors and Actuators
基金
国家自然科学基金项目(62075180)。
关键词
车辆检测与分车
二维激光雷达
中值滤波
边缘信息提取
凝聚层次聚类算法
最小二乘法
vehicle detection and separation
2D LiDAR
median filtering
edge information extraction
agglomerative hierarchical clustering algorithm
ordinary least squares
