基于路侧激光雷达的背景提取和地面分割方法  

Background Extraction and Ground Segmentation Method Based on Roadside LiDAR

作　　者：晋瑞河 任明武[1] JIN Ruihe;REN Mingwu(School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094)

机构地区：[1]南京理工大学计算机科学与工程学院,南京210094

出　　处：《计算机与数字工程》2023年第9期2033-2037,2054,共6页Computer & Digital Engineering

摘　　要：依据激光雷达工作原理,论文提出了一种基于路侧场景下的背景提取和地面点云分割方法。提出基于连续帧的背景提取方法,该方法以线号、角度值作为点云数据区域划分的依据,并通过空间密度聚类方法提取背景数据;提出基于射线特征的两阶段地面点云分割方法,对背景点云进行分割划分出地面点集以及非关键障碍物点集。在处理实时数据时,可以快速区分出关键障碍物、地面点以及非关键障碍物。最后,使用实际场景的路侧激光雷达数据进行方法性能测试。实验结果表明,论文提出的基于路侧激光雷达的背景提取和地面分割方法具有较高的实用性。According to the working principle of lidar,a method of background extraction and ground point cloud segmentation based on roadside scene is proposed in this paper.A background extraction method based on continuous frames is proposed,which uses line number and angle value as the basis of region division of point cloud data,and extracts background data by spatial density clustering method.A two-stage ground point cloud segmentation method based on ray features is proposed to divide the background point cloud into ground point sets and non-critical obstacle point sets.When processing real-time data,critical obstacles,ground points and non-critical obstacles can be quickly distinguished.Finally,the performance of the method is tested using the roadside liDAR data of the actual scene.Experimental results show that the proposed method of background extraction and ground segmentation based on roadside liDAR has high practicability.

关 键 词：激光雷达 背景提取 地面分割 

分 类 号：TN958.98[电子电信—信号与信息处理]