一种改进的车载激光扫描点云滤波算法  

An Improved Filtering Algorithm for Vehicle-borne Laser Scanning Point Cloud

作　　者：金芳芳 张菲 JIN Fangfang;ZHANG Fei(Hangzhou Fangyuan Surveying and Mapping Technology Service Co.,Ltd.,Hangzhou 310011,China;Hangzhou Zongyue Surveying and Mapping Technology Consulting Co.,Ltd.,Hangzhou 310000,China)

机构地区：[1]杭州方圆测绘技术服务有限公司,浙江杭州310011 [2]杭州纵越测绘技术咨询有限公司,浙江杭州310000

出　　处：《测绘与空间地理信息》2024年第9期161-164,共4页Geomatics & Spatial Information Technology

摘　　要：为了提高车载激光扫描点云滤波精度,同时解决目前点云滤波算法中存在的人机交互多、参数不易把控等问题,本文在经典不规则三角网(Triangulated Irregular Network,TIN)迭代加密滤波算法的基础上,优化地面种子点选取与进行格网大小自适应变化,提出了一种改进的TIN迭代加密滤波算法。该改进滤波算法实现非地面点滤波的主要途径为:首先,对原始车载点云数据构建格网,通过计算邻域卷积以及构造虚拟种子点实现地面种子点的确定;其次,根据设置阈值条件将待判断激光点加密至TIN中;最后,通过自适应改变格网大小并进行迭代滤波完成地面点提取。使用两段城市典型道路车载激光扫描点云数据进行实验,结果表明本文改进滤波算法的滤波性能更好,具有较好的地形适应性。In order to improve the filtering accuracy of vehicle-borne laser scanning point cloud,and at the same time,to improve the problems existing in the current point cloud filtering algorithm,such as many human-computer interactions and difficulty in controlling parameters,this paper proposes an improved Triangulated Irregular Network(TIN)iterative densifying filtering algorithm based on the classical TIN iterative densifying filtering algorithm,which optimizes the selection of ground seed points and adaptively changes the grid size.The main approaches of the improved filtering algorithm to realize non-ground point filtering are as follows:firstly,the grid is constructed for the original vehicle-borne point cloud data,and the ground seed points are determined by calculating the neighborhood convolution and constructing the virtual seed points;secondly,the laser point to be determined is densified into the TIN according to the setting threshold condition;finally,the ground points are extracted by adaptively changing the grid size and performing iterative filtering.The experiment is carried out on the point cloud data of two typical urban roads.The results show that the improved filtering algorithm has better filtering performance and better terrain adaptability.

关 键 词：车载激光扫描 点云滤波 不规则三角网 改进滤波算法 

分 类 号：P225[天文地球—大地测量学与测量工程]