基于激光雷达的建筑物轮廓信息识别提取研究  

Research on identification and extraction of building profile information based on lidar

作　　者：王震[1] 母德强[1] Wang Zhen;Mu Deqiang(Changchun University of Technology,Changchun 130000,China)

机构地区：[1]长春工业大学,吉林长春130000

出　　处：《无线互联科技》2022年第19期140-142,共3页Wireless Internet Technology

摘　　要：对于采用激光雷达扫描的建筑物进行点云重建的关键,在于建筑物轮廓信息的提取,并且要能够保证其精确度和完整性。文章提出了基于几何特征与语义描述的外轮廓信息改进识别方法,比较好地解决了目前建筑物轮廓提取中存在精度较低、后续处理麻烦的问题。最终建立了数学求解模型,并且以采集的单栋建筑物点云数据作为实验样本,进行了建筑物轮廓信息提取实验。结果表明,采用的改进方法不仅理论上能够实现,还大大提高了建筑物轮廓信息提取的清晰度,通过多次实验验证,此方法优越性明显。文中所述方法对点云数据分割,特别是建筑物与地面点的分割,建筑物的三维重建具有极大的意义。The main task of current point cloud reconstruction of buildings for lidar scanning is the extraction of contour information from buildings.The purpose of building outline extraction is to obtain a building profile with high integrity and high precision.In this paper,we propose an improved recognition method for outer contour information based on geometric features and semantic descriptions,which solves the problem of low precision and troublesome subsequent processing in current building contour extraction.Finally,the mathematical solution model was established,and the point cloud data of the single building was taken as the experimental sample,and the building profile information extraction experiment was carried out.The results show that the improved method can not only be realized theoretically,but also greatly improves the clarity of the contour information extraction of the building.It is proved by many experiments that the method has obvious advantages.The method described in this paper has great significance for point cloud data segmentation,especially the segmentation of buildings and ground points,and the three⁃dimensional reconstruction of buildings.

关 键 词：轮廓识别 信息提取 几何特征 点云重建 

分 类 号：TN958.98[电子电信—信号与信息处理] TU198[电子电信—信息与通信工程]