车载激光点云与全景影像匹配可视化  

Matching visualization of vehicle-borne laser point cloud and panoramic image

作　　者：韦景元 赵彦利 WEI Jingyuan;ZHAO Yanli(School of Geomatics and Urban Spatial Information,Beijing University of Civil Engineering and Architecture,Beijing 102616 China;Terra InfoTech(Beijing)Company Limited,Beijing 100192,China)

机构地区：[1]北京建筑大学测绘与城市空间信息学院,北京102616 [2]泰瑞数创科技(北京)股份有限公司,北京100192

出　　处：《北京测绘》2023年第7期975-980,共6页Beijing Surveying and Mapping

基　　金：国家自然科学基金(42171416)。

摘　　要：通过车载测量系统可获取海量的点云与全景影像,点云具有精确的三维空间坐标,全景具有丰富的纹理信息,将两者在进行匹配可视化可完成数据的优势互补,发挥更大作用。因此,本文根据道路点云的特征,设计了一种节点互斥的稀疏点云八叉树数据组织结构,利用Three.js三维引擎设计点云调度机制实现海量点云与全景在Web端的快速可视化;提出将点云与全景转换到同一模态并基于空间地物几何特征的点云全景匹配可视化流程。通过实验表明,本文技术流程能够实现车载点云与全景数据在Web端的精准匹配与流畅渲染,且匹配流程快速、简便,视觉效果较好,具有较高的实用价值。Massive point clouds and panoramic images can be obtained through the vehicle-mounted measurement system.The point clouds have accurate three-dimensional spatial coordinates and the panorama has rich texture information.The advantages of the two can complement each other and play a greater role in matching and visualization data.Therefore,according to the characteristics of road point clouds,this paper designed a sparse octree data organization structure of point clouds with mutually exclusive nodes,and applied three.js to design point cloud scheduling mechanism and realize rapid visualization of massive point clouds and panoramic views on the Web end.A visualization process of point cloud panorama matching was proposed,which converted point cloud and panorama into the same mode and was based on geometric features of spatial ground objects.Experiments showed that the technical process in this paper could achieve accurate matching and smooth rendering of vehiclemounted point cloud and panoramic data on the Web side,and the matching process was fast and simple,with good visual effect and high practical value.

关 键 词：车载测量系统 点云 全景影像 三维引擎 匹配可视化 

分 类 号：P258[天文地球—测绘科学与技术]