激光扫描点云中运动图像边界信息提取方法  

Extraction method of moving image boundary information in laser scanning point cloud

作　　者：刘健欣[1] 赵金凤[1] LIU Jian-xin;ZHAO Jin-feng(Pingdingshan University,Pingdingshan 467000,China)

机构地区：[1]平顶山学院,河南平顶山467000

出　　处：《激光与红外》2022年第10期1565-1570,共6页Laser & Infrared

基　　金：激光与物质相互作用国家重点开发及应用基金项目(No.SIKLM1697)资助。

摘　　要：使用距离地面较远的机载激光扫描获取运动图像边界信息数据,得到的数据点密度较低,无法精细提取运动图像边界信息,因此设计一种激光扫描点云中运动图像边界信息提取方法。首先获取激光扫描点云运动图像,对其中的三维激光扫描仪、惯性测量单元等装置进行组合并设置相关的参数,并对激光扫描点云得到的数据预处理,使用八叉树建立离散点之间的关系,通过点云体素化完成去噪,最后通过地物聚类完成运动图像特征点云三维信息的提取。实验结果表明,所设计方法提取准确性较高,完整性较好,提取质量效果较佳,均超过了97%,具有较好的提取效果。Using airborne laser scanning far from the ground to obtain moving image boundary information data,the data point density is low,and the moving image boundary information cannot be accurately extracted.Therefore,a moving image boundary information extraction method in laser scanning point cloud is designed.Firstly,the moving image of laser scanned point cloud is obtained,and the three-dimensional laser scanner,inertial measurement unit and other devices are combined and the relevant parameters are set,and the data obtained from laser scanned point cloud is preprocessed.The relationship between discrete points is established using octree,and the noise is removed through point cloud voxelization.Finally,the 3D information of feature point cloud of moving image is extracted by feature clustering.The experimental results show that the designed method has better extraction accuracy,better integrity and better extraction quality,all of which exceeded 97%,with better extraction effect.

关 键 词：激光扫描点云 运动图像 边界信息 八叉树 点云体素化 

分 类 号：P208[天文地球—地图制图学与地理信息工程] TP391.41[天文地球—测绘科学与技术]