基于UAV影像点云密度的植被稀疏区DEM精度分析  

DEM accuracy analysis of sparse vegetation area based on point cloud density of UAV images

作　　者：戴志林 郭辉[1] 包勤跃 DAI Zhilin;GUO Hui;BAO Qinyue(School of Space Information and Surveying and Mapping Engineering,Anhui University of Science&Technology,Huainan,Anhui 232001,China;Mineral Resources Center of Tongling Nonferrous Metals Group Company Limited,Tongling,Anhui 244000,China)

机构地区：[1]安徽理工大学空间信息与测绘工程学院,安徽淮南232001 [2]铜陵有色金属集团股份有限公司矿产资源中心,安徽铜陵244000

出　　处：《北京测绘》2025年第3期316-321,共6页Beijing Surveying and Mapping

基　　金：矿山环境与灾害协同监测煤炭行业工程研究中心开放基金(KSXTJC202202)。

摘　　要：本文以无人机(UAV)倾斜摄影影像点云为数据源,通过对构建植被稀疏区数字高程模型(DEM)的最佳点云密度进行分析研究,从中选取出最佳点云密度,以实现高效、快速获取DEM数据。使用随机采样算法将原始点云以10%~90%密度进行抽稀,同时采用中误差对生成的DEM进行精度评价分析。结果显示:①在点云抽稀10%~40%时,中误差随着点云密度的减小而增大,同时在点云抽稀30%时中误差与原始点云几乎相似;②在点云抽稀40%~60%时,中误差变化较为平缓,但总体呈上升趋势;③在点云抽稀60%~90%时,随着点云密度的进一步减小,中误差随着点云密度的减小而迅速增大。得出结论,点云密度与DEM精度呈正相关,抽稀30%的点云成为在同类型条件植被稀疏区UAV倾斜摄影点云生成DEM的最佳点云密度。This paper analyzed the optimal point cloud density for constructing a digital elevation model(DEM)in a sparse vegetation area by using the point cloud in images obtained by unmanned aerial vehicle(UAV)oblique photography as the data source.It selected the optimal point cloud density,so as to achieve efficient and fast acquisition of DEM data.The original point cloud was thinned at a density of 10%-90%by a random sampling algorithm,and the accuracy of the generated DEM was evaluated by using the mean square error.The results show that①When the point cloud is thinned by 10%-40%,the mean square error increases with the decrease in point cloud density,and when the point cloud is thinned by 30%,the mean square error is almost similar to that of the original point cloud.②When the point cloud is thinned by 40%-60%,the change of mean square error is relatively flat,but the overall trend is upward.③When the point cloud is thinned by 60%-90%,the mean square error increases rapidly with the decrease in the point cloud density.Therefore,it is concluded that the point cloud density is positively correlated with the accuracy of DEM,and the point cloud thinned by 30%becomes the optimal point cloud density for generating DEM from UAV oblique photography point clouds in the same type of sparse vegetation area.

关 键 词：无人机(UAV)倾斜摄影 影像点云 点云密度 抽稀 数字高程模型(DEM) 

分 类 号：P237[天文地球—摄影测量与遥感]