基于密度与局部统计的单光子点云去噪方法  

Single photon point cloud denoising method based on density and local statistics

作　　者：潘超 李凉海 曹海翊[2] 赵一鸣 袁逸飞 韩晓爽 PAN Chao;LI Lianghai;CAO Haiyi;ZHAO Yiming;YUAN Yifei;HAN Xiaoshuang(Beijing Research Institute of Telemetry,Beijing 100076,China;Institute of Remote Sensing Satellite,China Academy of Space Technology,Beijing 100094,China)

机构地区：[1]北京遥测技术研究所,北京100076 [2]中国空间技术研究院遥感卫星总体部,北京100094

出　　处：《中国科学院大学学报（中英文）》2024年第2期268-274,共7页Journal of University of Chinese Academy of Sciences

基　　金：中国航天科技集团自主研发项目;“十四五”民用航天预研项目(D040107)资助。

摘　　要：针对北京遥测技术研究所自主研发的64通道机载单光子激光雷达,提出一种基于密度与局部统计的二维剖面点云去噪方法:在确定信号点云的高程区间后,先使用基于密度的改进空间聚类算法粗去噪,然后使用基于局部统计的统计移除离群点算法精去噪,获取信号点云。实验结果表明,本方法可适用于多种地物类型点云,高程均方根误差为0.27 m,准确率90.78%,精度高于常规点云去噪算法,满足国产机载单光子激光雷达获取高精度地表三维轮廓的技术需求。In this paper,aiming at a 64-channel airborne single-photon LiDAR system developed by Beijing Research Institute of Telemetry,a two-dimensional profile point cloud denoising method based on density and local statistics was proposed.First,the elevation range of the point cloud was determined;then a modified DBSCAN algorithm was utilized for coarse denoising;finally,the statistical outlier removal algorithm was adapted for fine denoising and the valid signal point cloud was obtained.The experimental result shows that the method proposed in this paper can adapt to different surface types,the root mean square error of elevation is about 0.27 m,and the accuracy is 90.87%,which is better than the conventional point cloud denoising methods,and can meet the technical requirements of domestic airborne single-photon LiDAR to obtain high-precision three-dimensional surface contours.

关 键 词：单光子三维成像激光雷达 点云去噪 局部统计 密度聚类 

分 类 号：P237[天文地球—摄影测量与遥感]