李启才,赵闯姓

• 1:青海省遥感测绘院

摘要(Abstract)：

针对在复杂城市环境中难以有效自动提取树木信息的问题,该文首先基于移动激光扫描(MLS)数据,利用超体聚类对点云数据结构进行组织管理;然后从超体素局部上下文信息中提取去趋势几何特征,结合去趋势几何上下文特征,采用随机森林分类器对树木进行初始语义标记;接着,基于局部上下文信息进行迭代正则化,在全局图模型上进行整体优化,从而对初始语义分类结果进行空间平滑;最后根据语义标记结果,基于图割算法实现单木分割。该方法基于超体素的结构可以有效地保持场景中目标的几何边界,而且提升了处理效率。去趋势几何特征可以克服局部上下文中的冗余和显著性信息,使得获取的特征更具代表性。实验结果表明,该方法在3个数据集的树木语义标记结果达到90%左右,对结构简单且稀疏分布的树木都能正确提取。

关键词(KeyWords)： 点云分割;移动激光扫描系统;超体聚类;图割算法;局部上下文特征;单木分割

Abstract:

Keywords:

基金项目(Foundation):

作者(Author): 李启才,赵闯姓

DOI: 10.16251/j.cnki.1009-2307.2020.09.018

参考文献(References)：

• [1] LANDRIEU L,SIMONOVSKY M.Large-scale point cloud semantic segmentation with superpoint graphs[C]//31th Conference on Computer Vision and Pattern Recognition.Salt Lake City,USA:IEEE,2018:4558-4567.
• [2] HACKEL T,WEGNER J,SCHINDLER K.Fast semantic segmentation of 3D point clouds with strongly varying density[J].ISPRS Annals of Photogrammetry Remote Sensing & Spatial Information Sciences,2016,3(3):177-184.
• [3] WEINMANN M,BORIS J.Semantic point cloud interpretation based on optimal neighborhoods,relevant features and efficient classifiers[J].ISPRS Journal of Photogrammetry and Remote Sensing,2015(105):286-304.
• [4] 于永涛.大场景车载激光点云三维目标检测算法研究[D].厦门:厦门大学,2015.(YU Xiantao.Research on 3D object detection algorithm for mobile laser scanning point cloud in large scene[D].Xiamen:Xiamen University,2015.)
• [5] ZHOU Y,WANG D,XIE X.A fast and accurate segmentation method for ordered LiDAR point cloud of large-scale scenes[J].IEEE Geoscience & Remote Sensing Letters,2014,11(11):1981-1985.
• [6] YANG B,DONG Z.A shape-based segmentation method for mobile laser scanning point clouds[J].ISPRS Journal of Photogrammetry & Remote Sensing,2013,81(7):19-30.
• [7] YAO W,WEI Y.Detection of 3D individual trees in urban areas by combining airborne LiDAR data and imagery[J].IEEE Geoscience & Remote Sensing Letters,2013,10(6):1355-1359.
• [8] LI J,HU B,THOMAS N.Classification of tree species based on structural features derived from high density LiDAR data[J].Agricultural & Forest Meteorology,2013,171/172:104-114.
• [9] YANG B,DONG Z,ZHAO G,DAI W.Hierarchical extraction of urban objects from mobile laser scanning data[J].ISPRS Journal of Photogrammetry & Remote Sensing,2015(99):45-57.
• [10]PAPON J,ABRAMOV A,SCHOELER M. Voxel cloud connectivity segmentation—supervoxels for point clouds[C]∥26th Conference on Computer Vision and Pattern Recognition.Portland,USA:IEEE,2013:2027-2034.
• [11]LIN Y,WANG C.Toward better boundary preserved supervoxel segmentation for 3D point clouds[J].ISPRS Journal of Photogrammetry&Remote Sensing,2018,143:39-47.
• [12]SUN Z,XU Y,HOEGNER L.Classification of MLS point clouds in urban scenes using detrended geometric features from supervoxel-based local context[J].ISPRS Annals of the Photogrammetry,Remote Sensingand Spatial Information Sciences,2018,4(2):271-278.
• [13]LEHTOMAKI M,ANTTONI J,JUHA H. Object classification and recognition from mobile laser scanning point clouds in a road environment[J].IEEE Transactions on Geoscience and Remote Sensing,2015,54(2):1226-1239.
• [14]ZHONG R,WEI J,SUN W.A method for extracting trees from vehicle-borne laser scanning data[J].Mathematical&Computer Modelling,2013,58(9/10):733-742.
• [15]BAO F,WU R.Delineation of individual tree crowns from mobile laser scanning data for intelligent transportation[J].Journal of Xiamen University,2017,56(1):117-122.