范保青,姚剑敏,林志贤,严群,李成跃

• 1:福州大学物理与信息工程学院

摘要(Abstract)：

针对在三维点云环境下分离目标物体所出现的过度分割问题,提出一种结合随机抽样一致性和颜色差值区域聚类的分割方法。首先利用RANSAC算法去除场景中大部分平面,使得目标物体和连成片的点云脱离,然后结合点云的距离阈值和目标颜色差值,得到目标点云数据。针对L_1中值算法对曲率较大模型的骨架提取存在的不足,进行了改进。通过L_1中值算法对点云模型进行骨架提取,得到点云的骨架点,然后沿端点方向向外进行最大内切球的球心提取,最后连接多个球心及骨架末端点,得到符合人类视觉效果的骨架。改进的算法提高了L_1中值对曲率较大点云骨架提取的准确性。

关键词(KeyWords)： Kinect深度图像;三维点云;点云分割;改进骨架提取

Abstract:

Keywords:

基金项目(Foundation): 国家重点研发计划课题项目（2016YFB0401503）;; 广东省科技重大专项（2016B090906001）;; 福建省科技重大专项（2014HZ0003-1）;; 广东省光信息材料与技术重点实验室开放基金资助项目（2017B030301007）

作者(Author): 范保青,姚剑敏,林志贤,严群,李成跃

DOI: 10.16251/j.cnki.1009-2307.2021.01.022

参考文献(References)：

• [1]HAMMOUDI K,DORNAIKA F,SOHEILIAN B,et al.Extracting outlined planar clusters of street facades from 3D point clouds[C]∥2010 Canadian Conference on Computer and Robot Vision.[S.l.].IEEE,2010:122-129．
• [2]BASSIER M,BONDUEL M,VAN GENECHTEN B,et al.Segmentation of large unstructured point clouds using octree-based region growing and conditional random fields[J].The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences,2017,42(2W8):25-30．
• [3]李娜，马一薇，杨洋，等．利用RANSAC算法对建筑物立面进行点云分割[J]．测绘科学，2011,36(5):144-145.(LI Na,MA Yiwei,YANG Yang,et al.Segmentation of building facade point clouds using RANSAC[J].Science of Surveying and Mapping,2011,36(5):144-145．）
• [4]周平华，熊彪．半自动机载LiDAR点云建筑物三维重建方法[J]．测绘科学，2017,42(5):128-130.(ZHOUPinghua,XIONG Biao.3Dreconstruction method of buildings of semi-automatic airborne LiDAR point clouds[J].Science of Surveying and Mapping,2017,42(5):128-130．）
• [5]BLUM H.Biological shape and visual science[J].Journal of Theoretical Biology,1973,38(2):205-287．
• [6]TAGLIASACCHI A,ZHANG H,COHENOR D.Curve skeleton extraction from incomplete point cloud[J].ACM Transactions on Graphics,2009,28(3):1-9．
• [7]CAO J,TAGLIASACCHI A,OLSON M,et al.Point cloud skeletons via laplacian based contraction[C]∥2010Shape Modeling International Conference.IEEE,2010:187-197．
• [8]LI G,HUI H,WU S,et al.L1-medial skeleton of point cloud[J].ACM Transactions on Graphics,2013,32(4):1-8．
• [9]王晓洁，周元峰，潘晓，等．一种鲁棒的三维点云骨架提取方法[J]．中国科学（信息科学），2017,47(7):832-845.(WANG Xiaojie,ZHOU Yuanfeng,PANXiao,et al.A robust 3Dpoint cloud skeleton extraction method[J].Scientia Sinica Informationis,2017,47(7):832-845．）
• [10]晁莹，耿国华，张雨禾，等．基于区域分割的点云骨架提取算法[J]．计算机工程，2017,43(10):222-227.(CHAO Ying,GENG Guohua,ZHANG Yuhe,et al.Point cloud skeleton extraction algorithm based on region segmentation[J].Computer Engineering,2017,43(10):222-227．）
• [11]王腾，卜乐平，黄忠义．一种深度图像障碍物检测新方法[J]．测绘科学，2017,42(6):104-111.(WANG Teng,BU Leping,HUANG Zhongyi.A new method for obstacle detection based on Kinect depth image[J].Science of Surveying and Mapping,2017,42(6):104-111．）
• [12]杨晶东，孙磊明，邵雨婕，等．一种有效的自主机器人迭代最近点定位算法[J]．机械工程学报，2018,54(15):15-20.(YANG Jingdong,SUN Leiming,SHAOYujie,et al.Efficient iterative closest point localisation algorithm for autonomous robots[J].Journal of Mechanical Engineering,2018,54(15):15-20．）
• [13]FISCHLER M A,BOLLES,ROBERT C.Random sample consensus:aparadigm for model fitting with applications to image analysis and automated cartography[J].Communications of the ACM,1981,6(6):381-395．
• [14]朱磊，王健，许开辉，等．采用聚类分析的车载点云地物分类[J]．测绘科学，2016,41(4):77-82.(ZHU Lei,WANG Jian,XU Kaihui,et al.Classification method of vehicle-borne LiDAR point cloud based on clustering[J].Science of Surveying and Mapping,2016,41(4):77-82．）
• [15]缪永伟，陈程，孙瑜亮，等．基于最大内切球拟合的网格模型骨架提取[J]．计算机辅助设计与图形学学报，2018,30(10):20-28.(MIAO Yongwei,CHEN Cheng,SUN Yuliang,et al.Skeleton extraction of 3D meshes by fitting maximal inscribed spheres[J].Journal of Computer-Aided Design&Computer Graphics,2018,30(10):20-28．）
• [16]LIEUTIER A.Any open bounded subset of Rn has the same homotopy type as its medial axis[C]∥Eighth ACM Symposium on Solid Modeling&Applications.2003:1029-1046.