树木的骨架模型能够抽象的反映出树木的几何形状和拓扑结构等特征信息，已经广泛应用于树木的三维建模、数字林业以及景观设计等领域。激光雷达扫描技术具有全天候工作、测距能力高效精确以及穿透能力较强等优点，在林业应用中发展迅速。激光雷达扫描仪可以获取树木表面准确的三维点云数据，并通过对点云进行处理，得到树木的生长参数，已经成为近年来林业研究的热点之一。本文以杨树为研究对象，利用机载激光雷达和背包式激光雷达融合的点云数据，提出了基于单木分割的配准算法、基于主成分分析与区域生长的点云枝叶分类算法和基于聚类的树木骨架提取方法。本文的研究重点为基于多源融合的点云数据构建出单木的骨架模型。主要工作和结论如下：

（1）多源点云数据融合算法：提出了一种基于单木分割的配准算法，将背包式激光雷达和机载激光雷达的扫描数据进行融合，以提高点云数据的质量。将提出的算法与迭代最近点和改进的正态分布变化配准算法方法进行对比，在三块样地上进行实验，利用均方根误差测试算法的性能。结果表明：该方法下三块样地的均方根误差分别为2.27 cm、2.94 cm、1.98 cm，与另外两种算法的配准结果相比，本文提出的方法对于两组点云数据的配准具有更高的精度。

（2）树木生长参数及枝干点云提取：通过几何计算、最小二乘拟合圆等算法实现了树高、胸径以及冠幅的提取。本文提出了一种基于主成分分析与区域生长的点云枝叶分类算法，并将该算法在榆树和杨树上进行实验，结果表明该算法具有较高的精度。

（3）骨架点提取和链接：提出了一种改进的基于聚类的树木骨架提取方法，并对主干与细枝分别进行处理。主干采用的是随机采样一致性圆柱拟合方法，而细枝采用的是最小二乘直线拟合方法。最后对所有提取到的骨架点进行链接，得到了树木比较真实的骨架模型。

本文构建的骨架模型可以简单有效的描述树木的几何信息和拓扑连接信息，在构建树木的三维模型以及研究森林的碳循环、叶面积指数、生物量估算等方面具有重要意义。

The skeleton model of trees can abstractly reflect the characteristic information such as geometric shape and topological structure of trees. Therefore, it has been widely used in three-dimensional model of trees, digital forestry, landscape design and other fields. LiDAR scanning technology has the characteristics of all-weather operation, high efficiency and strong penetration. Its application in forestry investigation is developing rapidly. LiDAR scanners can obtain accurate 3D point cloud data on the surface of trees. By processing the point cloud data, the growth parameters of trees are obtained. This has become one of the hotspots of forestry research in recent years. The object of this research is poplar. The data used in this research is the point cloud data of airborne LiDAR and knapsack LiDAR. In this study, a registration algorithm based on single tree segmentation, a branch and leaf classification algorithm based on principal component analysis and region growth, and a tree skeleton extraction method based on clustering were proposed. The focus of this paper is to build a single tree skeleton based on multi-source point cloud data.

The main research works and conclusions are as follows:

(1) Automatic registration algorithms for multi-source point cloud data

In this paper, a new registration algorithm based on single tree segmentation was proposed, which combined the scanning data of knapsack LiDAR and airborne LiDAR. This can improve the quality of point cloud data. This paper chose to compare it with the iterative nearest point method and the improved normal distribution transformation method. Experiments were carried out on three sample plots, and the performance of algorithms was tested by root mean square error. The root mean square errors of the proposed method for three sample plots were 2.27 cm, 2.94 cm and 1.98 cm respectively. Compared with the registration results of the other two algorithms, the proposed method had higher accuracy.

(2) Tree growth parameters and branch point cloud extraction

The tree height, DBH and crown width were extracted by geometric calculation and least square fitting circle algorithm. In this paper, a point cloud branch and leaf classification algorithm based on principal component analysis and region growth was proposed. Then the proposed algorithm was tested on the data of elm and poplar, and the results showed that the proposed algorithm had high accuracy.

(3) Skeleton point extraction and linking

In this paper, an improved tree skeleton extraction method based on clustering was proposed. The trunk and twigs were treated separately. The trunk adopted the random sampling consistent cylindrical fitting method, while the twig used the least square straight line fitting method. All the extracted skeleton points were linked, and finally a more real skeleton model of the tree was obtained.

The skeleton model can effectively and simply describe the geometric information and topological connection information of trees. It is of great significance in building an accurate three-dimensional model of trees and studying the carbon cycle, leaf area index and biomass estimation of forests.

[1] 林清山, 洪伟. 中国森林碳储量研究综述[J]. 中国农学通报, 2009, 25(6): 220-224.

[2] Pan Y D, Birdsey R A, Fang J Y, et al. A large and persistent carbon sink in the world's forests[J]. Science, 2011, 333(6045): 988-993.

[3] 汤旭光，刘殿伟，王宗明，等. 森林地上生物量遥感估算研究进展[J]. 生态学杂志, 2012, 31(5): 1311-1318.

[4] 胡凯龙. 不同空间尺度森林参数多源遥感反演方法[D]. 北京: 中国矿业大学(北京), 2018.

[5] 李德仁, 王长委, 胡月明, 等. 遥感技术估算森林生物量的研究进展[J]. 武汉大学学报：信息科学版, 2012, 37(6): 631-635.

[6] Myneni R B, Dong J, Tucker C J, et al. A large carbon sink in the woody biomass of Northern forests[J]. Proceedings of the National Academy of Sciences, 2001, 98(26): 14784-14789.

[7] Zhang C, Zhou Y, Qiu F. Individual Tree Segmentation from LiDAR Point Clouds for Urban Forest Inventory[J]. Remote Sensing, 2015, 7(6): 7892-7913.

[8] Wulder, M. Optical remote-sensing techniques for the assessment of forest inventory and biophysical parameters[J]. Progress in Physical Geography, 1998, 22(4): 449.

[9] 刘鲁霞, 庞勇. 机载激光雷达和地基激光雷达林业应用现状[J]. 世界林业研究, 2014, 27(1): 49-56.

[10] 张文军. 三维激光扫描技术及其应用[J]. 测绘标准化, 2016, 32(2): 42-44.

[11] Lefsky M A, Cohen W B, Parker G G. Lidar Remote Sensing for Ecosystem Studies[J]. BioScience, 2002, 52(1): 19-30.

[12] Wehr A, Lohr U. Airborne laser scanning—an introduction and overview[J]. Isprs Journal of Photogrammetry Remote Sensing, 1999, 54(2-3): 68-82.

[13] Charlton M E, Large A R G, Fuller I C. Application of airborne LiDAR in river environments: the River Coquet, Northumberland, UK[J]. Earth Surface Processes Landforms, 2010, 28(3): 299-306.

[14] Marenco F, Johnson B, Turnbull K, et al. Airborne lidar observations of the 2010 Eyjafjallajökull volcanic ash plume[J]. Journal of Geophysical Research, 2011(116): D20.

[15] Hudnut K W. High-Resolution Topography along Surface Rupture of the 16 October 1999 Hector Mine, California, Earthquake (Mw 7.1) from Airborne Laser Swath Mapping[J]. Bulletin of the Seismological Society of America, 2002, 92(4): 1570-1576.

[16] Jensen J L R, Humes K S, Vierling L A. Discrete return lidar-based prediction of leaf area index in two conifer forests[J]. Remote Sensing of Environment, 2008, 112(10): 3947-3957.

[17] Hyyppa J, Kelle O, Lehikoinen M, et al. A segmentation-based method to retrieve stem volume estimates from 3-D tree height models produced by laser scanners[J]. IEEE Transactions on geoscience and remote sensing, 2001, 39(5): 969-975.

[18] Liang X, Juha H, Harri K, et al. International benchmarking of terrestrial laser scanning approaches for forest inventories[J]. Isprs Journal of Photogrammetry Remote Sensing, 2018, 144(2):137-179.

[19] 钟涛. 基于三维激光扫描测量技术的露天矿山变形监测研究[J]. 科技与创新, 2015(15): 1-9.

[20] 张亚. 三维激光扫描技术在三维景观重建中的应用研究[D]. 西安: 长安大学, 2011.

[21] 彭文博, 杨武年, 王鹏. 三维激光扫描技术在古建筑模型重建中的应用[J]. 地理空间信息, 2016, 14(3): 94-99.

[22] 索俊锋, 刘勇, 蒋志勇, 等. 基于三维激光扫描点云数据的古建筑建模[J]. 测绘科学, 2017, 42(3): 179-185.

[23] 徐云乾, 袁明道, 刘建文, 等. 地面三维激光扫描技术及其在水利工程中的应用[J]. 广东水利电力职业技术学院学报, 2016, 14(1): 13-16.

[24] Erikson M, Karin V. Finding tree-stems in laser range images of young mixed stands to perform selective cleaning[J]. Proceedings of the ScandLaser Scientific Workshop on Airborne Laser Scanning of Forest, 2003, 244-250.

[25] Hopkinson C, Chasmer L, Young-Pow C, et al. Assessing forest metrics with a ground-based scanning lidar[J]. Canadian Journal of Forest Research, 2004, 34(3): 573-583.

[26] Parker G, Berger H. A Portable LIDAR System for Rapid Determination of Forest Canopy Structure[J]. Journal of Applied Ecology, 2004, 41(4): 755-767.

[27] Watt P J, Donoghue D N M. Measuring forest structure with terrestrial laser scanning[J]. International Journal of Remote Sensing, 2005, 26(7): 1437-1446.

[28] Liang X, Kankare V, Yu X. Automated Stem Curve Measurement Using Terrestrial Laser Scanning[J]. IEEE Transactions on Geoscience Remote Sensing, 2013, 52(3): 1739-1748.

[29] Liang C, Lihua T, Yu W, et al. Extraction of Urban Power Lines from Vehicle-Borne LiDARData[J]. Remote Sensing, 2014, 6(4): 3302-3320.

[30] Jochem A, Höfle B, Rutzinger M. Extraction of Vertical Walls from Mobile Laser Scanning Data for Solar Potential Assessment[J]. Remote Sensing, 2011, 3(4): 650-667.

[31] Liang X, Hyyppa J, Kukko A, et al. The Use of a Mobile Laser Scanning System for Mapping Large Forest Plots[J]. IEEE Geoscience Remote Sensing Letters, 2014, 11(9): 1504-1508.

[32] Yang B, Fang L, Li J. Semi-automated extraction and delineation of 3D roads of street scene from mobile laser scanning point clouds[J]. Isprs Journal of Photogrammetry Remote Sensing, 2013, 79(5): 80-93.

[33] 杨猛, 刘杰, 杨锋, 等. 车载激光雷达技术在城市1:500地形图测绘工作中的应用[J]. 测绘与空间地理信息, 2020, 43(8): 5-14.

[34] 满其霞. 激光雷达和高光谱数据融合的城市土地利用分类方法研究[D]. 上海: 华东师范大学, 2015.

[35] Bauwens S, Bartholomeus H, Calders K, et al. Forest Inventory with Terrestrial LiDAR: A Comparison of Static and Hand-Held Mobile Laser Scanning[J]. Forests, 2016, 7(6): 127-138.

[36] Seki S, Tsubouchi T, Saratat S, et al. Forest mapping and trunk parameter measurement on slope using a 3D-LIDAR[C]. 2017 IEEE/SICE International Symposium on System Integration (SII). IEEE, 2017:380-389.

[37] Marselis S M, Yebra M, Jovanovic T. Deriving comprehensive forest structure information from mobile laser scanning observations using automated point cloud classification[J]. Environmental Modelling Software, 2016(82):142-151.

[38] Del Perugia B, Giannetti F, Chirici G. Influence of Scan Density on the Estimation of Single-Tree Attributes by Hand-Held Mobile Laser Scanning[J]. Forests, 2019, 10(3): 277-289.

[39] Holmgren J, Tulldahl H M, Nordlöf J, et al. Estimation of Tree Position and STEM Diameter Using Simultaneous Localization and Mapping with Data from a Backpack-Mounted Laser Scanner[J]. ISPRS - International Archives of the Photogrammetry, Remote Sensing Spatial Information Sciences, 2017(W3):59-63.

[40] Chen S, Liu H, Feng Z, et al. Applicability of personal laser scanning in forestry inventory[J]. PLoS ONE, 2019, 14(2): e0211392..

[41] Joseph R, Emily W, Martin S, et al. Assessing Handheld Mobile Laser Scanners for Forest Surveys[J]. Remote Sensing, 2015, 7(1): 1095-1111.

[42] Liang X, Kukko A, Kaartinen H, et al. Possibilities of a Personal Laser Scanning System for Forest Mapping and Ecosystem Services[J]. Sensors, 2014, 14(1): 1228-1248.

[43] Francesca G, Nicola P, Valerio Q, et al. Integrating terrestrial and airborne laser scanning for the assessment of single-tree attributes in Mediterranean forest stands[J]. European Journal of Remote Sensing, 2018, 51(1): 795-807.

[44] Cabo C, Pozo S D, Rodríguez-Gonzálvez P, et al. Comparing Terrestrial Laser Scanning (TLS) and Wearable Laser Scanning (WLS) for Individual Tree Modeling at Plot Level[J]. Remote Sensing, 2018, 10(4): 540-553.

[45] Ivar O, Marius H, Francesca G, et al. Comparing Three Different Ground Based Laser Scanning Methods for Tree Stem Detection[J]. Remote Sensing, 2018, 10(4): 538-542.

[46] Hu L, Xiao J, Wang Y. An automatic 3D registration method for rock mass point clouds based on plane detection and polygon matching[J]. The Visual Computer, 2019( 36):669-681.

[47] Besl P J, Mckay H D. A method for registration of 3-D shapes[J]. IEEE Transactions on Pattern Analysis Machine Intelligence, 1992, 14(2): 239-256.

[48] Bouaziz S, Tagliasacchi A, Pauly M. Sparse Iterative Closest Point[J]. Computer Graphics Forum, 2013, 32(5): 113–123.

[49] Mohammadzade H, Hatzinakos D. Iterative Closest Normal Point for 3D Face Recognition[J]. IEEE Transactions on Software Engineering, 2013, 35(2): 381-397.

[50] Habib A, Ghanma M, Morgan M, et al. Photogrammetric and LiDAR data registration using linear feature [J]. 2005, 71(121): 699–707.

[51] Al-Durgham M, Habib A. A Framework for the Registration and Segmentation of Heterogeneous Lidar Data[J]. International Conference on Computer Vision in Remote Sensing, 2013(79):135–145.

[52] Yang B, Zang Y. Automated registration of dense terrestrial laser-scanning point clouds using curves[J]. ISPRS Journal of Photogrammetry Remote Sensing, 2014, 95(3): 109–121.

[53] Thirion J P. The extremal mesh and the understanding of 3D surfaces[J]. International Journal of Computer Vision, 1993, 19(2): 115-128.

[54] Dorai C, Jain A K. COSMOS-A representation scheme for 3D free-form objects[J]. IEEE Transactions on Pattern Analysis Machine Intelligence, 1997, 19(10): 1115-1130.

[55] Bae K-H, Lichti D. A method for automated registration of unorganised point clouds[J]. Isprs Journal of Photogrammetry Remote Sensing, 2008, 63(1): 36-54.

[56] 朱延娟, 周来水, 张丽艳. 散乱点云数据配准算法[J]. 计算机辅助设计与图形学学报, 2006, 18(4): 475-481.

[57] 徐金亭, 孙玉文, 刘伟军. 复杂曲面加工检测中的精确定位方法[J]. 机械工程学报 2007, 43(6): 175-179.

[58] Rusu R B, Marton Z C, Blodow N, et al. Persistent Point Feature Histograms for 3D Point Clouds[C]. In Proceedings of the 10th International Conference on Intelligent Autonomous Syst (IAS-10). Baden-Baden. 2008: 119–128.

[59] Rusu R B, Blodow N, Beetz M. Fast Point Feature Histograms (FPFH) for 3D registration[C]. In Proceedings of the 2009 IEEE International Conference on Robotics and Automation. Kobe. 2009: 3212–3217.

[60] Fischler M. Random sample consensus : A paradigm for model fitting with applications to image analysis and automated cartography[J]. 1981(24):381-395.

[61] Chen C S, Hung Y P, Cheng J B. RANSAC-Based DARCES: A New Approach to Fast Automatic Registration of Partially Overlapping Range[J]. IEEE Transactions on Pattern Analysis Machine Intelligence, 1999, 21(11):1229-1234.

[62] Aiger D, Mitra N J. 4-points Congruent Sets for Robust Surface Registration[J]. 2008, 27(1): 1-10.

[63] Meng Y, Zhang H. Registration of point clouds using sample-sphere and adaptive distance restriction[J]. The Visual Computer, 2011, 27(6-8): 543-553.

[64] Chung D H, Yun I D, Sang U L. Registration of multiple-range views using the reverse-calibration technique[J]. 1998, 31(4): 457-464.

[65] Magnusson M, Lilienthal A, Duckett T. Scan registration for autonomous mining vehicles using 3D-NDT[J]. 2007, 24(10): 803-827.

[66] Segal A, Hhnel D, Thrun S. Generalized-ICP[C]. Robotics: Science and Systems V, University of Washington, Seattle, USA, June 28 - July 1, 2009.

[67] Censi A. An ICP variant using a point-to-line metric[C]. IEEE International Conference on Robotics & Automation. Pasadena, CA. 2008: 19-25.

[68] Fitzgibbon A W. Robust registration of 2D and 3D point sets[J]. Image and Vision Computing, 2003, 21(13–14): 1145-1153.

[69] Sandhu R, Dambreville S, Tannenbaum A R. Point Set Registration via Particle Filtering and Stochastic Dynamics[J]. IEEE Computer Society, 2010, 32(8):1459-1473.

[70] Ko K H, Maekawa T, Patrikalakis N. An algorithm for optimal free-form object matching[J]. Computer-Aided Design, 2003, 35(10): 913-923.

[71] Verroust A, Lazarus F. Extracting Skeletal Curves from 3D Scattered Data[J]. Visual Computer, 2000, 16(1): 15-25.

[72] Delagrange, Sylvain. Reconstruction and analysis of a deciduous sapling using digital photographs or terrestrial-LiDAR technology[J]. Annals of Botany, 2011, 108(6): 991-1003.

[73] Yan D M, Wintz J, Mourrain B, et al. Efficient and robust reconstruction of botanical branching structure from laser scanned points[C]. 11th IEEE International Conference on Computer-Aided Design and Computer Graphics. Huangshan, China, 2009:572-575.

[74] Bucksch A, Menenti L M. SkelTreRobust Skeleton Extraction from Imperfect Point Clouds[J]. The Visual Computer, 2010(26):1283–1300

[75] Hu S, Li Z, Zhang Z, et al. Efficient Tree Modeling from Airborne LiDAR Point Clouds[J]. Computers Graphics, 2017, 67(2): 1-13.

[76] Au O K-C, Tai C L, Chu H K, et al. Skeleton extraction by mesh contraction[J]. ACM Transactions on Graphics, 2008, 27(3):1–10.

[77] Tagliasacchi A, Zhang H, Cohen-Or D. Curve skeleton extraction from incomplete point cloud[J]. ACM Transactions on Graphics, 2009, 28(3):1–9.

[78] Cao J, Tagliasacchi A, Olson M, et al. Point Cloud Skeletons via Laplacian Based Contraction[C]. SMI 2010, Shape Modeling International Conference, Aix en Provence, France, June 21-23 2010. Aix-en-Provence, France, 2010.

[79] Flood M. Lidar activities and research priorities in the commercial sector[J]. International Archives of the Photogrammetry Remote Sensing Spatial Information Sciences, 2001, 34(W4): 3-7.

[80] Zhao X Q, Guo Q H, Su Y J, et al. Improved progressive TIN densification filtering algorithm for airborne LiDAR data in forested areas[J]. ISPRS Journal of Photogrammetry Remote Sensing, 2016, 117(1):79-91.

[81] Shi X, Tao L, Han X. Improved Iterative Closest Point(ICP) 3D point cloud registration algorithm based on point cloud filtering and adaptive fireworks for coarse registration[J]. International Journal of Remote Sensing, 2020, 41(8): 3197-3220.

[82] Chen Q J, Wang X, Hang M R, et al. Research on the improvement of single tree segmentation algorithm based on airborne LiDAR point cloud[J]. Open Geosciences, 2021, 13(1): 705-716.

[83] Goutte C, Gaussier E. A Probabilistic Interpretation of Precision, Recall and F-Score, with Implication for Evaluation[C]. European Conference on Information Retrieval. France, 2005:345–359.

[84] Yang B, San-Bao L, Wang L. Fast Point Cloud Registration Algorithm Based on NDT and ICP[J]. Science Technology and Engineering, 2017(17):91-95.

[85] Zhang W, Chen Y, Wang H, et al. Efficient registration of terrestrial LiDAR scans using a coarse-to-fine strategy for forestry applications[J]. Agricultural Forest Meteorology, 2016(225):8-23.

[86] Lin L, Yang B, Dai W, et al. Automatic Registration of Multi-View Terrestrial Laser Scanning Point Clouds in Forest Environments[J]. Journal of Geomatics, 2015(40):86-89.

[87] Solberg S, Naesset E, Bollandsas O M. Single tree segmentation using airborne laser scanner data in a structurally heterogeneous spruce forest[J]. Photogrammetric Engineering and Remote Sensing, 2006, 72(12): 1369-1378.

[88] Morsdorf F, Meier E, Tz B, et al. LIDAR-based geometric reconstruction of boreal type forest stands at single tree level for forest and wildland fire management[J]. Remote Sensing of Environment, 2004, 92(3): 353-362.

[89] Côté J F, Widlowski J L, Fournier R, et al. The structural and radiative consistency of three-dimensional tree reconstructions from terrestrial lidar[J]. Remote Sensing of Environment, 2009, 113(5): 1067-1081.

[90] Yao T, Yang X Y, Zhao F, et al. Measuring forest structure and biomass in New England forest stands using Echidna ground-based lidar [J]. Remote Sensing of Environment, 2011, 115(1): 2965-2974.

[91] Mei J, Zhang L Q, Wu S H, et al. 3D tree modeling from incomplete point clouds via optimization and L1-MST[J]. International Journal of Geographical Information Science, 2017, 31(5-6): 999-1021.

[92] Xu H, Gossett N, Chen B Q. Knowledge and heuristic-based modeling of laser-scanned trees[J]. Acm Transactions on Graphics, 2007, 26(4): 19-32.

[93] Shihua L, Leiyu D, Hongshu W, et al. Estimating Leaf Area Density of Individual Trees Using the Point Cloud Segmentation of Terrestrial LiDAR Data and a Voxel-Based Model[J]. Remote Sensing, 2017, 9(11): 1202-1213.