点云语义引导的无人机激光雷达单木分割与参数估算

doi:10.11707/j.1001-7488.LYKX20250136

摘要/Abstract

摘要：

目的: 针对无人机激光雷达（ULS）点云在桉树人工纯林单木分割与参数估算中存在树冠重叠、树干点稀疏和噪声干扰等问题，构建一种基于语义引导的单木分割与参数提取方法，以提高ULS点云的单木分割精度和参数提取准确性。方法: 以广西南宁市高峰林场为研究区，构建完整的点云处理流程。采用深度学习模型对点云进行语义分割，将其划分为树干、树叶和地面等语义类别，同时集成基于密度的空间聚类（DBSCAN）与最近邻分配（KNN）的混合算法，利用语义信息完成单木分割。针对ULS点云在树干处点云稀疏的问题，运用树干曲线拟合方法估算胸径，并采用高度伪波形方法估算树高。在不同结构复杂度的样地中对基于语义引导的单木分割与参数提取方法进行验证，以评估其适用性和精度。结果: 该方法在桉树人工纯林中实现了较高的单木分割精度，总体召回率为0.92，精确率为0.95，平均F分数为0.93。在单木参数提取方面，树高估算的决定系数（R²）为0.98，均方根误差（RMSE）为1.03 m；胸径估算的R²为0.81，RMSE为2.96 cm。结论: 基于语义引导的单木分割与参数提取方法能够较高精度地实现桉树人工纯林ULS点云的单木分割与参数提取，证明语义引导框架在提升ULS点云独立应用能力方面的有效性,可为森林资源监测中ULS数据的高效应用提供参考。

关键词: 单木分割, 单木参数, 激光雷达, 无人机, 点云

Abstract:

Objective: To address crown overlap, sparse stem points, and noise interference in unmanned aerial vehicle laser scanning (ULS) point clouds for individual tree segmentation (ITS) and parameter estimation in eucalyptus plantations, a semantic-guided method for individual tree segmentation and parameter extraction was developed to improve segmentation accuracy and parameter estimation performance. Method: Gaofeng Forest Farm in Nanning, Guangxi was selected as the study area, and a complete point-cloud processing workflow was established. A deep learning model was used to perform semantic segmentation on point clouds, the point clouds were classified into semantic categories, including stems, leaves, and ground. Subsequently, a hybrid algorithm combining density-based spatial clustering of applications with noise (DBSCAN) and K-nearest neighbors (KNN) was used for individual tree segmentation by incorporating semantic information. To counter the sparsity of ULS point clouds at the stem level, stem curve fitting was adopted for diameter at breast height (DBH) estimation, and a height pseudo-waveform method was employed for tree height estimation. The proposed method was validated across plots with varying structural complexities to assess its applicability and accuracy. Result: Experimental results showed that high ITS accuracy was achieved in the eucalyptus plantations, with an overall recall of 0.92, precision of 0.95, and an average F-score of 0.93. For individual tree parameter estimation, tree height estimation showed a coefficient of determination (R²) of 0.98 with a root mean square error (RMSE) of 1.03 m. DBH estimation yielded an R² of 0.81 and an RMSE of 2.96 cm. Conclusion: The proposed method enables accurate individual tree segmentation and parameter extraction from ULS point clouds in eucalyptus plantations, indicating that semantic guidance can improve the applicability of ULS point clouds for individual-tree-level analysis. This study provides a reference for the efficient use of ULS data in forest resource monitoring.

Key words: individual tree segmentation, individual tree parameters, light detectionand ranging (LiDAR), unmanned aerial vehicle (UAV), point cloud

中图分类号:

S758.5

练一宁,卢昊,淮永建,徐海峰,霍朗宁,王智超. 点云语义引导的无人机激光雷达单木分割与参数估算[J]. 林业科学, 2026, 62(4): 106-117.

Yining Lian,Hao Lu,Yongjian Huai,Haifeng Xu,Langning Huo,Zhichao Wang. Point Cloud Semantic-Guided Individual Tree Segmentation and Parameter Estimation Using UAV Laser Scanning[J]. Scientia Silvae Sinicae, 2026, 62(4): 106-117.

图/表 11

表1

图1

图2

图3

图4

图5

表2

图6

图7

图8

表3

参考文献 0

	陈思宇, 刘宪钊, 王懿祥, 等. 基于机载LiDAR的高郁闭度华北落叶松林单木识别. 浙江农林大学学报, 2022, 39 (4): 800- 806. doi: 10.11833/j.issn.2095-0756.20210399
	Chen S Y, Liu X Z, Wang Y X, et al. Individual tree detection in the high canopy density Larix principis-rupprechtii forest based on airborne LiDAR. Journal of Zhejiang A& F University, 2022, 39 (4): 800- 806. doi: 10.11833/j.issn.2095-0756.20210399
	丁志文, 邢艳秋, 尹伯卿, 等. 融合无人机和地基激光雷达点云数据估测单木结构参数. 森林工程, 2024, 39 (1): 142- 151.
	Ding Z W, Xing Y Q, Yin B Q, et al. Fusion of UAV and TLS LiDAR point cloud data for estimating individual tree structure parameters. Forest Engineering, 2024, 39 (1): 142- 151.
	冯仲科, 杜鹏志, 闫宏伟, 等. 创建新一代森林资源调查监测技术体系的实践与探索. 林草资源研究, 2018, (3): 5- 14. doi: 10.13466/j.cnki.lyzygl.2018.03.002
	Feng Z K, Du P Z, Yan H W, et al. Practice and exploration for establishing new generation technology system of forest resources investigation and monitoring. Forest Resources Management, 2018, (3): 5- 14. doi: 10.13466/j.cnki.lyzygl.2018.03.002
	耿　林, 李明泽, 范文义, 等. 基于机载 LiDAR 的单木结构参数及林分有效冠的提取. 林业科学, 2018, 54 (7): 62- 72.
	Geng L, Li M Z, Fan W Y, et al. Individual tree structure parameters and effective crown of the stand extraction base on airborn LiDAR data. Scientia Silvae Sinicae, 2018, 54 (7): 62- 72.
	孔　丹, 庞　勇, 梁晓军, 等. 基于分层叠加的机载LiDAR点云单木分割. 林业科学, 2024, 60 (3): 87- 99. doi: 10.11707/j.1001-7488.LYKX20220303
	Kong D, Pang Y, Liang X J, et al. Individual tree segmentation from ALS point clouds based on layers stacking algorithm. Scientia Silvae Sinicae, 2024, 60 (3): 87- 99. doi: 10.11707/j.1001-7488.LYKX20220303
	李增元, 陈尔学. 中国林业遥感发展历程. 遥感学报, 2021, 25 (1): 292- 301. doi: 10.11834/jrs.20211016
	Li Z Y, Chen E X. Development course of forestry remote sensing in China. National Remote Sensing Bulletin, 2021, 25 (1): 292- 301. doi: 10.11834/jrs.20211016
	李增元, 刘清旺, 庞　勇. 激光雷达森林参数反演研究进展. 遥感学报, 2016, 20 (5): 1138- 1150.
	Li Z Y, Liu Q W, Pang Y. Review on forest parameters inversion using LiDAR. Journal of Remote Sensing, 2016, 20 (5): 1138- 1150.
	刘金成, 郭义军, 曾　嘉, 等. 基于树冠最高点和地面树干中心的森林影像点云配准方法. 农业工程学报, 2024, 40 (15): 127- 134.
	Liu J C, Guo Y J, Zeng J, et al. Forest point cloud registration using the tree top and the ground-level tree center. Transactions of the Chinese Society of Agricultural Engineering, 2024, 40 (15): 127- 134.
	朱泊东, 罗洪斌, 金　京, 等. 高郁闭度人工林无人机激光雷达单木分割方法优化. 林业科学, 2022, 58 (9): 48- 59. doi: 10.11707/j.1001-7488.20220905
	Zhu B D, Luo H B, Jin J, et al. Optimization of individual tree segmentation methods for high canopy density plantation based on UAV LiDAR. Scientia Silvae Sinicae, 2022, 58 (9): 48- 59. doi: 10.11707/j.1001-7488.20220905
	Al-Ali Z M, Abdullah M M, Asadalla N B, et al. 2020. A comparative study of remote sensing classification methods for monitoring and assessing desert vegetation using a UAV-based multispectral sensor. Environmental Monitoring and Assessment, 192 (6): 389. doi: 10.1007/s10661-020-08330-1
	Brede B, Lau A, Bartholomeus H, et al. Comparing RIEGL RiCOPTER UAV LiDAR derived canopy height and DBH with terrestrial LiDAR. Sensors, 2017, 17 (10): 2371. doi: 10.3390/s17102371
	Calders K, Newnham G, Burt A, et al. Nondestructive estimates of above-ground biomass using terrestrial laser scanning. Methods in Ecology and Evolution, 2015, 6 (2): 198- 208. doi: 10.1111/2041-210X.12301
	Cao L, Liu H, Fu X Y, et al. Comparison of UAV LiDAR and digital aerial photogrammetry point clouds for estimating forest structural attributes in subtropical planted forests. Forests, 2019, 10 (2): 145. doi: 10.3390/f10020145
	Cao Y J, Ball J G C, Coomes D A. et al. Benchmarking airborne laser scanning tree segmentation algorithms in broadleaf forests shows high accuracy only for canopy trees. International Journal of Applied Earth Observation and Geoinformation, 2023, 123, 103490. doi: 10.1016/j.jag.2023.103490
	Chen Q, Baldocchi D, Gong P, et al. Isolating individual trees in a savanna woodland using small footprint LIDAR data. Photogrammetric Engineering & Remote Sensing, 2006, 72 (8): 923- 932. doi: 10.14358/PERS.72.8.923
	Chen Q D, Gao T, Zhu J J, et al. Individual tree segmentation and tree height estimation using leaf-off and leaf-on UAV-LiDAR data in dense deciduous forests. Remote Sensing, 2022, 14 (12): 2787. doi: 10.3390/rs14122787
	Dai W X, Yang B S, Dong Z, et al. A new method for 3D individual tree extraction using multispectral airborne LiDAR point clouds. ISPRS Journal of Photogrammetry and Remote Sensing, 2018, 144 (10): 400- 411. doi: 10.1016/j.isprsjprs.2018.08.010
	Erdody T L, Moskal L M. Fusion of LiDAR and imagery for estimating forest canopy fuels. Remote Sensing of Environment, 2010, 114 (4): 725- 737. doi: 10.1016/j.rse.2009.11.002
	Fu Y W, Niu Y F, Wang L, et al. Individual-tree segmentation from UAV-LiDAR data using a region-growing segmentation and supervoxel-weighted fuzzy clustering approach. Remote Sensing, 2024, 16 (4): 608. doi: 10.3390/rs16040608
	Henrich J, Delden J V. 2024,. Towards general deep-learning-based tree instance segmentation models. ICLR 2024 Machine Learning for Remote Sensing Workshop. doi: 10.48550/arXiv.2405.02061
	Hu B X, Li J L, Jing L H, et al. 2014,. Improving the efficiency and accuracy of individual tree crown delineation from high-density LiDAR data. International Journal of Applied Earth Observation and Geoinformation, 26: 145- 155. doi: 10.1016/j.jag.2013.06.003
	Hyypä E, Kukko A, Kaartinen H, et al. Direct and automatic measurements of stem curve and volume using a high-resolution airborne laser scanning system. Science of Remote Sensing, 2022, 5, 100050. doi: 10.1016/j.srs.2022.100050
	Hyypä J, Yu X W, Hyypä H, et al. Advances in forest inventory using airborne laser scanning. Remote Sensing, 2012, 4 (5): 1190- 1207. doi: 10.3390/rs4051190
	Jaakkola A, Hyyppä J, Yu X W, et al. Autonomous collection of forest field reference: the outlook and a first step with UAV laser scanning. Remote Sensing, 2017, 9 (8): 785. doi: 10.3390/rs9080785
	Jarahizadeh S, Salehi B. Advancing tree detection in forest environments: a deep learning object detector approach with UAV LiDAR data. Urban Forestry & Urban Greening, 2025, 105, 128695. doi: 10.1016/j.ufug.2025.128695
	Liang X L, Kankare V, Hyyppä J, et al. Terrestrial laser scanning in forest inventories. ISPRS Journal of Photogrammetry and Remote Sensing, 2016, 115, 63- 77. doi: 10.1016/j.isprsjprs.2016.01.006
	Lin Y Y, Li H, Jing L H, et al. Individual tree crown delineation using airborne LiDAR data and aerial imagery in the Taiga-tundra ecotone. Remote Sensing, 2024, 16 (21): 3920. doi: 10.3390/rs16213920
	Lu H, Li B W, Yang G, et al. Towards a point cloud understanding framework for forest scene semantic segmentation across forest types and sensor platforms. Remote Sensing of Environment, 2025, 318, 114591. doi: 10.1016/j.rse.2024.114591
	Muss J D, Mladenoff D J, Townsend P A. A pseudo-waveform technique to assess forest structure using discrete LiDAR data. Remote Sensing of Environment, 2011, 115 (3): 824- 835. doi: 10.1016/j.rse.2010.11.008
	Ni W J, Ranson K J, Zhang Z Y, et al. Features of point clouds synthesized from multi-view ALOS/PRISM data and comparisons with LiDAR data in forested areas. Remote Sensing of Environment, 2014, 149, 47- 57. doi: 10.1016/j.rse.2014.04.001
	Pu Y H, Xu D D, Wang H B, et al. A new strategy for individual tree detection and segmentation from leaf-on and leaf-off UAV-LiDAR point clouds based on automatic detection of seed points. Remote Sensing, 2023, 15 (6): 1619. doi: 10.3390/rs15061619
	Qin H M, Zhou W Q, Yao Y, et al. Individual tree segmentation and tree species classification in subtropical broadleaf forests using UAV-based LiDAR, hyperspectral, and ultrahigh-resolution RGB data. Remote Sensing of Environment, 2022a, 280, 113143. doi: 10.1016/j.rse.2022.113143
	Wieser M, Mandlburger G, Hollaus M, et al. A case study of UAS borne laser scanning for measurement of tree stem diameter. Remote Sensing, 2017, 9 (11): 1154. doi: 10.3390/rs9111154
	Wielgosz M, Puliti S, Xiang B B, et al. SegmentAnyTree: a sensor and platform agnostic deep learning model for tree segmentation using laser scanning data. Remote Sensing of Environment, 2024, 313, 114367. doi: 10.1016/j.rse.2024.114367
	Wu X Q, Shen X, Cao L, et al. Assessment of individual tree detection and canopy cover estimation using unmanned aerial vehicle based light detection and ranging (UAV-LiDAR) data in planted forests. Remote Sensing, 2019, 11 (8): 908. doi: 10.3390/rs11080908
	Xiang B B, Wielgosz M, Kontogianni T, et al. Automated forest inventory: analysis of high-density airborne LiDAR point clouds with 3D deep learning. Remote Sensing of Environment, 2024, 305, 114078. doi: 10.1016/j.rse.2024.114078
	Zhang Z N, Wang T J, Skidmore A K, et al. An improved area-based approach for estimating plot-level tree DBH from airborne LiDAR data. Forest Ecosystems, 2023, 10, 100089. doi: 10.1016/j.fecs.2023.100089

样地 Plot	地形 Terrain	树种 Tree species	林分类型 Stand type	株数密度 Density/ (tree?hm^?2)	点云密度 Point cloud density/ (pts·m^?2)	平均胸径 Average DBH/cm	平均树高 Average tree height/m	平均冠幅（南北） Average crown width (N-S)/m
1	缓坡 Gentle slope	桉树 Eucalyptus spp.	阔叶林 Broad-leaved forest	650	1 242	25.1	33.7	3.2
2	陡坡 Steep slope	桉树 Eucalyptus spp.	阔叶林 Broad-leaved forest	2 100	1 040	9.79	12.1	2.2
3	陡坡 Steep slope	桉树 Eucalyptus spp.	阔叶林 Broad-leaved forest	2 867	756	11.6	15.8	2.8

样地 Plot	语义引导的单木分割 Semantic-guided individual tree segmentation			标记控制的分水岭单木分割 Marker-controlled watershed individual tree segmentation
样地 Plot	召回率 Recall (Re)	精确率 Precision (Pr)	F分数 F-score (F)	召回率 Recall (Re)	精确率 Precision (Pr)	F分数 F-score (F)
1	0.97	1.00	0.98	0.67	0.15	0.25
2	0.92	0.95	0.94	1.00	0.24	0.39
3	0.86	0.90	0.87	0.91	0.18	0.30

组Group	RMSE/cm	RMSE%	Bias/cm	Bias%
无人机?调查 ULS-survey	2.96	20.27	?0.86	?5.88
地基?调查 TLS-survey	0.90	6.17	0.06	0.41
无人机?地基 ULS-TLS	3.20	22.00	?0.92	?6.31

[1]	曾伟生,温雪香,付涵,孙乡楠,吕康梅,刘樯漪,王甜. 基于激光雷达的北京市主要林分因子联立模型构建与应用[J]. 林业科学, 2026, 62(4): 68-80.
[2]	范光鹏,许亮亮,蔡会德,徐占勇,孟想,邵亚奎,卢峰. 基于地基激光雷达点云的桉树单木参数提取与地上碳储量测定[J]. 林业科学, 2026, 62(3): 88-99.
[3]	王天璨,格根塔娜,李晓松,月亮高可,沈通,陈超超,智育博,赵立成,姬翠翠. 基于Sentinel-2卫星影像与梯度提升树回归模型的疏林郁闭度精准监测——以内蒙古退耕还林工程为例[J]. 林业科学, 2026, 62(2): 173-185.
[4]	陈洪立,王冠凯,张文福,张建,黄宏亮,杨振华,赵胤,杜小强. 基于无人机的竹产品与疫木下山吊运系统设计与试验[J]. 林业科学, 2026, 62(1): 177-187.
[5]	张宇娇,赵恒谦,付含聪,刘哿,皇甫霞丹,刘轩绮. 基于无人机遥感多特征的单木地上生物量反演模型[J]. 林业科学, 2025, 61(8): 129-141.
[6]	熊晓燕,李彩霞,柴国奇,陈龙,贾翔,雷令婷,张晓丽. 联合UAV-LiDAR和GEDI数据的区域森林地上生物量估算[J]. 林业科学, 2025, 61(8): 142-153.
[7]	杨晨辉,程寿民,高谢雨,董利虎,郝元朔. 基于干形拟合的TLS长白落叶松树干参数提取[J]. 林业科学, 2025, 61(8): 154-163.
[8]	吴家敏,王亚欣,孙斌,马志杰,孙维娜,洪亮. 基于无人机的柠条锦鸡儿生物量遥感估测[J]. 林业科学, 2025, 61(6): 13-24.
[9]	刘鲁霞,胡波,桑国庆,刘玉玉. 激光雷达森林结构指标在森林植物多样性评估中的研究进展[J]. 林业科学, 2025, 61(1): 176-196.
[10]	谢婉滢,刘文萍,王晗. 无人机松林图像早期松材线虫病害检测[J]. 林业科学, 2024, 60(9): 124-133.
[11]	胡中洋,陕亮,陈翔宇,余坤勇,刘健. CHM与DSM相结合的无人机激光雷达单木分割[J]. 林业科学, 2024, 60(8): 14-24.
[12]	张宇,张怀清,安锋,蒋玲,云挺. 基于计算机模拟模型的林木冠层太阳短波辐射定量分析方法[J]. 林业科学, 2024, 60(4): 16-30.
[13]	艾萨迪拉·玉苏甫,玉米提·哈力克,巴比尔江·迪力夏提,雷诚,魏建新,阿不都拉·阿不力孜,崔健泺,何熙祥. 基于激光雷达数据的塔里木河下游胡杨种群空间分布格局和种内竞争关系[J]. 林业科学, 2024, 60(4): 31-39.
[14]	孔丹,庞勇,梁晓军,杜黎明,白羽. 基于分层叠加的机载LiDAR点云单木分割[J]. 林业科学, 2024, 60(3): 87-99.
[15]	周相贝, 李春干, 代华兵, 余铸, 李振, 苏凯. 点云密度对机载激光雷达大区域亚热带森林参数估测精度的影响[J]. 林业科学, 2023, 59(9): 23-33.