Three-dimensional laser scanning technology has become a crucial method for obtaining high-precision forest stand parameters in forestry surveys. However, the expression of forest three-dimensional scenes based on massive laser point clouds relies on professional software for processing, and it primarily focuses on object-level semantic understanding, lacking explicit expression of element relationship information. To address this issue, a three-dimensional scene graph conceptual model oriented to forest scenes was proposed, and the methods of element hierarchical classification, geometric description,semantic expression, and relationship description of forest scenes were systematically introduced. On this basis, the three-dimensional scene graph was combined with semantic association analysis, and a large language model was introduced as a query and analysis tool. A three-level universal evaluation framework was designed to comparatively evaluate ChatGPT-4o, DeepSeek-R1, and Grok4 in terms of operating speed, input limitations, and visual effects, and the efficacy, potential, and limitations of this model were discussed. Experimental results based on the public point cloud dataset ForestSemantic show that the three-dimensional scene graph has a good data-carrying capacity in forest scenes and can effectively organize and connect forest relationships; the accuracy of Grok4 is above 90%,which is superior to the other two large language models. This study can provide assistance for the spatial expression and analysis, operation management, and decision-making of forestry resources.
[1] 自然资源部办公厅.自然资源部办公厅关于印发《自然资源三维立体时空数据库建设总体方案》的通知:自然资办发[2021] 21号[EB/OL].(2021-02-08)[2026-01-07].https://www.gov.cn/zhengce/zhengceku/2021-02/11/content_5586814.htm.
[2] Lin Y, Wiegand K. Towards 3D tree spatial pattern analysis:setting the cornerstone of LiDAR advancing 3D forest structural and spatial ecology[J]. International Journal of Applied Earth Observation and Geoinformation, 2021, 103:102506.
[3] Zhang W M, Qi J B, Wan P, et al. An easy-to-use airborne LiDAR data filtering method based on cloth simulation[J]. Remote Sensing, 2016, 8(6):501.
[4] Axelsson P. DEM generation from laser scanner data using adaptive TIN models[J]. International Archives of Photogrammetry and Remote Sensing, 2000, 33(4):110-117.
[5] Hui Z, Hu Y, Yevenyo Y Z, et al. An improved morphological algorithm for filtering airborne LiDAR point cloud based on multi-level kriging interpolation[J]. Remote Sensing, 2016, 8(1):35.
[6] 杜诗韵,陈茂霖,潘建平,等.利用相对密度迭代分析的地基点云过滤方法[J].激光与红外, 2024,54(9):1358-1365.Du S Y, Chen M L, Pan J P, et al. A filtering method for terrestrial point clouds using relative density iterative analysis[J]. Laser&Infrared, 2024, 54(9):1358-1365.(in Chinese)
[7] Liu Y, Chen D, Fu S, et al. Segmentation of individual tree points by combining markercontrolled watershed segmentation and spectral clustering optimization[J]. Remote Sensing,2024, 16(4):610.
[8] Wang D. Unsupervised semantic and instance segmentation of forest point clouds[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 165:86-97.
[9] Xi Z, Hopkinson C. 3D graph-based individual-tree isolation(Treeiso)from terrestrial laser scanning point clouds[J]. Remote Sensing, 2022, 14(23):6116.
[10] 姬翠翠,月亮高可,李晓松,等.基于GEDI、Sentinel-2和机载激光雷达的森林冠层高度反演方法[J].应用科学学报, 2025, 43(4):694-708.Ji C C, Yueliang G K, Li X S, et al. Forest canopy height inversion method based on GEDI,Sentinel-2 and airborne LiDAR[J]. Journal of Applied Sciences, 2025, 43(4):694-708.(in Chinese)
[11] Kukkonen M, Maltamo M, Korhonen L, et al. Fusion of crown and trunk detections from airborne UAS based laser scanning for small area forest inventories[J]. International Journal of Applied Earth Observation and Geoinformation, 2021, 100:102327.
[12] Fan G, Lu F, Cai H, et al. A new method for reconstructing tree-level aboveground carbon stocks of eucalyptus based on TLS point clouds[J]. Remote Sensing, 2023, 15(19):4782.
[13] Wang H, Li D, Duan J, et al. ALS-based, automated, single-tree 3D reconstruction and parameter extraction modeling[J]. Forests, 2024, 15(10):1776.
[14] Delagrange S, Rochon P. Reconstruction and analysis of a deciduous sapling using digital photographs or terrestrial-LiDAR technology[J]. Annals of Botany, 2011, 108(6):991-1000.
[15] Xie D, Wang X, Qi J, et al. Reconstruction of single tree with leaves based on terrestrial LiDAR point cloud data[J]. Remote Sensing, 2018, 10(5):686.
[16] 万里红,曹振宇,田志林,等.一种基于地面激光雷达点云的树木三维建模方法[J].自然资源遥感,2025, 37(1):62-67.Wan L H, Cao Z Y, Tian Z L, et al. A 3D modeling method for trees based on terrestrial laser scanning point clouds[J]. Remote Sensing for Natural Resources, 2025, 37(1):62-67(in Chinese).
[17] Kok E, Wang X, Chen C. Obscured tree branches segmentation and 3D reconstruction using deep learning and geometrical constraints[J]. Computers and Electronics in Agriculture, 2023,210:107884.
[18] Liu Z, Wu K, Guo J, et al. Single image tree reconstruction via adversarial network[J]. Graphical Models, 2021, 117:101115.
[19] Tian G, Chen C, Huang H. Comparative analysis of novel view synthesis and photogrammetry for 3D forest stand reconstruction and extraction of individual tree parameters[J]. Remote Sensing, 2025, 17(9):1520.
[20] Ling S, Li J, Ding L, et al. Multi-view jujube tree trunks stereo reconstruction based on UAV remote sensing imaging acquisition system[J]. Applied Sciences, 2024, 14(4):1364.
[21] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[J]. Advances in Neural Information Processing systems, 2017:30.
[22] Ying S, Li Z, Yu M. Beyond words:evaluating large language models in transportation planning[J]. Geo-spatial Information Science, 2025:1-23.
[23] Chen Y, Zhang S, Han T, et al. Chat3D:interactive understanding 3D scene-level point clouds by chatting with foundation model for urban ecological construction[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2024, 212:181-192.
[24] Li Z, Ning H. Autonomous GIS:the next-generation AI-powered GIS[J]. International Journal of Digital Earth, 2023, 16(2):4668-4686.
[25] Xiao Z, Ma J. LLM agent framework for intelligent change analysis in urban environment using remote sensing imagery[J]. Automation in Construction, 2025, 177:106341.
[26] Liang X, Qi H, Deng X, et al. ForestSemantic:a dataset for semantic learning of forest from close-range sensing[J]. Geo-spatial Information Science, 2024:1-27.
