English

应用科学学报 ›› 2026, Vol. 44 ›› Issue (4): 644-656.doi: 10.3969/j.issn.0255-8297.2026.04.009

• 智能视觉感知 • 上一篇    下一篇

森林点云的三维场景图表达与LLM应用

彭锦鸿1, 陈茂霖1,2,3, 薛梅3, 李汝峰4   

  1. 1. 重庆交通大学 智慧城市学院, 重庆 400074;
    2. 自然资源部地理国情监测重点实验室, 重庆 400074;
    3. 重庆市测绘科学技术研究院, 重庆 400021;
    4. 北部战区陆军参谋部第30分队, 山东 济南 250000
  • 收稿日期:2026-01-07 发布日期:2026-08-01
  • 通信作者: 陈茂霖,副教授,研究方向为三维激光点云处理及其行业应用领域。E-mail:maolinchen@cqjtu.edu.cn E-mail:maolinchen@cqjtu.edu.cn
  • 基金资助:
    国家自然科学基金(No.42301459);重庆市技术创新与应用发展专项面上项目(No.CSTB2024TIADGPX0057);重庆市自然科学基金(No.CSTB2024NSCQ-MSX0652);自然资源部地理国情监测重点实验室开放基金(No.2025NGCM04)

Three-Dimensional Scene Graph Representation of Forest Point Clouds and LLM Applications

PENG Jinhong1, CHEN Maolin1,2,3, XUE Mei3, LI Rufeng4   

  1. 1. School of Smart City, Chongqing Jiaotong University, Chongqing 400074, China;
    2. Key Laboratory of National Geographic Condition Monitoring, Ministry of Natural Resources, Chongqing 400074, China;
    3. Chongqing Academy of Surveying and Mapping, Chongqing 400021, China;
    4. 30th Unit of the General Staff Department of the Northern Theater Command Army, Jinan 250000, Shandong, China
  • Received:2026-01-07 Published:2026-08-01

PDF

4

摘要： 三维激光扫描技术已成为林业调查中获取高精度林分参数的重要手段，但基于海量激光点云的森林三维场景表达依赖专业软件处理，且主要关注对象级语义理解，缺少要素关系信息的显示表达。针对于此，提出了一种面向森林场景的三维场景图概念模型，系统介绍了森林场景的要素分层分类、几何描述、语义表达及关系描述方法。在此基础上，将三维场景图与语义关联分析相结合，引入大语言模型作为查询与分析工具，设计了三级通用评估框架，对比评估了ChatGPT-4o、DeepSeek-R1和Grok4的运行速度、输入限制和可视化效果等方面，探讨了该模型的效能、潜力和局限性。基于公开点云数据集ForestSemantic的实验结果表明：三维场景图在森林场景具有良好的数据承载能力，能够有效组织和联系森林关系；Grok4的准确率在90%以上，优于其余两种大语言模型。本研究可为林业资源的空间表达与分析、经营管理与决策制定提供辅助。

关键词: 激光雷达, 林业调查, 三维场景图, 大语言模型, 森林空间关系, 林业应用

Abstract: 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.

Key words: light detection and ranging, forestry survey, three-dimensional scene graph, large language model, forest spatial relation, forestry application

中图分类号: