林草行业大模型构建思路与应用前景

doi:10.11707/j.1001-7488.LYKX20250188

摘要/Abstract

摘要：

目的: 大模型作为发展新质生产力的重要引擎，对传统行业的转型升级以及基础科学研究起到重要支撑作用。林草行业具有地域广阔、类型复杂和工作难度大等特点，当前林草专业模型在通用性、适应性、复杂问题处理和协同决策等方面存在不足，难以满足行业需求，制约林草行业高质量发展，林草行业亟需以行业大模型为代表的人工智能（AI）技术进行深度融合，实现创新赋能。本研究旨在探索构建林草行业大模型的有效路径，突破行业发展瓶颈，全面提升林草行业新质生产力，推动林草行业智能化升级。方法: 阐述大模型的发展现状，提出林草行业大模型的建设思路，构建涵盖基础设施层、数据资源层、模型构建层、应用服务层的林草行业大模型框架；基于林草行业应用场景特点，设计林草大语言模型、林草视觉大模型、林草时空大模型和林草多模态大模型4类林草大模型协同的林草行业大模型技术体系。面向智慧林草的发展需求和趋势，针对林草AI智能体的应用前景，深入探讨林草行业大模型在林草资源监测、林草育种与培育、林草经营管理、林草生态系统管理、林草资源保护、林草资源利用、林草生态工程、林草行业管理8大林草行业典型场景中的应用，提出林草行业大模型与林草实景三维重建、林草数字孪生、林草具身智能、林草元宇宙等前沿技术的创新融合发展思路。结果: 通过构建林草行业大模型与专业模型协同的林草AI智能体，强化不同业务场景下大小模型的协同决策效能，提升林草行业复杂问题的协同分析处理水平，为智慧林草提供专业化、智能化和精准化解决方案。结论: 林草行业大模型是智慧林草建设的核心驱动力和林草信息化发展的智能引擎，不仅能变革林草科学研究范式，重塑行业管理体系，更能为林草新质生产力的发展注入核心动力，为智慧林草全行业垂直领域的科研、应用和产业发展提供有力支撑，全面推动林草智能化转型与AI产业创新应用，赋能林草行业高质量发展。

关键词: 人工智能, 大模型, 林草行业大模型, 智慧林草, 新质生产力

Abstract:

Objective: As an important engine for developing new quality productivity, large models play a crucial role in supporting the transformation and upgrading of traditional industries and basic scientific research. The forestry and grassland industry is characterized by vast geographical coverage, complex types, and high work difficulty. However, current professional models in this industry suffer from insufficient generality, adaptability, complex problem-solving capabilities, and collaborative decision-making abilities, making them unable to meet the industry's demands and restricting the high-quality development of the forestry and grassland industry. Therefore, it is imperative for the forest and grassland industry to deeply integrate artificial intelligence (AI) technologies represented by industry large models with forestry and grassland business operations to achieve innovative empowerment. This study aims to explore effective paths for constructing large models in the forestry and grassland industry, break through development bottlenecks, comprehensively enhance the informatization management level of the forestry and grassland industry, and promote the intelligent upgrade of the forestry and grassland industry. Method: This paper elaborates on the current development status of large models, proposes the construction ideas of large models in the forestry and grassland industry, and constructs a framework for large models in the forestry and grassland industry covering infrastructure layer, data resource layer, model construction layer, and application service layer. Based on the characteristics of application scenarios in the forestry and grassland industry, it designs a technical system for large models in the forestry and grassland industry, including four types of large models: forestry and grassland large language models, forestry and grassland visual large models, forestry and grassland spatiotemporal large models, and forestry and grassland multimodal large models. Focusing on the development needs and trends of smart forestry and grassland and the application prospects of forestry and grassland intelligent agents, it deeply explores the application of large models in the forestry and grassland industry in eight typical scenarios, including forestry and grassland resource monitoring, forestry and grassland breeding and cultivation, forestry and grassland operation and management, forestry and grassland ecosystem management, forestry and grassland resource protection, forestry and grassland resource utilization, forestry and grassland ecological engineering, and forestry and grassland industry management. It also proposes innovative and integrated development ideas for large models in the forestry and grassland industry with cutting-edge technologies such as forestry and grassland real-scene 3D reconstruction, forestry and grassland digital twins, forestry and grassland embodied intelligence and forestry and grassland metaverse. Result: By constructing a forestry and grassland AI intelligent agent that integrates large models and professional models, the collaborative decision-making efficiency of large and small models in different business scenarios is strengthened, and the collaborative analysis and processing level of complex problems in the forestry and grassland industry is improved, providing professional, intelligent, and precise solutions covering the entire business chain for smart forestry and grassland. Conclusion: Large models in the forestry and grassland industry are the core driving force for the construction of smart forestry and grassland and the intelligent engine for the development of forestry and grassland informatization. They can not only transform the research paradigm of forestry and grassland science but also reshape the industry management system. Moreover, they can inject core power into the development of new quality productivity in the forestry and grassland industry, provide strong support for scientific research, application, and industrial development in vertical fields of the entire smart forestry and grassland industry, and comprehensively promote the intelligent transformation of the forestry and grassland industry and the innovative application of AI industries, empowering the high-quality development of the forestry and grassland industry.

Key words: artificial intelligence, large models, forestry-grassland industry large models, smart forestry-grassland, new quality productive forces

中图分类号:

S757
TP399

谭晶维,张怀清,郭梦蕾,朱学岩,刘洋,杨廷栋. 林草行业大模型构建思路与应用前景[J]. 林业科学, 2025, 61(7): 170-181.

Jingwei Tan,Huaiqing Zhang,Menglei Guo,Xueyan Zhu,Yang Liu,Tingdong Yang. Construction Ideas and Application Prospects of Large Models in Forestry and Grassand Industry[J]. Scientia Silvae Sinicae, 2025, 61(7): 170-181.

图/表 7

图1

图2

图3

图4

图5

图6

图7

参考文献 0

	葛晓宁, 许新桥, 张怀清, 等. 2025. 林木基因型-环境互作算法研究进展与思考. 林业科学, 61(3): 1−15.
	Ge X N, Xu X Q, Zhang H Q, et al. 2025. Progress and reflection on genotype-environment interaction algorithms in forest tree breeding. Scientia Silvae Sinicae, 61(3): 1−15. [in Chinese]
	何　江, 梁　正, 韩希佳. 2024. AI产业化与产业AI化:AI大模型产业生态的行动实践、变革趋势与前沿议题. 西安财经大学学报, 37(6): 49−63.
	He J, Liang Z, Han J X. 2024. AI industrialization and industrial AI-ization: action practices, transformation trends,and frontiertopics for the AI large model industrial ecosystem. Journal of Xi’an University of Finance and Economics, 37(6): 49−63. [in Chinese]
	黄再胜. AI大模型赋能新质生产力加快发展: 内在机理、现实障碍与实践进路. 改革与战略, 2024, 40 (2): 1- 12.
	Huang Z S. Empowering the accelerated development of new-quality productivity by AI large models: internal mechanisms, realistic obstacles, and practical approaches. Reformation & Strategy, 2024, 40 (2): 1- 12.
	刘海军, 温赞玲. 2025. 深度求索DeepSeek:人工智能、技术创新与新质生产力. 当代经济管理, 47(6): 1−11.
	Liu H J, Wen Z L. 2025. DeepSeek: artificial intelligence, technological innovation and new quality productive forces. Contemporary Economic Management, 47(6): 1−11. [in Chinese]
	舒全英, 马　媛, 陈　亮, 等. 数字孪生水利建设中的人工智能大模型应用探索. 中国水利, 2025, (6): 14- 30. doi: 10.3969/j.issn.1000-1123.2025.06.002
	Shu Q Y, Ma Y, Chen L, et al. Exploration of artificial intelligence large model applications in digital twin water conservancy construction. China Water Resources, 2025, (6): 14- 30. doi: 10.3969/j.issn.1000-1123.2025.06.002
	谭晶维, 张怀清, 刘　洋, 等. 问答式林业预训练语言模型ForestBERT. 林业科学, 2024, 60 (9): 99- 110. doi: 10.11707/j.1001-7488.LYKX20240435
	Tan J W, Zhang H Q, Liu Y, et al. Question-answering forestry pre-trained language model: ForestBERT. Scientia Silvae Sinicae, 2024, 60 (9): 99- 110. doi: 10.11707/j.1001-7488.LYKX20240435
	燕　琴, 顾海燕, 杨　懿, 等. 2024. 智能遥感大模型研究进展与发展方向. 测绘学报, 53(10): 1967−1980.
	Yan Q, Gu H Y, Yang Y, et al. 2024. Research progress and trend of intelligent remote sensing large model. Acta Geodaetica et Cartographica Sinica, 53(10): 1967−1980. [in Chinese]
	朱教君, 王高峰, 张怀清, 等. 关于“气候智慧林业”研究的思考. 林业科学, 2024, 60 (7): 1- 7. doi: 10.11707/j.1001-7488.LYKX20240433
	Zhu J J, Wang G F, Zhang H Q, et al. On the research of climate-smart forestry. Scientia Silvae Sinicae, 2024, 60 (7): 1- 7. doi: 10.11707/j.1001-7488.LYKX20240433
	Awais M, Naseer M, Khan S, et al. 2025. Foundation models defining a new era in vision: A survey and outlook. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(4): 2245−2264.
	Benegas G, Ye C, Albors C, et al. 2025. Genomic language models: Opportunities and challenges. Trends in Genetics, 41(4): 286−302.
	Bi K, Xie L, Zhang H, et al. Accurate medium-range global weather forecasting with 3d neural networks. Nature, 2023, 619 (7970): 533- 538. doi: 10.1038/s41586-023-06185-3
	Bodnar C, Bruinsma W P, Lucic A, et al. 2025. A foundation model for the earth system. Nature, 641(8065): 1180−1187.
	Borowiec M L, Dikow R B, Frandsen P B, et al. Deep learning as a tool for ecology and evolution. Methods in Ecology and Evolution, 2022, 13 (8): 1640- 1660. doi: 10.1111/2041-210X.13901
	Brixi G, Durrant M G, Ku J, et al. 2025. Genome modeling and design across all domains of life with Evo 2. bioRxiv, 2025.2002. 2018.638918.
	Chen L, Zhong X, Li H, et al. A machine learning model that outperforms conventional global subseasonal forecast models. Nature Communications, 2024, 15 (1): 6425. doi: 10.1038/s41467-024-50714-1
	Cheng G, Chen X, Wang C, et al. Visual fire detection using deep learning: A survey. Neurocomputing, 2024b, 596, 127975. doi: 10.1016/j.neucom.2024.127975
	Cheng Y H, Zhang C Y, Zhang Z W, et al. 2024a. Exploring large language model based intelligent agents: definitions, methods, and prospects. arXiv Preprint arXiv: 2401.03428.
	Curtis P S, Gough C M. Forest aging, disturbance and the carbon cycle. New Phytologist, 2018, 219 (4): 1188- 1193. doi: 10.1111/nph.15227
	Du S Q, Tang S J, Wang W X, et al. 2023. Tree-GPT: modular large language model expert system for forest remote sensing image understanding and interactive analysis. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XLVIII-1/W2-2023: 1729−1736.
	Gan L, Chu W, Li G, et al. Large models for intelligent transportation systems and autonomous vehicles: a survey. Advanced Engineering Informatics, 2024, 62, 102786. doi: 10.1016/j.aei.2024.102786
	Goodge A, Ng W S, Hooi B, et al. 2025. Spatio-temporal foundation models: Vision, challenges, and opportunities. arXiv Preprint arXiv: 2501.09045.
	Gu J, Stevens S, Campolongo E G, et al. 2025. BioCLIP 2: Emergent properties from scaling hierarchical contrastive learning. arXiv Preprint arXiv:2505.23883.
	Guo X, Lao J, Dang B, et al. 2023. SkySense: A multi-modal remote sensing foundation model towards universal interpretation for earth observation imagery. arXiv Preprint arXiv: 2312.10115.
	Harris N L, Gibbs D A, Baccini A, et al. Global maps of twenty-first century forest carbon fluxes. Nature Climate Change, 2021, 11 (3): 234- 240. doi: 10.1038/s41558-020-00976-6
	Hong D, Zhang B, Li X, et al. 2023. SpectralGPT: Spectral remote sensing foundation model. arXiv Preprint arXiv:2311.07113.
	Jiang S X, Liang J F, Wang J Y, et al. 2024. From specific-MLLMs to omni-MLLMs: a survey on MLLMs aligned with multi-modalities. arXiv Preprint arXiv: 2412.11694.
	Latif S, Shoukat M, Shamshad F, et al. 2023. Sparks of large audio models: A survey and outlook. arXiv Preprint arXiv: 2308.12792.
	Lei K X, Zhang H Q, Qiu H Q, et al. A two-dimensional four-quadrant assessment method to explore the spatiotemporal coupling and coordination relationship of human activities and ecological environment. Journal of Environmental Management, 2024, 370, 122362. doi: 10.1016/j.jenvman.2024.122362
	Li Z, Xia L, Tang J, et al. 2024. UrbanGPT: Spatio-temporal large language models. Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Barcelona, Spain, 5351–5362.
	Lian G, Zhang H, Lei K, et al. 2025. A novel collaborative planning framework for artificial forest harvesting and replanting. Biosystems Engineering, 256: 104156.
	Marcus G. 2018. Deep learning: A critical appraisal. arXiv Preprint arXiv: 1801.00631.
	Mendoza-Revilla J, Trop E, Gonzalez L, et al. A foundational large language model for edible plant genomes. Communications Biology, 2024, 7 (1): 835. doi: 10.1038/s42003-024-06465-2
	Qiu H, Zhang H, Kexin L, et al. 2023. Forest digital twin: A new tool for forest management practices based on spatio-temporal data, 3d simulation engine, andintelligent interactive environment. Computers and Electronics in Agriculture, 215: 108416.
	Qiu H Q, Zhang H Q, Lei K X, et al. A novel method for forest spatial structure heterogeneity evaluation of plantation utilizing point-wise vector network and neighborhood index. Computers and Electronics in Agriculture, 2025, 229, 109774. doi: 10.1016/j.compag.2024.109774
	Seidl R, Thom D, Kautz M, et al. Forest disturbances under climate change. Nature Climate Change, 2017, 7 (6): 395- 402. doi: 10.1038/nclimate3303
	Shanahan M. Talking about large language models. Communications of the ACM, 2024, 67 (2): 68- 79. doi: 10.1145/3624724
	Shao M, Basit A, Karri R, et al. Survey of different large language model architectures: trends, benchmarks, and challenges. IEEE Access, 2024, 12, 188664- 188706. doi: 10.1109/ACCESS.2024.3482107
	Simon E, Swanson K, Zou J. Language models for biological research: A primer. Nature Methods, 2024, 21 (8): 1422- 1429. doi: 10.1038/s41592-024-02354-y
	Sun X, Wang P, Lu W, et al. RingMo: A remote sensing foundation model with masked image modeling. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61, 1- 22.
	Tan J, Zhang H, Yang J, et al. ForestryBERT: A pre-trained language model with continual learning adapted to changing forestry text. Knowledge-Based Systems, 2025, 320, 113706. doi: 10.1016/j.knosys.2025.113706
	Tang Y, Bi J, Xu S, et al. 2023. Video understanding with large language models: A survey. arXiv Preprint arXiv: 2312.17432.
	Thirunavukarasu A J, Ting D S J, Elangovan K, et al. Large language models in medicine. Nature Medicine, 2023, 29 (8): 1930- 1940. doi: 10.1038/s41591-023-02448-8
	Wang J L, Zhang D. Intelligent pest forecasting with meteorological data: an explainable deep learning approach. Expert Systems with Applications, 2024, 252, 124137. doi: 10.1016/j.eswa.2024.124137
	Wu Z, Zhang C, Gu X, et al. Deep learning enables satellite-based monitoring of large populations of terrestrial mammals across heterogeneous landscape. Nature Communications, 2023, 14 (1): 3072. doi: 10.1038/s41467-023-38901-y
	Xing Z, Feng Q, Chen H, et al. 2024. A survey on video diffusion models. ACM Computing Surveys, 57(2): 1−42.
	Yang J, Jin H, Tang R, et al. 2024. Harnessing the power of LLMs in practice: A survey on ChatGPT and beyond. ACM Transactions on Knowledge Discovery from Data, 18(6): 1−32.
	Yuan Y, Ding J, Feng J, et al. 2024. UniST: A prompt-empowered universal model for urban spatio-temporal prediction. arXiv Preprint arXiv: 2402.11838.
	Yun T, Li J, Ma L F, et al. Status, advancements and prospects of deep learning methods applied in forest studies. International Journal of Applied Earth Observation and Geoinformation, 2024, 131, 103938. doi: 10.1016/j.jag.2024.103938
	Zhang K, Zhou R, Adhikarla E, et al. A generalist vision–language foundation model for diverse biomedical tasks. Nature Medicine, 2024a, 30 (11): 3129- 3141. doi: 10.1038/s41591-024-03185-2
	Zhang W, Cai M, Zhang T, et al. 2024b. EarthMarker: A visual prompting multi-modal large language model for remote sensing. IEEE Transactions on Geoscience and Remote Sensing. DOI: 10.1109/TGRS.2024.3523505.
	Zhu H, Qin S, Su M, et al. 2024. Harnessing large vision and language models in agriculture: A review. arXiv preprint arXiv: 240719679.

[1]	崔泽宇,张怀清,刘洋,张京,杨廷栋,傅汝饶. 树木三维建模与可视化模拟技术进展与应用[J]. 林业科学, 2025, 61(6): 1-12.
[2]	徐彩瑶,钱一莹,孔凡斌,柯水发,王浩. 乡村数字化影响林业新质生产力发展的理论逻辑与作用机制[J]. 林业科学, 2025, 61(6): 159-182.
[3]	张少鹏,付瑶,朱洪革,赵梦涵. 天然林资源保护政策对重点国有林区林业新质生产力的影响——基于林业绿色全要素生产率视角[J]. 林业科学, 2025, 61(6): 183-195.
[4]	朱洪革,赵梦涵,张少鹏. 国有林管理体制转型对林业新质生产力发展的影响机制[J]. 林业科学, 2025, 61(5): 207-221.
[5]	柯水发,万深玮,孔凡斌,朱洪革. 中国林业新质生产力的形成逻辑、时空演变与障碍因子分析[J]. 林业科学, 2025, 61(3): 199-213.
[6]	朱教君,王高峰,张怀清,高添. 关于“气候智慧林业”研究的思考[J]. 林业科学, 2024, 60(7): 1-7.