[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-146":3,"related-146":72},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":8,"published_at":11,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":23,"tags":28,"primary_topic":34,"secondary_topics":40,"classification":41,"search_phrases":65,"slug":68,"view_count":69,"doi":70,"paper":8,"created_at":71},146,"AgriAgent: 面向农业环境控制的端到端大模型智能体系统架构","https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042","裘佳颖、刘应昌、高星杰、黄远、张红雨、田芳、李万理、冯在文提出端到端工具增强型大模型智能体系统架构AgriAgent:通过知识检索与工具调用闭环,在农业场景下实现\"环境感知-动态决策-执行反馈\"的全周期闭环控制。实验表明,番茄产量分别提升926.20%、979.55%、1537.46%(三种配置);卷心菜产量分别提升501.67%、504.65%、1185.14%;当参数量达7B时智能体展现卓越的灵活性与适应性,能解决大部分复杂农业生产决策问题,基于DSSAT作物生长模型构建农业评测平台。",null,"欢迎您访问《智慧农业（中英文）》官方网站！[English](https:\u002F\u002Fwww.smartag.net.cn\u002FEN\u002F2096-8094\u002Fhome.shtml)\n\n[![Image 1](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002F2096-8094\u002Fimages\u002Fbanner.png)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F2096-8094\u002Fhome.shtml)\n\n[![Image 2](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002F2096-8094\u002Fimages\u002Fbanner_mobile.png)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F2096-8094\u002Fhome.shtml)\n\n[![Image 3](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002F2096-8094\u002Fimages\u002Fyyzkj.png)](https:\u002F\u002Fpano.bookan.com.cn\u002F?id=60704&cid=352)\n\nToggle navigation[智慧农业](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042)\n\n*   [网站首页](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F2096-8094\u002Fhome.shtml)\n*   [期刊介绍](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fcolumn\u002Fcolumn133.shtml)\n*   [编委会](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fcolumn\u002Fcolumn153.shtml)\n*   [投稿指南](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fcolumn\u002Fcolumn135.shtml)\n*   [期刊服务](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fcolumn\u002Fcolumn124.shtml)\n*   [下载中心](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fcolumn\u002Fcolumn126.shtml)\n*   [联系我们](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fcolumn\u002Fcolumn127.shtml)\n\n[Smart Agriculture](https:\u002F\u002Fwww.smartag.net.cn\u002F) ›› [2026](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Farticle\u002FshowTenYearVolumnDetail.do?nian=2026), [Vol. 8](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Farticle\u002FshowTenYearVolumnDetail.do?nian=2026) ›› [Issue (2)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fvolumn\u002Fvolumn_838.shtml): 220-236.doi: [10.12133\u002Fj.smartag.SA202507042](https:\u002F\u002Fdoi.org\u002F10.12133\u002Fj.smartag.SA202507042)\n\n• 智能装备与系统 •[上一篇](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fabstract\u002Fabstract22626.shtml)[下一篇](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Fabstract\u002Fabstract22589.shtml)\n\n### AgriAgent：面向农业环境控制的端到端大模型智能体系统架构\n\n裘佳颖 1, 刘应昌 1, 高星杰 1, 黄远 2, 张红雨 1, 田芳 1,3, 李万理 1([![Image 4](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Femail.png)](mailto:liwanli@mail.hzau.edu.cn)), 冯在文 1,3,4([![Image 5](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Femail.png)](mailto:zaiwen.Feng@mail.hzau.edu.cn))\n\n1.   1. 华中农业大学信息学院，湖北 武汉 430070，中国\n\n2. 华中农业大学园艺林学学院，湖北 武汉 430070，中国\n\n3. 云南现代农业产业研究院有限公司，云南 昆明 650032，中国\n\n4. 农业智能技术教育部工程研究中心，湖北 武汉 430070，中国\n\n*   `收稿日期:`2025-07-30 `出版日期:`2026-03-30 \n*   `基金项目:`湖北省技术创新计划项目(2024BBB055); 湖北省科技厅援藏项目(2024EIA009); 云南省重大科技专项计划(202502AE090003); 中央高校基本科研业务费专项资金(2662025XXPY005); 中央高校基本科研业务费专项资金(2662023XXQD002); 农业融合出版知识挖掘与知识服务重点实验室开放课题(2025KMKS01) \n*   `作者简介:`裘佳颖，本科，研究方向为自然语言处理。E-mail：[19560533416@163.com](mailto:19560533416@163.com)； 刘应昌，硕士研究生，研究方向为自然语言处理。E-mail：[liuyingchang@webmail.hzau.edu.cn](mailto:liuyingchang@webmail.hzau.edu.cn) 裘佳颖，本科，研究方向为自然语言处理。E-mail：[19560533416@163.com](mailto:19560533416@163.com)； 裘佳颖和刘应昌并列第一作者 \n*   `通信作者:`李万理，博士，讲师，研究方向为自然语言处理、知识图谱、元学习。E-mail：[liwanli@mail.hzau.edu.cn](mailto:liwanli@mail.hzau.edu.cn)； 冯在文，博士，副教授，研究方向为农业AI大模型。E-mail：[zaiwen.Feng@mail.hzau.edu.cn](mailto:zaiwen.Feng@mail.hzau.edu.cn) \n\n### AgriAgent: End-to-End Large Model Agent System Architecture for Agricultural Environment Control\n\nQIU Jiaying 1, LIU Yingchang 1, GAO Xingjie 1, HUANG Yuan 2, ZHANG Hongyu 1, TIAN Fang 1,3, LI Wanli 1([![Image 6](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Femail.png)](mailto:liwanli@mail.hzau.edu.cn)), FENG Zaiwen 1,3,4([![Image 7](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Femail.png)](mailto:zaiwen.Feng@mail.hzau.edu.cn))\n\n1.   1. College of Informatics, Huazhong Agricultural University, Wuhan 430070, China\n\n2. College of Horticulture and Forestry, Huazhong Agricultural University, Wuhan 430070, China\n\n3. Yunnan Modern Agricultural Industry Research Institute, Kunming 650032, China\n\n4. Engineering Research Center of Agricultural Intelligent Technology, Ministry of Education, Wuhan 430070, China\n\n*   `Received:`2025-07-30 `Online:`2026-03-30 \n*   `Foundation items:`Hubei Provincial Technology Innovation Plan Project(2024BBB055); Tibet Assistance Project of Department of Science and Technology of Hubei Province(2024EIA009); Yunnan Provincial Major Science and Technology Special Project Plan(202502AE090003); Fundamental Research Funds for the Chinese Central Universities(2662025XXPY005); Open Project of Key Laboratory of Knowledge Mining and Knowledge Service for Agricultural Integrated Publishing(2025KMKS01) \n*   `About author:`QIU Jiaying, E-mail: [19560533416@163.com](mailto:19560533416@163.com);  LIU Yingchang, E-mail: [liuyingchang@webmail.hzau.edu.cn](mailto:liuyingchang@webmail.hzau.edu.cn) \n*   `Corresponding author:`LI Wanli, E-mail: [liwanli@mail.hzau.edu.cn](mailto:liwanli@mail.hzau.edu.cn); FENG Zaiwen, E-mail: [zaiwen.Feng@mail.hzau.edu.cn](mailto:zaiwen.Feng@mail.hzau.edu.cn) \n\n## [在线阅读 99](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)\n\n## [知网下载](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042)\n\n## [本地下载 86](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#1)\n\n[_![Image 8](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Fknowledge\\_map1.png)_ ## 可视化 0](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Flexeme\u002FshowArticleByLexeme.do?articleID=22689)\n\n#### [摘要\u002FAbstract](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042)\n\n**摘要：**\n\n**【目的\u002F意义】** 为突破大语言模型在农业生产决策领域应用中存在的物理世界交互能力不足、专业工具调用受限，以及“幻觉”生成等问题，构建了一个工具增强型环控智能体系统。 **【方法】** 首先，基于开放气象数据与作物生长模型构建农业数字孪生评测平台，提供标准化闭环控制验证环境。其次，提出了AgriAgent架构，通过“传感器-记忆库-检索器-大语言模型-工具执行器”闭环流程，实现了田间环境的动态感知、知识检索、推理决策与精准执行的端到端闭环控制。该端到端架构的优势在于实现了“环境感知-动态决策-执行反馈”的全流程无缝衔接，极大地提升了系统在复杂、动态农业环境下的实时响应能力和决策精度，解决了传统应用缺乏“感知-决策-执行”闭环能力的问题。为验证系统有效性，在数字孪生环境中，对玉米、小米、甜菜、番茄、卷心菜5种作物在常规、扰动、极端3种情景下开展实验。 **【结果和讨论】** 以Qwen2.5_7B为核心的AgriAgent-7B模型表现最优，常规情景、扰动情景和极端情景下玉米产量较基线模型分别提升562.20%、459.70%、463.60%；小米产量分别提升646.80%、558.70%、351.20%；甜菜产量分别提升228.20%、118.20%、125.40%；番茄产量分别提升926.20%、979.55%、1 537.46%；卷心菜产量分别提升501.67%、504.65%、1 185.14%。此外，智能体参数量显著影响决策的质量，当参数量达到7 B时，智能体展现了卓越的灵活性与适应性，能够解决大部分复杂的农业生产决策问题。 **【结论】** 本研究验证了工具增强型智能体在农业动态决策中的有效性，提出的AgriAgent端到端架构，通过知识检索与工具调用闭环，在农业场景下实现了“环境感知-动态决策-执行反馈”的全周期闭环控制，为农业智能体研究提供了可靠技术路径和创新范式。\n\n**关键词:**[工具增强型大语言模型,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[农业智能体,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[数字孪生,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[农业评测平台,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[DSSAT\u002F作物生长模型](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)\n\n**Abstract:**\n\n**[Objective]** Large language models (LLMs) have demonstrated strong capabilities in natural language understanding, knowledge integration, and complex reasoning, offering new opportunities for intelligent decision-making in agriculture. However, their direct application in agricultural production and facility environment control remains challenging due to strong physical constraints and high operational risks. The lack of real-world interaction and executable decision grounding limits the practical effectiveness of conventional LLMs in such scenarios. To address these challenges, a tool-augmented LLM-based agricultural intelligent agent system, termed AgriAgent, was proposed, and a digital-twin-based evaluation platform for agricultural decision-making was developed. By integrating a high-fidelity digital twin environment with an end-to-end agent architecture, the decision-making performance of agricultural intelligent agents with different parameter scales was systematically evaluated across multiple crops and climate scenarios. **[Methods]** A high-fidelity agricultural digital twin evaluation platform was constructed using the decision support system for agrotechnology transfer (DSSAT) v4.8 crop growth model as the core simulation engine to model crop growth under diverse environmental conditions and management strategies. Meteorological driving data were obtained from the Seoul Historical Weather Data dataset. Through data cleaning, missing-value imputation, unit normalization, and time-series reconstruction, the raw meteorological data were transformed into standardized inputs compatible with DSSAT. Three climate scenarios representing different environmental complexities were designed, including a regular scenario, a perturbed scenario, and an extreme scenario. The regular scenario employed historical observations, the perturbed scenario introduced stochastic disturbances to simulate short-term climate variability, and the extreme scenario incorporated multi-factor coupled stresses such as high temperatures and excessive precipitation during sensitive growth stages. In total, 90 annual climate driving sequences were generated. Fixed soil profile parameters calibrated by domain experts were applied across all simulations to minimize confounding effects. Within this digital twin environment, a tool-augmented agricultural intelligent agent, AgriAgent, was implemented using a modular architecture consisting of a sensor module, memory module, retriever, large language model, and tool executor, forming a closed-loop decision-making framework. In each decision cycle, the agent perceived environmental and crop state information, including soil moisture and nutrient status, meteorological conditions, crop growth stages, and stress indicators. State summaries and historical decisions were stored in memory, while agronomic knowledge was retrieved through a retrieval-augmented generation mechanism. Based on integrated information, the LLM generated structured environmental control commands in JSON format, which were validated and constrained by the tool executor before updating the DSSAT environment. The system supported irrigation, supplementary lighting, ventilation, heating, fertilization, and CO 2 enrichment. Five representative crops: maize, millet, sugar beet, tomato, and cabbage were simulated under the three climate scenarios over complete growing seasons, resulting in 450 crop-scenario combinations. An unmanaged DSSAT simulation served as the baseline. AgriAgent models with three parameter scales (1.5B, 3B, and 7B), built on the Qwen2.5 series, were evaluated. Crop economic yield expressed as dry matter at physiological maturity was adopted as the evaluation metric. **[Results and Discussions]** The results showed that AgriAgent consistently outperformed the baseline across all crops and climate scenarios, with model scale exerting a significant influence on decision-making performance. AgriAgent-7B achieved the best overall performance under regular, perturbed, and extreme scenarios, demonstrating strong generalization ability and environmental adaptability. By dynamically adjusting water, nutrient, light, and thermal management strategies, the agent effectively mitigated environmental stresses even under multi-factor coupled extreme climate conditions. Under extreme scenarios, AgriAgent-7B increased yields by 463.60% for maize, 351.20% for millet, 125.40% for sugar beet, 1 537.46% for tomato, and 1 185.14% for cabbage compared with the baseline. Particularly large gains were observed for high-value crops such as tomato and cabbage, highlighting the advantages of the proposed framework for precision-controlled facility agriculture. In contrast, AgriAgent-1.5B exhibited performance comparable to the baseline, while AgriAgent-3B achieved moderate improvements but remained inferior to the 7B model. These findings indicate a clear scaling effect, suggesting that larger models possess stronger capabilities in multi-source information integration, long-term temporal reasoning, and adaptation to complex environments. **[Conclusions]** This study developed a digital-twin-based agricultural decision evaluation platform and proposed a tool-augmented, end-to-end agricultural intelligent agent named AgriAgent. Experiments across multiple crops and climate scenarios verified the effectiveness and robustness of the proposed framework for dynamic agricultural decision-making. The results demonstrate that integrating knowledge retrieval, reasoning, and tool execution within a closed-loop LLM-based agent enables stable, reliable, and adaptive environmental control, providing a feasible technical pathway and standardized evaluation paradigm for intelligent agriculture.\n\n**Key words:**[tool-augmented large language model,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[agricultural agent,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[digital twin,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[agricultural evaluation platform,](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)[DSSAT\u002Fcrop growth model](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)\n\n**中图分类号:**\n\n*   [S126](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)\n\n*   [TP18](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#)\n\n#### 引用本文\n\n裘佳颖, 刘应昌, 高星杰, 黄远, 张红雨, 田芳, 李万理, 冯在文. AgriAgent：面向农业环境控制的端到端大模型智能体系统架构[J]. 智慧农业(中英文), 2026, 8(2): 220-236.\n\nQIU Jiaying, LIU Yingchang, GAO Xingjie, HUANG Yuan, ZHANG Hongyu, TIAN Fang, LI Wanli, FENG Zaiwen. AgriAgent: End-to-End Large Model Agent System Architecture for Agricultural Environment Control[J]. Smart Agriculture, 2026, 8(2): 220-236.\n\n#### 使用本文\n\n**[](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#) \u002F[推荐](javascript:void(null))**\n**导出引用管理器**[EndNote](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Farticle\u002FgetTxtFile.do?fileType=EndNote&id=22689)|[Ris](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Farticle\u002FgetTxtFile.do?fileType=Ris&id=22689)|[BibTeX](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002Farticle\u002FgetTxtFile.do?fileType=BibTeX&id=22689)\n\n**链接本文:**[https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042)\n\n[https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002FY2026\u002FV8\u002FI2\u002F220](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002FY2026\u002FV8\u002FI2\u002F220)\n\n#### [图\u002F表 14](javascript:;)\n\n[![Image 9](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002Fthumbnail\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F001.jpg)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#figureClass13975)\n\n**图1**\n\n工具增强型环控智能体系统架构示意图\n\n![Image 10](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F001.jpg)\n\n 图1 \n\n[![Image 11](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Ftable-icon.gif)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#inline_content13976)\n\n**表1**\n\nAgriAgent控制系统建模\n\n| 设备 | 控制动作符号 | 范围\u002F分辨率 | 主要影响 |\n| :---: | :---: | :---: | :---: |\n| 灌溉系统 | _I_ | 0~30 mm\u002Fh | 提升 W s o i l$W_{s o i l}$ |\n| 补光系统 | _L_ | 0~200% | 提升光合有效辐射 |\n| 通风\u002F加温 | _V_，_H_ | 0\u002F1（开关） | 调节 T a v g,W S$T_{a v g} , W S$ |\n| 施肥系统 | _F_ | 0~50 kg\u002Fhm 2 | 增加 N t o t a l$N_{t o t a l}$​ |\n| CO 2 施放 | _C_ | 0~800 μmol\u002Fmol\u002Fh | 提升室内CO₂ |\n\n 表1 \n\n[![Image 12](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002Fthumbnail\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F002.jpg)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#figureClass13977)\n\n**图2**\n\n数字孪生农场功能示意图\n\n![Image 13](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F002.jpg)\n\n 图2 \n\n[![Image 14](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002Fthumbnail\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F003.jpg)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#figureClass13978)\n\n**图3**\n\nAgriAgent架构示意图\n\n![Image 15](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F003.jpg)\n\n 图3 \n\n[![Image 16](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Ftable-icon.gif)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#inline_content13979)\n\n**表2**\n\nAgriAgent系统组件核心职能与关键实现要点\n\n| 模块 | 功能摘要 | 实现要点 |\n| :---: | :---: | :---: |\n| 传感器 | 从仿真环境读取状态向量 s t$\\mathbf{\\mathit{s}}_{t}$ | 以API形式模拟真实传感器 |\n| 记忆库 | 持久化存储全局状态摘要 | 以文本形式保存在System prompt中 |\n| 格式转换器 | 将结构化状态拼接为自然语言片段 | 以固定规则实现 |\n| 检索器（Retriever） | 面向知识库检索作物生理模型、控制经验 | 以Faiss检索实现 |\n| LLM | 解析环境信息并生成JSON格式工具调用 | 经过预训练和后训练的大语言模型 |\n| 工具执行器 | 校验、裁剪并调用修改DSSAT 的接口 | Python实现 |\n\n 表2 \n\n[![Image 17](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Ftable-icon.gif)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#inline_content13980)\n\n**表3**\n\n数字孪生农场土壤的理化性质与特征参数\n\n| 土层深度\u002Fcm | ［0，20） | ［20，40） | ［40，60） | ［60，80） | ［80，100） |\n| :---: | :---: | :---: | :---: | :---: | :---: |\n| 颗粒组成\u002F% | 黏粒 | 21.6 | 21.4 | 20.8 | 11.2 | 10.4 |\n| 粉粒 | 78.4 | 78.6 | 78 | 78.1 | 69.3 |\n| 砂粒 | 0.0 | 0.0 | 1.2 | 10.7 | 20.3 |\n| 土壤容重\u002F（g\u002Fcm 3） | 1.47 | 1.44 | 1.52 | 1.54 | 1.46 |\n| 田间持水量\u002F（cm 3\u002Fcm 3） | 0.306 0 | 0.274 9 | 0.319 1 | 0.375 6 | 0.328 8 |\n| 永久凋萎点\u002F（cm 3\u002Fcm 3） | 0.080 6 | 0.095 2 | 0.093 1 | 0.072 3 | 0.093 1 |\n| 土壤有机碳\u002F% | 1.0 | 0.5 | 0.5 | 0.4 | 0.3 |\n| pH | 8.47 | 8.80 | 8.82 | 8.84 | 8.85 |\n| 饱和含水量\u002F（cm 3\u002Fcm 3） | 0.410 5 | 0.420 9 | 0.257 8 | 0.426 0 | 0.429 2 |\n| 饱和导水率\u002F（cm\u002Fd） | 45.60 | 25.13 | 55.56 | 7.49 | 130.30 |\n\n 表3 \n\n[![Image 18](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002Fthumbnail\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F004.jpg)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#figureClass13981)\n\n**图4**\n\nAgriAgent模型（1.5B， 3B， 7B）与基线模型在5种作物和3种情景下的性能对比（以最终作物干物质产量（kg\u002Fhm2）衡量，数据为30个独立气候轨迹的平均值）\n\n![Image 19](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F004.jpg)\n\n 图4 \n\n[![Image 20](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Ftable-icon.gif)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#inline_content13982)\n\n**表4**\n\n不同参数量AgriAgent模型在5种作物与3种情景下的平均最终产量干重 ( kg\u002Fhm2)\n\n| 作物 | 模型 | 常规 | 扰动 | 极端 |\n| :---: | :---: | :---: | :---: | :---: |\n| 玉米 | Baseline | 990.10±144.55 | 997.60±142.00 | 600.67±524.62 |\n| AgriAgent-1.5B | 990.10±144.55 | 997.60±142.00 | 600.67±524.62 |\n| AgriAgent-3B | 5 440.80±1 520.62 | 5 287.33±1 471.01 | 3 304.53±2 950.12 |\n| AgriAgent-7B | 6 557.83±950.12 | 5 584.17±1 200.45 | 3 385.20±2 900.34 |\n| 小米 | Baseline | 867.20±268.49 | 894.63±268.79 | 721.97±426.74 |\n| AgriAgent-1.5B | 867.20±268.49 | 894.63±268.79 | 721.97±426.74 |\n| AgriAgent-3B | 5 281.03±1 174.62 | 5 291.87±1 428.79 | 3 431.40±1 948.76 |\n| AgriAgent-7B | 6 474.03±1 400.56 | 5 892.70±1 200.78 | 3 257.83±1 800.45 |\n| 甜菜 | Baseline | 8 587.90±6 250.11 | 8 853.70±6 459.85 | 8 845.07±7 099.96 |\n| AgriAgent-1.5B | 8 592.87±6 265.19 | 7 835.5±6 465.73 | 9 081.30±7 113.27 |\n| AgriAgent-3B | 16 755.30±5 911.93 | 13 818.70±6 247.26 | 15 365.40±7 188.85 |\n| AgriAgent-7B | 28 186.00±2 745.67 | 19 315.33±2 400.50 | 19 938.23±2 800.25 |\n| 番茄 | Baseline | 39.43±57.93 | 35.20±135.13 | 22.13±91.36 |\n| AgriAgent-1.5B | 39.43±57.93 | 35.20±135.13 | 22.13±91.36 |\n| AgriAgent-3B | 367.97±178.01 | 352.67±142.79 | 300.03±194.80 |\n| AgriAgent-7B | 404.63±120.34 | 380.00±150.67 | 362.17±200.12 |\n| 卷心菜 | Baseline | 88.87±6.52 | 84.97±7.32 | 35.20±25.99 |\n| AgriAgent-1.5B | 88.87±6.52 | 84.97±7.32 | 35.20±25.99 |\n| AgriAgent-3B | 398.53±127.86 | 336.60±158.31 | 311.57±142.41 |\n| AgriAgent-7B | 534.70±80.12 | 513.77±70.45 | 452.27±90.23 |\n\n 表4 \n\n[![Image 21](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002Fthumbnail\u002F005.jpg)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#figureClass13983)\n\n**图5**\n\nAgriAgent-7B模型调控下玉米、小米和甜菜收获物干重随时间的变化曲线（常规情景）\n\n![Image 22](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002F005.jpg)\n\n 图5 \n\n[![Image 23](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002Fthumbnail\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F006.jpg)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#figureClass13984)\n\n**图6**\n\nAgriAgent 在设施环境控制中的“感知-决策-执行”闭环流程样例：基于实时环境数据的工具调用决策\n\n![Image 24](https:\u002F\u002Fwww.smartag.net.cn\u002Ffileup\u002F2096-8094\u002FFIGURE\u002F2026-8-2\u002FImages\u002F2096-8094-2026-8-2-220\u002FF1A06695-387E-4f44-9B6A-ED4922CA7FF3-F006.jpg)\n\n 图6 \n\n[![Image 25](https:\u002F\u002Fwww.smartag.net.cn\u002Fimages\u002Ftable-icon.gif)](https:\u002F\u002Fwww.smartag.net.cn\u002FCN\u002F10.12133\u002Fj.smartag.SA202507042#inline_content13985)\n\n**表5**\n\n高影响幻觉（权重1.0）类型及核心特征：决策错误、天气条件与产量影响关联\n\n| 决策错误 | 天气条件 | 对产量的影响 |\n| :---: | :---: | :---: |\n| 极端高温下灌溉不足 | 温度>作物临界高温 | 作物热胁迫，光合作用受阻，直接减产 |\n| 极端高温下大量施肥 | 温度>作物临界高温+5 ℃ | 烧苗，根系损伤，不可逆伤害 |\n| 干旱条件灌溉不足 | 降水少+温度高 | 水分胁迫，生长停滞，严重减产 |\n| 低湿度高温灌溉不足 | 湿度低+温度高 | 过度蒸","智慧农业(中英文) 2026 8(2): 220-236","2026-07-28T00:00:00Z","论文",25,false,71,{"impact":17,"substance":18,"depth":17,"authority":19,"freshness":20,"relevant":21,"comment":22},18,20,13,2,1,"提出农业环境控制大模型智能体架构，方法新颖，但发表于3月，时效性低。",[24,25],{"name":10,"url":6},{"name":26,"url":27},"智慧农业(中英文)","https:\u002F\u002Fcastjournals.cast.org.cn\u002Fjoweb\u002Fzhny\u002FCN",[29,30,31,32,33],"大语言模型","智能体","数字孪生","DSSAT","农业环境控制",{"code":35,"slug":36,"name":37,"path":38,"parent_code":39},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":42,"locked":14,"version":43,"method":44,"evidence":45},"classified","2026-10-22.2","rules",{"candidates":46,"agricultural_evidence":48},[47,53,60],{"code":35,"score":19,"supported":48,"title":49,"abstract":51,"tags":52},true,[50],"大模型",[50],[29,30],{"code":54,"score":55,"supported":48,"title":56,"abstract":57,"tags":59},"B1",3,[],[58],"番茄",[],{"code":61,"score":21,"supported":14,"title":62,"abstract":63,"tags":64},"I1",[],[],[31],[66,67],"农业环境控制 大语言模型 数字孪生 智能体","农业环境控制 大语言模型","农业环境控制大语言模型数字孪生智能体-146",12,"10.12133\u002Fj.smartag.sa202507042","2026-08-01T06:01:55.418382Z",{"total":73,"page":21,"page_size":73,"items":74},6,[75,167,238,292,349,437],{"id":76,"title":77,"url":78,"summary":79,"summary_zh":80,"content":8,"source_name":81,"source_url":78,"published_at":82,"category":12,"cover_url":8,"hotness":83,"is_selected":14,"score":84,"score_detail":85,"sources":89,"tags":91,"primary_topic":95,"secondary_topics":96,"classification":97,"search_phrases":146,"slug":149,"view_count":150,"doi":151,"paper":152,"created_at":166},5232,"Current applications of artificial intelligence for advancing agriculture, sustainability and climate mitigation","https:\u002F\u002Fdoi.org\u002F10.1007\u002Fs44163-026-02540-1","Artificial intelligence (AI) is reshaping agriculture by enabling data-driven decision-making across crop production systems and supporting the transition toward more sustainable, resilient, and climate-smart farming. This review critically synthesizes advances in AI applications spanning precision irrigation, nutrient and soil management, yield prediction, pest and disease monitoring, computer vision, autonomous agricultural robotics, and large language model (LLM)-based decision support systems. Beyond summarizing technological developments, the review evaluates their practical deployment, emphasizing that successful agricultural AI depends not only on predictive accuracy but also on model generalizability, explainability, economic feasibility, interoperability, and integration with existing agronomic knowledge. Emerging technologies including Vision Transformers, foundation models, Digital Twins, multimodal AI, Retrieval-Augmented Generation, and federated and continual learning—are discussed as important directions for overcoming limitations related to data scarcity, environmental variability, and scalable field implementation. The review highlights that responsible AI adoption requires transparent governance frameworks addressing data ownership, privacy, algorithmic fairness, cybersecurity, and equitable access, particularly for smallholder farming systems. Although numerous AI models report outstanding performance, widespread agricultural adoption remains constrained by cross-region validation, fragmented datasets, infrastructure requirements, and insufficient real-world deployment. Future progress will depend on multidisciplinary collaboration integrating advances in artificial intelligence, sensing technologies, robotics, crop science, and agricultural engineering to develop robust, trustworthy, and farmer-centered decision-support systems. By critically evaluating technological innovations and implementation challenges, this review provides a framework for understanding the current state of AI in agriculture while identifying research priorities that will shape the next generation of sustainable and climate-resilient agricultural systems.","人工智能（AI）正在重塑农业，通过在作物生产系统中实现数据驱动的决策，并支持向更可持续、更具韧性和气候智能型农业转型。本综述批判性地综合了AI应用的进展，涵盖精准灌溉、养分与土壤管理、产量预测、病虫害监测、计算机视觉、自主农业机器人以及基于大语言模型（LLM）的决策支持系统。除总结技术发展外，本综述还评估了其实际部署情况，强调农业AI的成功不仅取决于预测准确性，还取决于模型泛化能力、可解释性、经济可行性、互操作性以及与现有农学知识的整合。文中讨论了包括视觉Transformer、基础模型、数字孪生、多模态AI、检索增强生成以及联邦学习和持续学习在内的新兴技术，将其视为克服数据稀缺、环境变异性和田间可扩展实施等局限的重要方向。本综述强调，负责任的AI采用需要透明的治理框架，以解决数据所有权、隐私、算法公平性、网络安全和公平获取等问题，尤其是对小农耕作系统而言。尽管众多AI模型报告了卓越的性能，但农业领域的广泛采用仍受限于跨区域验证、碎片化数据集、基础设施要求以及现实世界部署不足。未来的进展将取决于多学科协作，整合人工智能、传感技术、机器人技术、作物科学和农业工程领域的进展，以开发稳健、可信且以农民为中心的决策支持系统。通过批判性评估技术创新和实施挑战，本综述提供了一个理解AI在农业中现状的框架，同时确定了将塑造下一代可持续和气候韧性农业系统的研究优先事项。","Discover Artificial Intelligence","2026-10-17T00:00:00Z",10,83,{"impact":86,"substance":86,"depth":17,"authority":69,"freshness":87,"relevant":21,"comment":88},22,9,"该综述系统梳理了AI在农业中的应用与落地瓶颈，覆盖精准灌溉、机器人、大模型决策支持等方向，信息密度高，但属综述类论文，产业级突破性有限，可入精选。",[90],{"name":81,"url":78},[92,93,29,31,94],"智慧农业","农业人工智能","气候智慧农业",{"code":35,"slug":36,"name":37,"path":38,"parent_code":39},[],{"status":42,"locked":14,"version":43,"method":44,"evidence":98},{"candidates":99,"agricultural_evidence":48},[100,110,117,124,129,135,141],{"code":35,"score":101,"supported":48,"title":102,"abstract":104,"tags":109},19,[103],"artificial intelligence",[105,103,106,107,108],"计算机视觉","computer vision","large language model","decision support",[93,29],{"code":111,"score":73,"supported":48,"title":112,"abstract":113,"tags":116},"H2",[],[114,115],"土壤","soil",[],{"code":118,"score":73,"supported":48,"title":119,"abstract":120,"tags":123},"H3",[],[121,122],"灌溉","irrigation",[],{"code":61,"score":125,"supported":48,"title":126,"abstract":127,"tags":128},5,[],[31],[31,92],{"code":130,"score":55,"supported":48,"title":131,"abstract":132,"tags":134},"F6",[],[133],"农业机器人",[],{"code":136,"score":55,"supported":48,"title":137,"abstract":138,"tags":140},"G2",[],[139],"smallholder",[],{"code":142,"score":21,"supported":14,"title":143,"abstract":144,"tags":145},"H7",[],[],[94],[147,148],"气候智慧农业 AI","农业人工智能 气候智慧农业 大语言模型 数字孪生","气候智慧农业AI-5232",0,"10.1007\u002Fs44163-026-02540-1",{"doi":151,"openalex_id":153,"authors":154,"venue":81,"cited_by_count":150,"oa_url":78,"card":158,"direction":164,"ingested_from":165},"W7222823256",[155],{"name":156,"orcid":157},"Mostafa Farajpour","https:\u002F\u002Forcid.org\u002F0000-0003-4223-502X",{"tldr":159,"method":160,"finding":161,"direction":162,"opportunity":163},"综述农业人工智能在精准灌溉、产量预测、病虫害监测等应用及落地挑战。","文献综述，覆盖视觉Transformer、基础模型、数字孪生、多模态AI与联邦学","农业AI落地受限于跨区域验证、数据碎片化与基础设施，需兼顾可解释性与公平治理。","农业人工智能与决策模型","可研究小农户场景下跨区域可泛化、可解释且低成本的农业大模型与联邦学习部署。","智慧农业 \u002F 农业物联网","openalex","2026-10-18T23:30:07.666127Z",{"id":168,"title":169,"url":170,"summary":171,"summary_zh":172,"content":8,"source_name":173,"source_url":170,"published_at":174,"category":12,"cover_url":8,"hotness":83,"is_selected":14,"score":175,"score_detail":176,"sources":180,"tags":182,"primary_topic":185,"secondary_topics":190,"classification":192,"search_phrases":212,"slug":215,"view_count":150,"doi":216,"paper":217,"created_at":237},5376,"Agri-Synergy: A Hardware–Software Co-Designed Hybrid IoT System for Sensor Interface Standardization and LLM-Driven Interactive Decision Support","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fagriengineering8100446","This paper presents Agri-Synergy, a hybrid Internet of Things (IoT) system featuring a hardware–software co-design to tackle sensor data heterogeneity and the barriers to effective AI-driven decision support in smart agriculture. Unlike software-based data harmonization, we propose a universal sensor printed circuit board (PCB) with RS485\u002FModbus interface for hardware unification, complemented by on-node edge processing that ensures standardized data from source. To lower the entry barrier for farmers, the system integrates the DeepSeek large language model (LLM) API, combining static agricultural knowledge with real-time sensor data, for context-aware, natural language-based decision support, functioning without local model deployment. Field deployment showed reference-verified sensor accuracy predominantly within manufacturer specifications, stable 4G connectivity during the 4-h benchmark but intermittent dropouts during the 20-day deployment, and remote-control success rates of 96% (normal) and 80% (weak signal at 300 m). Multi-node concurrent transmission resulted in average latency of 238.0 ± 27.1 ms and packet loss below 0.5%. The system achieves a 3–5× cost reduction compared to commercial platforms. These results demonstrate that addressing data heterogeneity from source through hardware unification and software edge processing, while leveraging AI to lower the barrier to system adoption, provides a more reliable foundation for data synchronization and accessible AI-driven analytics.","本文提出Agri-Synergy，一种采用软硬件协同设计的混合物联网（IoT）系统，旨在解决智慧农业中传感器数据异构性问题以及有效AI驱动决策支持的障碍。与基于软件的数据协调方式不同，我们提出了一种带有RS485\u002FModbus接口的通用传感器印刷电路板（PCB）以实现硬件统一，并辅以节点端边缘处理，确保从源头获得标准化数据。为降低农民的使用门槛，该系统集成了DeepSeek大语言模型（LLM）API，将静态农业知识与实时传感器数据相结合，提供上下文感知的、基于自然语言的决策支持，且无需本地部署模型即可运行。现场部署表明，经参考验证的传感器精度基本处于制造商规格范围内；在4小时基准测试期间4G连接稳定，但在20天部署期间出现间歇性断连；远程控制成功率分别为96%（正常信号）和80%（300 m处弱信号）。多节点并发传输的平均延迟为238.0 ± 27.1 ms，丢包率低于0.5%。与商业平台相比，该系统实现了3–5倍的成本降低。这些结果表明，通过硬件统一和软件边缘处理从源头解决数据异构性问题，同时利用AI降低系统采用门槛，可为数据同步和可及的AI驱动分析提供更可靠的基础。","AgriEngineering","2026-10-19T00:00:00Z",76,{"impact":177,"substance":178,"depth":17,"authority":69,"freshness":87,"relevant":21,"comment":179},16,21,"硬件统一接口与边缘处理结合大模型决策支持的农业物联网系统论文，方法新颖、有实地部署数据，但属细分技术进展，公共影响有限。",[181],{"name":173,"url":170},[92,93,183,29,184],"农业物联网","传感器标准化",{"code":186,"slug":187,"name":188,"path":189,"parent_code":39},"I5","agricultural-iot","物联网传感与边缘计算","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-iot",[191],{"code":35,"slug":36,"name":37,"path":38,"parent_code":39},{"status":42,"locked":14,"version":43,"method":44,"evidence":193},{"candidates":194,"agricultural_evidence":48},[195,202,206],{"code":186,"score":17,"supported":48,"title":196,"abstract":198,"tags":201},[197],"IoT",[199,200,197],"物联网","internet of things",[183],{"code":35,"score":177,"supported":48,"title":203,"abstract":204,"tags":205},[108],[107,108],[93,29],{"code":61,"score":207,"supported":48,"title":208,"abstract":209,"tags":211},7,[],[92,210],"smart agriculture",[92],[213,214],"Agri-Synergy 物联网 传感器","DeepSeek 决策支持","Agri-Synergy物联网传感器-5376","10.3390\u002Fagriengineering8100446",{"doi":216,"openalex_id":218,"authors":219,"venue":173,"cited_by_count":150,"oa_url":170,"card":232,"direction":164,"ingested_from":165},"W7222920424",[220,222,224,226,228,230],{"name":221,"orcid":8},"Leqi Lai",{"name":223,"orcid":8},"Langtao Duan",{"name":225,"orcid":8},"Yuhao Chen",{"name":227,"orcid":8},"Yuhang Zheng",{"name":229,"orcid":8},"Jinyu Yang",{"name":231,"orcid":8},"Mingyao Wen",{"tldr":233,"method":234,"finding":235,"direction":164,"opportunity":236},"提出软硬件协同的混合物联网系统，统一传感器接口并用大模型提供交互式决策支持。","通用RS485\u002FModbus传感器PCB、节点边缘处理、DeepSeek大模型A","传感器精度达标，多节点延迟238ms、丢包低于0.5%，成本降低3-5倍。","可研究弱信号下通信可靠性提升与边缘-云端LLM协同的长期田间鲁棒性。","2026-10-20T23:30:08.319698Z",{"id":239,"title":240,"url":241,"summary":242,"summary_zh":243,"content":8,"source_name":244,"source_url":241,"published_at":174,"category":12,"cover_url":8,"hotness":83,"is_selected":14,"score":245,"score_detail":246,"sources":248,"tags":250,"primary_topic":253,"secondary_topics":254,"classification":255,"search_phrases":275,"slug":278,"view_count":150,"doi":279,"paper":280,"created_at":291},5315,"PLANT-GPK v4.0: Multi-Stage Plant Phenotyping with Time-Discovery and Harvest Prediction","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23461121","This paper presents a comprehensive extension to the PLANT-GPK framework, demonstrating its robust capability to analyze plants across entirely different developmental stages from single images. By formulating an inverse problem that simultaneously optimizes physiological parameters and the hidden time dimension (t), the model successfully discovered the exact age of two distinct wheat samples: a vegetative (green) stage and a senescent (golden) stage. The framework bridges computer vision, spatial heterogeneity analysis, and ODE-based physics to provide a complete digital twin capable of forecasting the exact day of harvest regardless of the plant's current state.","本文对PLANT-GPK框架进行了全面扩展，展示了其仅凭单张图像即可分析处于完全不同发育阶段植物的稳健能力。通过构建一个同时优化生理参数与隐藏时间维度（t）的反问题，该模型成功推断出两个不同小麦样本的精确年龄：一个处于营养期（绿色阶段），另一个处于衰老期（金色阶段）。该框架融合了计算机视觉、空间异质性分析以及基于常微分方程（ODE）的物理建模，提供了一个完整的数字孪生，无论植物当前处于何种状态，均能预测确切的收获日期。","Zenodo (CERN European Organization for Nuclear Research)",81,{"impact":17,"substance":86,"depth":17,"authority":19,"freshness":83,"relevant":21,"comment":247},"该研究提出从单张图像反演植物生育阶段与收获日的数字孪生框架，方法新颖且具产业应用潜力，值得进入每日精选。",[249],{"name":244,"url":241},[92,93,31,251,252],"植物表型","收获预测",{"code":35,"slug":36,"name":37,"path":38,"parent_code":39},[],{"status":42,"locked":14,"version":43,"method":44,"evidence":256},{"candidates":257,"agricultural_evidence":48},[258,264,268],{"code":61,"score":259,"supported":48,"title":260,"abstract":261,"tags":263},8,[],[31,262],"digital twin",[31,92],{"code":35,"score":207,"supported":48,"title":265,"abstract":266,"tags":267},[],[105,106],[93],{"code":269,"score":73,"supported":48,"title":270,"abstract":271,"tags":274},"A2",[],[272,273],"小麦","wheat",[],[276,277],"PLANT-GPK 植物表型","小麦 收获期 预测","PLANT-GPK植物表型-5315","10.5281\u002Fzenodo.23461121",{"doi":279,"openalex_id":281,"authors":282,"venue":244,"cited_by_count":150,"oa_url":241,"card":285,"direction":289,"ingested_from":165},"W7222875391",[283],{"name":284,"orcid":8},"MOHAMMED TAHAR SAIM",{"tldr":286,"method":287,"finding":288,"direction":289,"opportunity":290},"扩展PLANT-GPK框架，从单张图像反演植物生理参数与隐藏时间维度，实现跨生育期表型分析与收获期预","基于ODE物理模型与逆问题优化，联合估计生理参数和隐藏时间t，用两个不同生育期小","模型能准确发现两个不同生育期小麦样本的精确年龄，并预测确切收获日期。","农业遥感与作物表型","可将单图时间反演扩展到多作物多生育期，结合时序遥感构建大尺度数字孪生与收获期预报。","2026-10-19T23:30:19.267920Z",{"id":293,"title":294,"url":295,"summary":296,"summary_zh":297,"content":8,"source_name":244,"source_url":295,"published_at":298,"category":12,"cover_url":8,"hotness":83,"is_selected":14,"score":299,"score_detail":300,"sources":302,"tags":304,"primary_topic":307,"secondary_topics":308,"classification":309,"search_phrases":325,"slug":328,"view_count":21,"doi":329,"paper":330,"created_at":348},5194,"A Comprehensive Review on Multi-Modal Large Language Models and Agentic Architectures for Precision Agricultural Decision Support","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23411181","This article reviews recent developments in multi-modal large language models and agentic architectures for agricultural decision support. Forty-five representative studies published between 2023 and 2026 were selected across perception, language, orchestration, soil intelligence and governance, organised under a four-layer taxonomy, and classified against nine functional capabilities using a stated, reproducible procedure. The comparative analysis exposes a pronounced imbalance. Crop and disease perception is covered by 36 per cent of the selected studies, while dose-bearing prescription appears in 7 per cent, profit-aware recommendation in 4 per cent, and persistent per-field memory in a single study. Advisory tools in this literature therefore remain unable to exchange state: image-based scanners are rarely conditioned on measured soil chemistry, and crop-recommendation models optimise agronomic suitability without reference to farm economics. On this basis the article synthesises a conceptual reference architecture from the reviewed literature. It is a literature-derived synthesis rather than an implemented system, and calibrated abstention, evaluation protocols for longitudinal memory, and vernacular voice coverage emerge as the most urgent gaps.","本文综述了面向农业决策支持的多模态大语言模型与智能体架构的最新进展。按照可复现的既定流程，选取2023年至2026年间发表的45项代表性研究，涵盖感知、语言、编排、土壤智能与治理等方向，将其纳入一个四层分类体系，并依据九项功能能力进行分类。比较分析揭示出明显的不均衡：作物与病害感知在所选研究中占36%，而含剂量处方的研究仅占7%，利润感知推荐占4%，持久性田块记忆仅见于一项研究。因此，该领域的咨询工具仍无法交换状态：基于图像的扫描器很少以实测土壤化学状况为条件，作物推荐模型在优化农艺适宜性时也未考虑农场经济性。在此基础上，本文综合所综述文献提出一个概念性参考架构。该架构是文献衍生的综合框架，而非已实现的系统；校准弃权、纵向记忆的评估协议以及方言语音覆盖被列为最紧迫的空白。","2026-10-16T00:00:00Z",63,{"impact":69,"substance":17,"depth":177,"authority":259,"freshness":87,"relevant":21,"comment":301},"系统性综述揭示处方决策与农场经济维度严重缺位，对农业AI研发方向有参考价值，但属文献综述而非实证突破，影响范围有限。",[303],{"name":244,"url":295},[92,93,30,305,306],"多模态大模型","农业决策支持",{"code":35,"slug":36,"name":37,"path":38,"parent_code":39},[],{"status":42,"locked":14,"version":43,"method":44,"evidence":310},{"candidates":311,"agricultural_evidence":48},[312,317,321],{"code":35,"score":313,"supported":48,"title":314,"abstract":315,"tags":316},14,[108],[108],[93,305,30],{"code":111,"score":73,"supported":48,"title":318,"abstract":319,"tags":320},[],[114,115],[],{"code":61,"score":21,"supported":14,"title":322,"abstract":323,"tags":324},[],[],[92],[326,327],"多模态大模型 农业决策支持","农业人工智能 农业决策支持 多模态大模型 智慧农业","多模态大模型农业决策支持-5194","10.5281\u002Fzenodo.23411181",{"doi":329,"openalex_id":331,"authors":332,"venue":244,"cited_by_count":150,"oa_url":295,"card":343,"direction":162,"ingested_from":165},"W7222605322",[333,335,337,339,341],{"name":334,"orcid":8},"Tanmay Bhamre",{"name":336,"orcid":8},"Yash Barde",{"name":338,"orcid":8},"Devendra Yeole",{"name":340,"orcid":8},"Gajanan Chaudhary",{"name":342,"orcid":8},"Amarsinh Verpe",{"tldr":344,"method":345,"finding":346,"direction":162,"opportunity":347},"综述多模态大模型与智能体架构在精准农业决策支持中的应用与不足。","筛选2023-2026年45篇代表性研究，按四层分类与九项功能能力归类比较。","感知研究占36%，处方仅7%、利润推荐4%、田块记忆仅1篇，系统难以交换状态。","可研究跨模态状态交换、纵向记忆评测协议与方言语音覆盖，填补处方与经济决策空白。","2026-10-17T23:30:55.282694Z",{"id":350,"title":351,"url":352,"summary":353,"summary_zh":354,"content":8,"source_name":355,"source_url":352,"published_at":356,"category":12,"cover_url":8,"hotness":83,"is_selected":14,"score":245,"score_detail":357,"sources":359,"tags":361,"primary_topic":364,"secondary_topics":365,"classification":372,"search_phrases":414,"slug":417,"view_count":20,"doi":418,"paper":419,"created_at":436},4656,"Artificial Intelligence, Foundation Models, and Digital Twins for Climate-Resilient Agriculture: A Review","https:\u002F\u002Fdoi.org\u002F10.54963\u002Fia.v2i2.103039","Climate change, increasing food demand, and resource constraints are accelerating the transition from conventional precision agriculture toward smart, autonomous, and climate-resilient agricultural systems. Artificial intelligence (AI), foundation models, digital twins (DTs), the Internet of Things (IoT), remote sensing, edge–cloud computing, and big data analytics are emerging as key technologies for data-driven agricultural management. This review critically evaluates recent advances in AI-driven intelligent agriculture, with emphasis on machine learning, deep learning, explainable AI (XAI), generative and agentic AI, multimodal foundation models, and digital twins, together with the sensing and computing infrastructures required for their deployment. Relevant literature was identified using keywords related to AI, foundation models, digital twins, IoT, precision agriculture, and climate-resilient agriculture, with emphasis on recent peer-reviewed studies addressing technological applications, performance, limitations, challenges, and future prospects. The selected literature was critically synthesized to identify technological trends, knowledge gaps, and emerging research opportunities. The review highlights that the greatest potential of intelligent agriculture lies in integrating heterogeneous data from sensors, UAVs (unmanned aerial vehicles), satellites, weather systems, field operations, and agricultural knowledge to support context-aware prediction, decision-making, and autonomous action. However, data quality, model generalizability, explainability, interoperability, cybersecurity, computational requirements, and large-scale field validation remain major challenges. This review further develops an integrated perspective connecting AI, foundation models, multimodal data infrastructures, digital twins, and autonomous agricultural systems within a unified cyber-physical ecosystem. Future priorities include trustworthy AI, agentic decision-making, semantic interoperability, federated learning, continuous learning, and self-learning digital twins toward Agriculture 6.0.","气候变化、粮食需求增长与资源约束正加速推动传统精准农业向智能、自主且气候适应型的农业系统转型。人工智能（AI）、基础模型、数字孪生（DT）、物联网（IoT）、遥感、边缘—云计算和大数据分析正成为数据驱动农业管理的关键技术。本综述批判性地评估了人工智能驱动智能农业的最新进展，重点关注机器学习、深度学习、可解释人工智能（XAI）、生成式与代理式人工智能、多模态基础模型和数字孪生，以及其部署所需的传感与计算基础设施。相关文献通过人工智能、基础模型、数字孪生、物联网、精准农业和气候适应型农业等关键词检索获得，重点纳入近期同行评审研究，涵盖技术应用、性能、局限性、挑战与未来前景。对所选文献进行了批判性综合，以识别技术趋势、知识空白和新兴研究机遇。本综述强调，智能农业的最大潜力在于整合来自传感器、无人机（UAV）、卫星、气象系统、田间作业和农业知识的异构数据，以支持情境感知预测、决策和自主行动。然而，数据质量、模型泛化能力、可解释性、互操作性、网络安全、计算需求和大规模田间验证仍是主要挑战。本综述进一步提出了一种整合视角，将人工智能、基础模型、多模态数据基础设施、数字孪生和自主农业系统连接于统一的网络物理生态系统中。未来优先方向包括可信人工智能、代理式决策、语义互操作性、联邦学习、持续学习以及面向农业6.0的自学习数字孪生。","Intelligent Agriculture","2026-10-10T00:00:00Z",{"impact":86,"substance":18,"depth":17,"authority":19,"freshness":259,"relevant":21,"comment":358},"系统综述AI、基础模型与数字孪生在气候韧性农业中的集成路径与挑战，学术价值高，对智慧农业方向有较强参考意义。",[360],{"name":355,"url":352},[92,93,362,31,363],"农业大模型","气候韧性农业",{"code":35,"slug":36,"name":37,"path":38,"parent_code":39},[366,371],{"code":142,"slug":367,"name":368,"path":369,"parent_code":370},"climate","气候适应与防灾减灾","\u002Ftopic\u002Fenvironment\u002Fclimate","H",{"code":186,"slug":187,"name":188,"path":189,"parent_code":39},{"status":42,"locked":14,"version":43,"method":44,"evidence":373},{"candidates":374,"agricultural_evidence":48},[375,383,390,397,401,408],{"code":35,"score":101,"supported":48,"title":376,"abstract":377,"tags":382},[103],[378,379,380,381,103],"深度学习","机器学习","machine learning","deep learning",[93,362],{"code":61,"score":384,"supported":48,"title":385,"abstract":386,"tags":389},11,[],[387,31,388],"精准农业","precision agriculture",[31,92],{"code":142,"score":83,"supported":48,"title":391,"abstract":392,"tags":396},[],[393,394,395],"气候变化","气候适应","climate change",[363],{"code":186,"score":87,"supported":48,"title":398,"abstract":399,"tags":400},[],[199,200,197],[],{"code":402,"score":73,"supported":48,"title":403,"abstract":404,"tags":407},"I3",[],[405,406],"遥感","remote sensing",[],{"code":409,"score":55,"supported":48,"title":410,"abstract":411,"tags":413},"A7",[],[412],"间作",[],[415,416],"农业基础模型 数字孪生","气候韧性农业 人工智能","农业基础模型数字孪生-4656","10.54963\u002Fia.v2i2.103039",{"doi":418,"openalex_id":420,"authors":421,"venue":355,"cited_by_count":150,"oa_url":352,"card":431,"direction":164,"ingested_from":165},"W7221321372",[422,425,428],{"name":423,"orcid":424},"Krishan Kumar","https:\u002F\u002Forcid.org\u002F0009-0005-7569-4590",{"name":426,"orcid":427},"Annu Khatri","https:\u002F\u002Forcid.org\u002F0009-0000-3103-9520",{"name":429,"orcid":430},"Indu Shekhar Thakur","https:\u002F\u002Forcid.org\u002F0000-0003-2382-1218",{"tldr":432,"method":433,"finding":434,"direction":162,"opportunity":435},"综述AI、基础模型与数字孪生如何支撑气候韧性农业，并指出关键挑战与未来方向。","文献综述，聚焦机器学习、深度学习、可解释AI、多模态基础模型与数字孪生。","最大潜力在于融合多源异构数据实现情境感知预测与自主决策，但数据质量与可解释性等仍是瓶颈。","可探索可信AI、联邦学习与自学习数字孪生在真实农田的规模化验证与语义互操作。","2026-10-11T23:30:07.465569Z",{"id":438,"title":439,"url":440,"summary":441,"summary_zh":442,"content":8,"source_name":443,"source_url":440,"published_at":444,"category":12,"cover_url":8,"hotness":83,"is_selected":14,"score":245,"score_detail":445,"sources":447,"tags":449,"primary_topic":451,"secondary_topics":452,"classification":455,"search_phrases":492,"slug":495,"view_count":20,"doi":496,"paper":497,"created_at":510},4213,"Agri digital twin–driven intelligent computing framework for real-time crop yield forecasting and sustainable decision support","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.engappai.2026.116418","In the context of climate change, resource constraints, and increasing food demand, sustainable agricultural systems must become more productive and efficient in their use of resources. The Internet of Things (IoT) sensing and Artificial Intelligence (AI) technologies are allowing data-driven agriculture, but a common platform for real-time prediction of crop yield and sustainable agricultural decision-making is still not available. We propose an Agri Digital Twin (ADT) framework that combines a Gated Recurrent Unit (GRU) model with an IoT-based sensing system for temporal crop yield prediction and climate-resilient farm management. The framework constantly updates soil and weather parameters in real time, such as soil moisture, air temperature, humidity, atmospheric pressure, wind speed, wind gust, and wind direction, to create a real-time digital model of the crop field. The data collected is used to derive temporal features to predict yield, which are then categorized as low, medium, and high, and then further into sustainable, warning, and unsustainable to enable irrigation scheduling and crop stress management. The experimental results show that the proposed framework outperforms the others by achieving the Mean Absolute Error (MAE) of 0.0243, Root Mean Square Error (RMSE) of 0.0459, and Coefficient of Determination ( R 2 ) of 0.9254. Furthermore, it achieves an average prediction accuracy of 97.41% (95% CI: 95.03%, 99.80%), representing improvements of 4.09%, 6.32%, and 8.66% over BiLSTM, LSTM, and 1D-CNN, respectively ( p \u003C 0.05 ). The decision-support module also achieves a classification accuracy of 94.03%, showing its potential for sustainable and climate-resilient smart agriculture.","在气候变化、资源约束和粮食需求不断增长的背景下，可持续农业系统必须提高生产力并更高效地利用资源。物联网（IoT）传感与人工智能（AI）技术正在推动数据驱动型农业发展，但目前仍缺乏一个用于作物产量实时预测和可持续农业决策的通用平台。我们提出了一种农业数字孪生（Agri Digital Twin, ADT）框架，该框架将门控循环单元（Gated Recurrent Unit, GRU）模型与基于物联网的传感系统相结合，用于时序作物产量预测和气候适应型农场管理。该框架实时持续更新土壤和气象参数，如土壤湿度、气温、湿度、大气压、风速、阵风和风向，从而构建作物田地的实时数字模型。所采集的数据用于提取时序特征以预测产量，随后将产量划分为低、中、高，并进一步划分为可持续、警告和不可持续，以实现灌溉调度和作物胁迫管理。实验结果表明，所提框架优于其他方法，其平均绝对误差（MAE）为0.0243，均方根误差（RMSE）为0.0459，决定系数（R²）为0.9254。此外，其平均预测准确率达到97.41%（95% CI：95.03%，99.80%），相较于BiLSTM、LSTM和1D-CNN分别提升了4.09%、6.32%和8.66%（p \u003C 0.05）。决策支持模块的分类准确率也达到94.03%，显示出其在可持续且气候适应型智慧农业中的应用潜力。","Engineering Applications of Artificial Intelligence","2026-10-04T00:00:00Z",{"impact":17,"substance":86,"depth":17,"authority":313,"freshness":87,"relevant":21,"comment":446},"提出农业数字孪生与GRU融合的实时产量预测与决策支持框架，方法新颖、指标扎实，对智慧农业研究有参考价值。",[448],{"name":443,"url":440},[92,93,199,31,450],"作物产量预测",{"code":35,"slug":36,"name":37,"path":38,"parent_code":39},[453,454],{"code":186,"slug":187,"name":188,"path":189,"parent_code":39},{"code":142,"slug":367,"name":368,"path":369,"parent_code":370},{"status":42,"locked":14,"version":43,"method":44,"evidence":456},{"candidates":457,"agricultural_evidence":48},[458,462,466,470,474,478,482,488],{"code":61,"score":101,"supported":48,"title":459,"abstract":460,"tags":461},[262],[92,31,210,262],[31,92],{"code":35,"score":69,"supported":48,"title":463,"abstract":464,"tags":465},[108],[103],[93],{"code":186,"score":83,"supported":48,"title":467,"abstract":468,"tags":469},[],[199,200,197],[199],{"code":142,"score":87,"supported":48,"title":471,"abstract":472,"tags":473},[],[393,394,395],[],{"code":111,"score":73,"supported":48,"title":475,"abstract":476,"tags":477},[],[114,115],[],{"code":118,"score":73,"supported":48,"title":479,"abstract":480,"tags":481},[],[121,122],[],{"code":483,"score":55,"supported":48,"title":484,"abstract":485,"tags":487},"H8",[],[486],"可持续农业",[],{"code":402,"score":21,"supported":14,"title":489,"abstract":490,"tags":491},[],[],[450],[493,494],"Agri Digital Twin GRU 产量预测","物联网 数字孪生 智慧农业","AgriDigitalTwinGRU产量预测-4213","10.1016\u002Fj.engappai.2026.116418",{"doi":496,"openalex_id":498,"authors":499,"venue":443,"cited_by_count":150,"oa_url":440,"card":505,"direction":164,"ingested_from":165},"W7219849006",[500,503],{"name":501,"orcid":502},"Rohit Kumar","https:\u002F\u002Forcid.org\u002F0009-0000-8484-6588",{"name":504,"orcid":8},"Munish Bhatia",{"tldr":506,"method":507,"finding":508,"direction":164,"opportunity":509},"提出农业数字孪生框架，结合GRU与物联网实现实时产量预测与可持续决策支持。","基于物联网传感实时采集土壤气象数据，用GRU模型做时序产量预测与分类。","预测MAE 0.0243、R² 0.9254，平均准确率97.41%，优于BiLSTM、LSTM和1","可探索多源遥感与数字孪生融合、跨区域迁移学习及决策闭环验证等空白。","2026-10-05T23:30:09.119891Z"]