农业大数据学报

• •    下一篇

农业大灾风险分析与集成智能决策平台设计与应用

孔莉莎1吴焕萍1*李玉1谢能付2*刘北1肖风劲1薛晓萍3郭萃4   

  1. 1.国家气候中心,北京 1000812.中国农业科学院农业信息研究所,北京 100081;3.山东省气候中心,济南 250031; 4.北京超图软件股份有限公司,北京 100015

  • 出版日期:2026-01-19 发布日期:2026-01-19

Design and Application of an Integrated Platform for Agricultural Catastrophe Risk Analysis and Intelligent Decision-Making

KONG LiSha1, WU HuanPing1, LI Yu1, XIE NengFu2, LIU Bei1, XIAO FengJin1, XUE XiaoPing3, Guo Cui4   

  1. 1.National Climate Centre, Beijing 100081, China; 2.Institute of Agricultural Information, Chinese Academy of Agricultural Sciences, Beijing 100081, China; 3.Shandong Climate Centre, Jinan 250031, China; 4.Beijing Supermap Software Co.,Ltd, Beijing 100015, China

  • Published:2026-01-19 Online:2026-01-19

摘要: 为满足农业大灾监测识别、动态预警、产量预测、损失评估及保险理赔等全链条需求,有效支撑面向多应用场景、多用户的农业大灾风险分析与智能决策服务,设计并建设了农业大灾风险分析与集成智能决策平台。该平台以一体化、集成化、智能化、中台化为设计原则,基于数据层、平台层、应用层、表现层组成总体架构,采用基于气象大数据云平台的数据集成方法、基于云原生的多源异构算法集成、基于地理处理自动化的低代码构建、基于微前端和微服务的架构设计、基于人工智能生成内容的智能产品制作、基于人工智能大模型的农业大灾风险可视化等关键技术,通过构建数据中台、技术中台和业务中台,能够实现气象要素监测、农气灾害识别、灾害影响评估、风险决策服务等功能。初步应用显示,该平台具有良好的业务能力和发展前景,有助于提升跨部门、多角色间的全流程协同分析、智能推理和决策支持能力,有效应对气候变化、保障粮食安全,提高农业现代化治理水平。

关键词: 农业, 大灾, 风险, 智能, 大模型, 评估, 决策

Abstract: To meet the full-chain demands of agricultural disaster monitoring and identification, dynamic early warning, yield prediction, loss assessment and insurance claims, and to effectively support multi-scenario, multi-user agricultural catastrophe risk analysis and intelligent decision-making services, an integrated platform for agricultural catastrophe risk analysis and intelligent decision-making has been designed and developed. Guided by the design principles of integration, modularization, intelligence and platformization, based on the data layer, platform layer, application layer, presentation layer composed of the overall architecture, the platform employs key technologies such as data integration based on the meteorological big data cloud platform, multi-source heterogeneous algorithm integration based on cloud-native, low-code development via geospatial processing automation, micro-frontend and microservices architecture, intelligent product generation based on artificial intelligence generated content, and visualization of major agricultural disaster risks based on large artificial intelligence models. By building a data middle platform, a technology middle platform and a business middle platform, it enables functions including meteorological element monitoring, identification of agricultural meteorological disasters, disaster impact assessment, and risk-informed decision support. Preliminary applications demonstrate that the platform exhibits strong operational capabilities and promising potential for development. It contributes to enhancing cross-departmental, multi-role collaborative analysis, intelligent reasoning and decision-making throughout the entire workflow, thereby effectively addressing climate change challenges, safeguarding food security and advancing the modernization of agricultural governance.

Key words: agriculture, great disaster, risk, intelligence, large model, evaluation, decision-making