| 摘要: |
| 【目的】 系统综述数据智能驱动的作物群体数字孪生研究进展，阐明其理论框架、关键技术与应用方向，为实现作物“群体—养分—水分”一体化精准调控与智慧农业发展提供参考。【方法】 文章基于国内外相关文献，梳理数字孪生在农业中的定义、层级结构及核心组成；重点分析多源信息感知、群体结构表征、机理—AI融合建模、数据同化及虚实交互等关键技术；结合典型案例，总结其在作物养分水分管理、健康监测、结构优化与智能决策中的应用。【结果】 （1）数字孪生通过多源数据融合与模型协同，实现作物群体的动态映射与智能调控。（2）数字孪生通过虚实融合实现智能决策，并为水肥管理、群体健康监测及产量优化等关键环节提供了决策依据。（3）数字孪生仍然存在数据异构、模型标准化及系统集成不足等问题。【结论】 数据智能驱动的作物群体数字孪生是实现农业精准化与农业绿色发展的关键路径。未来应加强数据标准建设、模型协同与平台集成，构建开放共享的孪生体系，推动农业智能化与可持续发展。 |
| 关键词: 作物群体数字孪生 数据智能 多源感知 机理—AI融合 虚实交互 智能决策 智慧农业 |
| DOI：10.12105/j.issn.1672-0423.20260104 |
| 分类号: |
| 基金项目:教育部基础学科和交叉学科突破计划（JYB2025XDXM702） |
|
| Data-driven digital twins and intelligent approaches for crop population systems：A review |
|
Xu Beibei1,2, Cheng Jie3, Ping Jinming1,2, Chen Guipeng3, Cong Wenfeng1,2
|
|
1State Key Laboratory of Nutrient Use and Management/College of Resources and Environmental Sciences，China Agricultural University/National Academy of Agriculture Green Development，Beijing 100193，China;2Hebei Quzhou Agricultural Green Development National Field Scientific Observation and Research Station，Handan 057250，Hebei，China;3Agricultural Economics and Information Institute，Jiangxi Academy of Agricultural Sciences/Jiangxi Province Engineering Research Center of Intelligent Perception in Agriculture，Nanchang 330200，Jiangxi，China
|
| Abstract: |
| [Purpose] This study systematically reviews recent advances in data intelligence-driven digital twins for crop populations，with the aim of clarifying their conceptual frameworks，key enabling technologies，and application domains. The review provides a theoretical and technical reference for realizing integrated and precise regulation of crop populations，nutrients，and water，thereby supporting the development of smart agriculture.[Method] Based on relevant domestic and international literature，the definitions，hierarchical architectures，and core components of digital twins in agriculture were summarized. Key technologies were analyzed with emphasis on multi-source sensing，crop population structural representation，mechanism-AI hybrid modeling，data assimilation，and cyber-physical interaction. Typical case studies were examined to synthesize applications in nutrient and water management，health monitoring，structural optimization，and intelligent decision-making.[Result] By integrating heterogeneous multi-source data with collaborative modeling，digital twins enabled dynamic representation and intelligent regulation of crop populations. Through tight coupling between virtual models and physical systems，digital twins supported data-driven decision-making and provided scientific guidance for critical processes such as water and fertilizer management，population health assessment，and yield optimization. Nevertheless，several challenges remained，including data heterogeneity，insufficient model standardization，and limited system-level integration.[Conclusion] Data intelligence-driven digital twins of crop populations constitute a key pathway toward precision-oriented，green，and high-efficiency agricultural production. Future research should focus on strengthening data standards，enhancing model interoperability，and advancing platform integration to establish open and shareable digital twin ecosystems，thereby accelerating agricultural intelligence and sustainable development. |
| Key words: crop population digital twin data intelligence multi-source sensing mechanism-AI integration cyber-physical interaction intelligent decision-making smart agriculture |
