Abstract:

[Significance] To provide a reference for advancing high-quality agricultural production driven by data, this paper focuses on the issues of inconsistent acquisition standards, incomplete data collection, and ambiguous governance mechanisms in China's agricultural production data, examines existing governance models for agricultural production big data, and clarifies the technical pathways for realizing the value of data elements through the integrated and innovative application of key big data governance technologies and tools in practical scenarios. [Progress] From the perspective of agricultural production big data governance, this paper explores 17 types of big data governance technologies and tools across six core processes: Data acquisition and processing, data storage and exchange, data management, data analysis, large models, and data security guarantee. It conducts in-depth research on the application methods of big data governance technologies in agricultural production, revealing that: Remote sensing, unmanned aerial vehicle(UAV), Internet of Things (IoT), and terminal data acquisition and processing systems are already reatively mature; data storage and exchange system are developing rapidly, data management technologies remain in the initial stage; data analysis technologies have been widely applied; large model technology systems have taken initially shape; and data security assurance systems are gradually being into parctice. The above technologies are effectively applied in scenarios through tools and middleware such as data matching, computing power matching, network adaptation, model matching, scenario matching, and business configuration. This paper also analyzes the data governance throughout the entire agricultural production chain, including pre-production, in-production, and post-production, stages, as well as service cases involving different types of agricultural parks, research institutes and universities, production entities, and farmers. It demonstrates that sound data governance can provide sufficient planning and input analysis prior to production, helping planting entities in making rational plans. In production, it can provide data-driven guidance for key scenarios such as agricultural machinery operations and agricultural technical services, thereby fully supporting decision-making in the production process; and based on massive data, it can achieve reliable results in yield assessment and production benefit evaluation. Additionally, the paper introduces governance experience from national-level industrial parks, provincial-level agricultural science and technology parks, and some single-product entities, and investigates domestic and international technologies, practices, and tools related to agricultural production big data governance, indicating that there is a need to break through the business chains and service model of agricultural production across regions, themes, and scenarios. [Conclusions and Prospects] This paper presents insights into the future development directions of agricultural production big data governance, encompassing the promotion of standard formulation and implementation for agricultural production big data governance, the establishment of a universal resource pool for such governance, the expansion of diversified application scenarios, adaptation to the new paradigm of large-model- and massive-data-driven agricultural production big data governance, and the enhancement of security and privacy protection for agricultural production big data.

Key words: agricultural big data, big data governance, big data acquisition and processing, metadata, big data security, agricultural large model