CropWatch, initiated in the 1990s, has undergone five major rounds of technological upgrades and has evolved into the CropWatch Cloud Platform for Global Agricultural Remote Sensing. The platform adopts a one stop service framework integrating data–function–computation–information. Built upon a microservices architecture and standardized interface encapsulation technology, it achieves the atomic decomposition of agricultural monitoring models and provides over 200 APIs (Fig. B.1). This modular design effectively reduces interdependence among system components, significantly enhancing system maintainability and scalability.

The platform integrates multiple functional modules, including global crop condition monitoring, multi-stress identification, crop area monitoring models, and crop yield estimation using remote sensing techniques. It also offers global agricultural remote sensing data mapping services and online programming capabilities, thereby lowering the technical threshold for users to build customized agricultural monitoring systems.

Fig B.1. CropWatch cloud service framework based on APIs

Through the API interfaces of the CropWatch platform, users can freely define their area of interest (AOI) and the major cereal and oilseed types within it, and flexibly configure agricultural meteorological conditions, crop phenology, and physiological parameters. Users can independently select data sources, monitoring indicators, model types, and time ranges to lightweight and rapidly construct customized crop monitoring systems (Fig. B.2).

This mechanism breaks away from the centralized and closed architecture of traditional agricultural monitoring systems. It enables users to independently conduct crop monitoring and product generation for specific regions, facilitating collaborative analysis, information sharing, and transparent publication. By dismantling the “information black box” of conventional systems, the platform significantly improves the transparency and traceability of agricultural information.

During the system customization process, users are not required to invest heavily in local computing infrastructure. Instead, they can leverage the dynamic cloud computing resources of platform for efficient crop conditions monitoring. The platform further supports edge computing extensions to meet local data processing needs and employs user level storage isolation to ensure privacy and data security. All data acquisition, processing, and analysis operations comply with relevant national laws and privacy protection standards, thereby safeguarding data sovereignty of users.

Fig B.2. CropWatch platform enabling user-defined crop classification and yield prediction. 

(a) Interactive monitoring of rice yield in the Nile Delta, 2025; (b) Interactive monitoring of crop types in typical regions of Western Europe, 2025;

(c) Customized monitoring of maize yield in Zambia, 2025; (d) Customized remote sensing monitoring of crop types in France, 2025.