Loading…
Google announced on 26 August 2026 that its India-first agricultural AI models are now generating insights for six African nations: Kenya, Uganda, Ghana, Rwanda, Nigeria and Zambia.
The two models, built by Google DeepMind's AnthroKrishi team, are Agricultural Landscape Understanding (ALU) and Agricultural Monitoring and Event Detection (AMED).
ALU maps agricultural field boundaries and land patterns from satellite imagery; AMED tracks dynamic changes such as sudden crop stress.
Their outputs are available as APIs and have been added to Google Earth as a data layer.
Indian users include TerraStack, an IIT Bombay-incubated startup that has mapped over 140 million hectares of farmland, and Telangana's Agriculture Data Exchange (ADeX).
Satellite images alone do not tell you where one farm ends and the next begins. ALU draws those field boundaries automatically, and AMED then watches each field over time for changes such as crop stress.
Simple Analogy: One model draws the plot map; the other keeps watch over it.
Google's artificial intelligence research division; its AnthroKrishi team built ALU and AMED
UN agency for food and agriculture; its geoAI4stats initiative received USD 2.5 million from Google.org to integrate these models globally
State-run data exchange that integrated the model outputs to support digital agriculture services for over 5 million farmers
| Feature | ALU | AMED |
|---|---|---|
| Full form | Agricultural Landscape Understanding | Agricultural Monitoring and Event Detection |
| What it does | Maps field boundaries, land patterns, landscape characteristics | Detects dynamic changes and events such as sudden crop stress |
| Data refresh | About every 6 months | About every 15 days |
| Nature of output | Static structure of the landscape | Time-sensitive signals over that landscape |
IIT Bombay-incubated startup that built a spatial intelligence platform on the ALU and AMED APIs, covering over 140 million hectares of Indian farmland
Combined the model outputs with weather data to manage 2.6 million hectares of irrigated area
Uses the ALU API with Gemini to advise on low-carbon rice cultivation, targeting 2 million hectares by 2030
Hosts the ALU output as a public data layer, one of the most widely used layers on the platform
GS Paper 3 > Science and Technology and Agriculture > IT and computers, e-technology for farmers
General Awareness > Science and Technology Current Affairs
Agricultural Landscape Understanding - AI model that maps agricultural field boundaries and land patterns from satellite data
Agricultural Monitoring and Event Detection - AI model that tracks changes such as sudden crop stress across fields
Application Programming Interface, through which software systems exchange data automatically