How Satellite Imagery and AI can enable targeted solutions for Indian farmers

Agriculture is the cornerstone of the Indian economy, providing direct or indirect employment to more than 42% of the population and contributing a significant 18% to the country’s Gross Domestic Product (GDP).

Yet, thousands of smallholder farmers in India have, for long, faced challenges with land record accuracy and completeness.

While technology, artificial intelligence in particular, has been proving it can be applied to real-world problems and with immense success, can agriculture be next in line? Can AI have an on-ground impact for India’s agriculture sector, making lives better for millions of Indians in the process?


The answer is yes, and not in a far-fetched future but now. And that is exactly what was explored in this latest episode of Blueprint to Bharat, powered by Google.

Senior Economic Times journalist Sonia Singh was in conversation with Alok Talekar, Research Lead, Agriculture & Sustainability at Google DeepMind, and Aaryan Dangi, co-founder of Terrastack, a startup incubated out of IIT-Bombay that has built a unique rural land intelligence system.

“In this episode, we are going to hear how satellite imagery and AI are helping map out farm boundaries and crop health in real time,” Singh said.

The experts discussed how these AI-powered technologies are helping the agri ecosystem to make quicker decisions and build targeted solutions that benefit farmers.

For Google DeepMind, agriculture has been one of the most consequential focus areas for AI research

“And that is because agriculture in India impacts a huge chunk of the population,” Alok Talekar said. “We have now organised information at an individual field level, identifying every single field, water body, tree, and then monitoring those fields across time.”

Terrastack has built a spatial intelligence platform on the ALU and AMED APIs, and integrated land records, farm-level satellite data, crop activity, climate signals, and market data for nearly all the land under cultivation across India. This platform has reduced the need for physical field visits and enabled Terrastack to generate key insights with the APIs that can enable India’s agri value chain to make faster and more accurate decisions that benefit farmers.

Talekar says Google's mission was to organise the world's agricultural information and make it useful, given the vulnerabilities of the sector, among them extreme weather events. “The focus was India first, not India only.”

The actionable satellite imagery:

On a macro level, the agri insights powered by these models can help government agencies in formulating targeted policies, because policymaking till now has been arbitrarily on a broader level.

The agencies now will have actionable data, giving them a headway into individual farmer requirements. “The data is intended to benefit and build an ecosystem like Aadhaar has done for identity, and UPI has done for payments,” Talekar said, “Personalised policy making to support individual farmers through private or government organisations is now very much possible.”

This shift to field-level insights can also enable the ecosystem to build more targeted innovations and make decisions based on the conditions of individual farms rather than on aggregated estimates.

Talekar feels it is a matter of bringing dignity with inclusion, without bias or favour. “Bringing dignity to services that farmers receive is really, really important with everybody getting equitable treatment,” he said, stressing that the data hence generated through satellite imagery is enabling farmers to have equitable access.

Enabling climate resilience:

The agriculture sector is highly vulnerable to climate change. Erratic monsoons, lack of irrigation, extreme weather events, dry spells, heavy wet spells, and more: every such event has a direct bearing on the farmer’s chances of getting a beneficial yield.

AI-powered support now opens meaningful opportunities in developing resilience. The satellite monitoring is enabling us to identify certain important aspects about the farms, including how cultivation patterns are changing, and more.

“We are very focused on building capabilities that increase climate resilience,” Talekar said. “I think we need to build solutions both for the short term and the long term, as in how do we give immediate relief through data-driven decision-making.”

In the longer run, this intelligence can support planning around crop diversification, water management, and climate adaptation strategies.

With these promising AI-enabled avenues, Aaryan feels, the biggest player in the agricultural ecosystem is the government. “Central and the state governments, who need ways to be able to monitor, distribute subsidies, offer disaster management programs efficiently, at scale, and in a transparent manner.”

Making it possible:

Talking about it now seems natural. However, looking back and analysing the scale of the problem gives a complete picture of how immense the task at hand was. GoogleMind has now got down to individual fields, individual trees, and water sources.

“Land records can be messy. There are disputes. There's trying to correlate land records which go back to, I don't know, how many hundreds of years. It's messy. How did you think AI could actually solve this huge problem?” Singh posed the question to Talekar.

Talekar said that Google has two models, one focused on segmentation, identifying individual fields, trees, water bodies, and more-and is called an agricultural landscape.

“The second model is called agricultural monitoring and event detection. So it monitors what crops are growing, when it was sown, and when it was harvested. The first model has 15 years of history, pretty much since Google started collecting high-res images for our satellite view on Google Maps. And then the other model has seven-ish years of history,” Talekar said.

Aaryan stressed the importance of critical players like Google being present, enabling data infrastructure or AI infrastructure at scale. “Without players like Google, it becomes extremely difficult to go and reach the last mile, right?”

Click here to discover how Google is partnering with the ecosystem to ensure AI delivers real impact, safely and responsibly, to drive progress for every Indian.