IDinsight partners with Radiant Earth Foundation to create high-quality ground-truthed baseline geospatial data for Machine Learning models in agriculture

January 18, 2022 – IDinsight’s Data on Demand team is collaborating with Radiant Earth Foundation to create ground-truthed baseline training data for Machine Learning models in agriculture, through high-quality data collection of crop types and field boundaries of agricultural plots across four states (Bihar, Odisha, Rajasthan, and Uttar Pradesh) in India. IDinsight, a data analytics, impact measurement, and advisory organization will support Radiant Earth Foundation, which works with open machine learning and Earth observation data, to improve agricultural monitoring practices.

Without accurate agriculture data, farmers aren’t able to estimate their yields, which affects the supply and demand of nutritious food locally, regionally, and globally. When farmers are better able to monitor their crops, they can adjust their practices as weather patterns change and other exterior forces impact their yield. Machine Learning techniques are increasingly being applied to satellite imagery (or Earth observations, EO), to detect issues in crops and crop yields. To create the algorithms that will detect patterns and lead to recommendations for farmers, crop-related data (aka labels) must be collected, curated, and prepared. 

IDinsight will use its Data on Demand (DoD) team to collect high-quality data on crop types and field boundaries of agricultural plots across Bihar, Odisha, Rajasthan, and Uttar Pradesh in India. The DoD team is working with its network of local surveyors to collect geospatial data on approximately 10,000 agricultural plots (and additional data on field coordinates, crop types, agricultural inputs, production). Radiant Earth Foundation will use this ground reference data to generate a training dataset and develop a baseline machine model focused on field boundary detection. All of these will be publicly available on Radiant MLHub to benefit agriculture agencies and practitioners.

IDinsight will also build a public-facing knowledge base on the methods and protocols to collect high-quality ground-truthing data. This includes detailed field protocols as well as technical outputs on comparing the accuracy of the various methods of geospatial data collection (like Garmin eTrex devices, in-built SurveyCTO GPS, and app-based geospatial data).

“We are truly excited to work with Radiant Earth on this project,” said Krishanu Chakraborty, IDinsight Associate Director. “This will help us both develop our internal capabilities to collect high-quality ground-truthing data and also contribute to the global learning agenda using machine learning models in agriculture.”

“We are excited to expand our partnership with IDinsight and collaborate on diversifying agricultural training data in India,” says Hamed Alemohammad, Executive Director and Chief Data Scientist of Radiant Earth Foundation. “The planned benchmark training dataset we’re creating will empower practitioners and decision-makers across the agriculture sector to deploy local solutions and enhance their data-driven policymaking. It is a notion that is central to our mission.”

This project is supported by Enabling Crop Analytics at Scale (ECAAS), an initiative managed by Tetra Tech with funding from the Bill & Melinda Gates Foundation.

About IDinsight

IDinsight is a mission-driven global advisory, data analytics, and research organization that helps global development leaders maximize their social impact. We tailor a wide range of data and evidence tools, including randomized evaluations and machine learning, to help decision-makers design effective programs and rigorously test what works to support communities. We work with governments, multilateral agencies, foundations, and innovative non-profit organizations in Asia and Africa. We work across a wide range of sectors, including agriculture, education, health, governance, sanitation, and financial inclusion. We have team members who are remote and have offices in Dakar, Lusaka, Manila, Nairobi, New Delhi, and Rabat.

www.idinsight.org

About Radiant Earth Foundation

Radiant Earth Foundation is a non-profit organization actively working to develop Earth observation machine learning libraries and models through an open-source hub that supports global missions like agriculture, conservation, and climate change. Radiant Earth also fosters a community of practice to develop standards around machine learning for Earth observation and provide information on the progress and innovation in the Earth observation marketplace.

www.radiant.earth