Dr. Ritvik Sahajpal, a NASA Harvest team member and GEOG researcher, gave an invited talk at a “Mapping the Future of Agriculture: Technology and Citizen Science for International Crop Monitoring” co-organized by the World-Wide Human Geography Data Working Group (WWHGD) and the U.S. Department of Agriculture Foreign Agriculture Service on 15 January 2020 at the U.S. Geological Survey in Reston. The WWHGD is co-led by the National Geospatial-Intelligence Agency and the U.S. Department of State, the WWHGD Working Group seeks to build voluntary partnerships across the Human Geography data and mapping communities with the goal of promoting Human Security.
In his talk, Dr. Sahajpal explained how machine-learning (ML) algorithms can reduce time and effort required to monitor crop performance from local to global scales by describing two NASA Harvest projects. The Field Data Collection Optimization project in collaboration with Swiss Re, which focuses on learning how to use ML-based techniques to reduce the field data collection costs while maintaining the capacity and accurately predicting crop yields and conditions. Results from applying this model in Ukraine indicate that the number of data samples required can be reduced by as much as 22% while maintaining model performance. Dr. Sahajpal provided an overview of the Forecasting Crop Yields project to describe how ML models are used on Earth Observation dataset to forecast crop yields for major commodity crop producers.