The Power of Public Research: USDA Brings AI to Agriculture

Making autonomous technology affordable and widely available to American farmers is the target of a new endeavor by USDA’s Agricultural Research Service. Until now, a primary obstacle to this goal was the absence of large, publicly available datasets of high-quality, annotated plant images that can train computer vision models.

Enter Dr. Steven Mirsky, a USDA-ARS researcher (and GROW co-founder), who has recently launched the Digital Agricultural Systems Hub (DASH), a new USDA platform that will help plant breeders and scientists turn their research and findings into real-world, deployable AI solutions for farmers.  Co-leaders of the system include USDA Computational Biologist Dr. Amanda Hulse-Kemp and NC State Plant Scientist Dr. Chris Reberg-Horton.

The first tools of this new platform include AgIR – the national Open-Access Plant Image Repository – built by public sector scientists around the country, and PlantMap3D, a modular camera system designed for both plant phenotyping and on-the-ground deployment on farming equipment for real-time biomass mapping and plant ID.

Mirsky was recently featured in a TEDx Talk, where he explained why USDA is well situated to take on the challenge of leading agriculture into the digital revolution. Watch it below!



Text and feature photo by Emily Unglesbee, GROW; banner image by Ubaldo Torres, Texas A&M