Sony x AI AgTech Challenge
Automated Weed Finding for Precision Agriculture
Sony x AI AgTech Challenge
May 19, 2025
Growers need better targeting than blanket spraying, but field-ready sensing has to be cost-effective and practical.
Our team researched grower pain points and found pesticide usage and rising costs were widespread issues, then built a Raspberry Pi camera system that connects detection output to a weed-density heat map.
It won first place at the Sony x AI AgTech Challenge among 15+ other teams.
It showed how lightweight computer vision could help growers focus treatment where it's needed instead of treating a whole field the same way.
Highlights
- Built and quantized a computer vision model to fit on a Raspberry Pi.
- Connected detection results to a heat-map data pipeline.
- Paired technical prototyping with grower viability research.