Digifarm: Detecting the world’s highest accuracy field boundaries to power precision agriculture
- /
- POC
- /
- Digifarm: Detecting the world’s...
DigiFarm uses Artificial Intelligence technology to detect accurate field boundaries and seeded acres for precision agriculture. They developed deep neural network models, achieving detection accuracies above 96% on 15+ million hectares.
Duration in months: 12
Problem Description
DigiFarm's solution optimizes agricultural operations, reduces seasonal uncertainty, and minimizes production costs, increasing crop revenue and enabling carbon capture, with pilot results implemented in corporate partners' commercial FMS-solutions.
Goals
New services
Challenges
DigiFarm developed deep neural network models for detecting field boundaries using Sentinel-2 satellite imagery. The model has been validated on 15+ million hectares. The project included training for entire-country regions like Germany, Austria, Belgium, and the UK.
Innovation results
Digifarm utilized EOSC services to train a deep neural network model for accurate agricultural field boundary detection using Sentinel-2 at 1m per pixel resolution Satellite Earth Observation data. The results were achievied using HPC and GPUs, coupled with internal team expertise.
Business impact
DigiFarm successfully delineated field boundaries and seeded acres across AOIs like Germany, Austria, Belgium, and the UK, achieving a targeted accuracy of 0.94+ across all regions, a 10-12% higher accuracy than LPIS Cadastral data.