AI Crop Health & Precision Agriculture

Manual crop-disease identification and fertiliser dosing caused roughly 35% avoidable seed waste and 25–30% fertiliser overuse.

I trained a multi-class CNN on 1,500+ field images, reaching 96%+ disease-detection accuracy and about 92% seed-classification accuracy.

Soil analysis (pH, NPK), a fertiliser calculator and weather-API modules cut fertiliser overuse by 25–30%.

The Django farmer portal includes an agri-marketplace with 50+ products and is live across 3+ farming deployments.