README.md

CropGuard GH 🌿

Image-Based Crop Disease Detection System for Smallholder Farmers in Ghana. Final Year Project — Oppong David, BTech Computer Technology, Kumasi Technical University.

Two ways to run it

1. Instant / no install — open cropguard.html on any phone or browser. Covers all 14 Ghanaian crops via the model; works offline as an estimate (English + Twi). Can also connect to the trained model — see Settings / docs.

2. Full production system — the real Chapter 3 stack:

   • backend/train.py — train the MobileNetV2 model
   • backend/app.py — FastAPI inference server
   • backend/recommendations.json — treatment knowledge base
   • frontend/src/CropGuard.jsx — React frontend

Quick start (backend)

cd backend
pip install -r requirements.txt
python train.py --data ./data          # train (needs a dataset)
uvicorn app:app --port 8000            # serve

Full details, API reference and report mapping: docs/DOCUMENTATION.md.