CropGuard — Indian Crop Disease Detection & Treatment

AI-powered crop disease detection system for Indian agriculture. Upload a leaf image → get instant disease diagnosis with treatment recommendations specific to the Indian market.

🎯 Features

• Image-based disease detection — CNN model (MobileNetV2) identifies diseases from leaf photos
• Smart treatment recommendations — Organic + chemical solutions with Indian brand names and dosage per acre
• 42 diseases & pests covered — Rice, Wheat, Tomato, Potato, Cotton, Maize, Sugarcane
• Prediction history — Save and review past analyses (requires login)
• JWT authentication — Secure register/login system
• Demo mode — Works without a trained model (returns realistic sample results)

🛠 Tech Stack

Layer	Technology
Frontend	React 19, MUI, React Router, Axios
Backend	FastAPI, Motor (async MongoDB), PyJWT
Database	MongoDB
ML	PyTorch, MobileNetV2 (transfer learning)
Deploy	Docker, Docker Compose

📁 Project Structure

CropGuard AI/
├── Server/
│   ├── app/
│   │   ├── main.py          # FastAPI entry
│   │   ├── config.py        # Environment config
│   │   ├── database.py      # MongoDB connection
│   │   ├── auth.py          # JWT + bcrypt
│   │   ├── models.py        # Pydantic schemas
│   │   ├── routes/          # API endpoints
│   │   └── services/        # ML inference + disease data
│   ├── model/               # Trained model files
│   ├── Dockerfile
│   └── requirements.txt
├── Frontend/
│   ├── src/
│   │   ├── api/client.js    # Axios + JWT
│   │   ├── context/         # Auth state
│   │   ├── pages/           # Login, Register, History, DiseaseDetail
│   │   └── components/      # UI components
│   └── package.json
├── docker-compose.yml
└── README.md

🚀 Quick Start

Prerequisites

• Node.js 18+
• Python 3.10+
• MongoDB (local or Docker)

1. Start MongoDB

# Option A: Docker
docker run -d -p 27017:27017 --name cropguard-mongo mongo:7

# Option B: Local install
mongod --dbpath /data/db

2. Start Backend

cd Server
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

3. Start Frontend

cd Frontend
npm install
npm run dev

Open http://localhost:5173 in your browser.

Docker (Alternative)

docker-compose up -d

🌐 Frontend + Backend Deployment (Vercel)

If your frontend is deployed at https://crop-guard-ai-ebon.vercel.app, set backend CORS to allow it:

CORS_ORIGINS=https://crop-guard-ai-ebon.vercel.app

If you also want local frontend access during development, keep both:

CORS_ORIGINS=http://localhost:5173,https://crop-guard-ai-ebon.vercel.app

In Vercel project settings, set your frontend environment variable to your backend URL:

VITE_API_URL=https://<your-backend-domain>

Parallel Deployments + Version Sync

This repository includes a GitHub Actions workflow that deploys frontend and backend in parallel:

• Workflow: .github/workflows/vercel-parallel-deploy.yml
• Trigger: push to main or manual run
• Benefit: frontend/backend deploy at the same time and use the same commit SHA for release alignment

Required GitHub repository secrets:

• VERCEL_TOKEN
• VERCEL_ORG_ID
• VERCEL_FRONTEND_PROJECT_ID
• VERCEL_BACKEND_PROJECT_ID

Release sync behavior:

• Frontend injects VITE_APP_RELEASE=<git sha> at build time
• Backend injects APP_RELEASE=<git sha> at deploy time
• Frontend sends X-Client-Release header and checks /api/status release value
• Mismatch is logged as release.mismatch in runtime logs

Build Speed Guidance

• For faster builds, set Vercel Build Machine to a larger size in Project Settings → Build and Deployment.
• Build machine size is a Vercel account/project setting and cannot be forced from repository code.

📡 API Endpoints

Method	Endpoint	Auth	Description
GET	/	—	Health check
POST	/api/register	—	Create account
POST	/api/login	—	Get JWT token
POST	/api/predict	Optional	Upload image → disease prediction
GET	/api/history	Required	User's prediction history
GET	/api/diseases	—	List all diseases
GET	/api/diseases/{key}	—	Disease detail
GET	/api/crops	—	List supported crops

Sample Request — Predict

curl -X POST http://localhost:8000/api/predict \
  -H "Authorization: Bearer <token>" \
  -F "file=@leaf_photo.jpg"

Sample Response

{
  "crop_name": "Tomato",
  "disease_name": "Tomato Late Blight",
  "confidence": 92,
  "severity": "High",
  "spread_risk": "High",
  "description": "Late blight is the most destructive disease...",
  "organic_treatment": ["Spray Bordeaux mixture (1%)...", "..."],
  "chemical_treatment": ["Metalaxyl 8% + Mancozeb 64% WP (Ridomil Gold)...", "..."],
  "dosage": "Ridomil Gold: 500 g/acre in 200 litres...",
  "prevention": ["Use disease-free transplants...", "..."],
  "status": "Diseased"
}

🌾 Supported Crops & Diseases (42 Classes)

Crop	Diseases & Pests
Rice	Blast, Brown Spot, Bacterial Leaf Blight, Tungro
Wheat	Leaf Rust, Powdery Mildew, Flag Smut, Leaf Smut, Black Rust, Yellow Rust, Leaf Blight, Scab, Stem Fly, Aphid, Mite
Tomato	Late Blight, Early Blight, Leaf Curl Virus, Septoria
Potato	Late Blight, Early Blight
Cotton	Bacterial Blight, Leaf Curl, Anthracnose, American Bollworm, Bollworm, Pink Bollworm, Aphid, Whitefly, Mealy Bug, Thrips, Boll Rot, Red Bug, Wilt
Maize	Northern Leaf Blight, Common Rust, Gray Leaf Spot, Armyworm, Fall Armyworm, Ear Rot, Stem Borer
Sugarcane	Red Rot, Smut, Mosaic, Red Rust, Yellow Rust

📊 Dataset

• 20K+ Multi-Class Crop Disease Images (42 classes)
• Pre-split Train/Validation folders
• Trained with PyTorch MobileNetV2 transfer learning

📝 License

MIT