🌱 TerraDetect – A Smart Agriculture Solution

🔗 Live Demo: https://terradetect.onrender.com

---

TerraDetect is a comprehensive smart farming platform that combines real-time soil monitoring, IoT sensor integration, and ML-powered crop & fertilizer recommendations. The system features a Go backend with ONNX inference, React Native mobile app with Expo, and ESP32 microcontrollers for environmental data collection.

---

🔧 Key Features

📱 Mobile App (React Native + Expo)

• � Modern Material Design 3 UI with dark mode support
• 📊 Real-time sensor data visualization with telemetry cards
• 🌾 ML-powered crop, fertilizer, and suitability predictions
• 📈 Historical data tracking and analysis
• � User authentication with JWT tokens
• 🎭 Guest Mode - Try predictions without registration
• � Offline support with request queuing
• � Cross-platform (iOS & Android)

�️ Backend (Go + MongoDB)

• ⚡ High-performance REST API with Gin framework
• 🧠 ONNX Runtime for ML inference (crop & fertilizer models)
• 🔐 JWT authentication with refresh tokens
• 👥 User and device management
• 📊 Sensor data storage and retrieval
• 🎭 Guest token system for anonymous predictions
• 🔒 Rate limiting and security middleware
• 📡 CORS support for cross-origin requests

🔌 IoT Integration (ESP32)

• 📡 WiFi connectivity with dynamic configuration
• 🔐 Device ID authentication (format: AB1234)
• 📈 HTTPS sensor data uploads with API key validation
• 🌿 Multi-sensor support:
  • Soil pH
  • Moisture
  • Temperature & Humidity
  • EC (Electrical Conductivity)
  • NPK (Nitrogen, Phosphorus, Potassium)

🤖 Machine Learning

• 🌾 Crop recommendation (23 crop types)
• 🧪 Fertilizer suggestion (7 fertilizer types)
• 📊 Crop suitability analysis
• ⚡ ONNX models for fast inference
• 🎯 Confidence scores for predictions

---

📁 Project Structure

TerraDetect/
├── backend/              # Go backend server
│   ├── handlers/         # API route handlers
│   ├── middleware/       # Auth, rate limiting, CORS
│   ├── models/           # Data models (User, Device, Sensor)
│   ├── inference/        # ONNX ML inference engine
│   ├── db/               # MongoDB connection
│   └── main.go           # Server entry point
├── mobile/               # React Native mobile app
│   ├── app/              # Expo Router pages
│   │   ├── (app)/        # Authenticated screens
│   │   └── (auth)/       # Auth screens (login, register, guest)
│   ├── components/       # Reusable UI components
│   ├── store/            # Zustand state management
│   ├── lib/              # API client & utilities
│   └── assets/           # Images, icons, fonts
├── ml/                   # Machine learning models
│   ├── crop-model.onnx   # Crop recommendation model
│   ├── fertilizer-model.onnx  # Fertilizer suggestion model
│   └── *.py              # Model training & conversion scripts
├── sketches/             # ESP32 Arduino firmware
│   └── esp32_terradetect.ino
└── docs/                 # API documentation

---

🚀 Getting Started

🧰 Prerequisites

• Backend: Go 1.21+, MongoDB 4.4+
• Mobile: Node.js 18+, npm/yarn, Expo CLI
• IoT: ESP32-WROOM-32, Arduino IDE or PlatformIO
• ML: Python 3.9+ (for model training/conversion)

⚙️ Backend Setup

# Navigate to backend directory
cd backend

# Install Go dependencies
go mod download

# Set up environment variables
cp .env.example .env
# Edit .env with your MongoDB URI and JWT secret

# Run the server
go run .

The backend will start on http://localhost:8080

📱 Mobile App Setup

# Navigate to mobile directory
cd mobile

# Install dependencies
npm install

# Set up environment variables
cp .env.example .env
# Edit .env with your API URL

# Start Expo development server
npx expo start

# Scan QR code with Expo Go app (iOS/Android)
# Or press 'a' for Android emulator, 'i' for iOS simulator

📡 ESP32 Setup

1. Open sketches/esp32_terradetect.ino in Arduino IDE
2. Install required libraries:
   • WiFiManager
   • HTTPClient
   • ArduinoJson
3. Copy secrets.h.template to secrets.h and configure:
   • API endpoint URL
   • Device ID
   • API key
4. Upload to ESP32 board
5. On first boot, connect to ESP32 AP to configure WiFi

🤖 ML Model Training (Optional)

# Navigate to ml directory
cd ml

# Train models (if needed)
python train_crop_model.py
python train_fertilizer_model.py

# Convert to ONNX format
python export_onnx.py

# Validate ONNX models
python validate_onnx.py

---

🔐 API Endpoints

Authentication

• POST /api/v1/auth/register - Register new user
• POST /api/v1/auth/login - User login
• POST /api/v1/auth/guest - Get guest token (30 min)
• POST /api/v1/auth/refresh - Refresh access token
• POST /api/v1/auth/logout - Logout user

Sensor Data (Authenticated)

• GET /api/v1/sensor/latest - Get latest sensor reading
• GET /api/v1/sensor/history - Get historical data (paginated)

Predictions (Authenticated)

• POST /api/v1/predict/crop - Crop recommendation
• POST /api/v1/predict/fertilizer - Fertilizer suggestion
• POST /api/v1/predict/suitability - Crop suitability analysis

Guest Predictions (No Auth Required)

• POST /api/v1/guest/predict/crop - Guest crop prediction
• POST /api/v1/guest/predict/fertilizer - Guest fertilizer prediction
• POST /api/v1/guest/predict/suitability - Guest suitability analysis

Device Management

• POST /api/v1/device/check - Check device availability
• POST /api/esp32 - ESP32 sensor data upload (API key required)

See docs/api.md for detailed API documentation.

---

🔒 Security Features

• 🔑 JWT-based authentication with access & refresh tokens
• 🎭 Guest tokens with limited permissions (30-minute expiry)
• 🔐 API key validation for ESP32 devices
• 🛡️ Rate limiting on all endpoints
• 🚫 CORS protection with configurable origins
• 🔒 Password hashing with bcrypt
• 📊 User-specific data isolation
• 🧠 Device identity validation (format: AB1234)

---

🧑‍🌾 Use Cases

• 🌾 Farmers: Real-time soil monitoring and crop recommendations
• 🎓 Students: Learning platform for Agri-IoT and ML applications
• 🔬 Researchers: Field data collection and analysis
• 🏢 Agribusinesses: Precision agriculture and resource optimization
• 🌱 Home Gardeners: Smart gardening with data-driven insights
• 👥 Consultants: Try predictions as guest without registration

---

🎯 Key Improvements in v2.0

• ✅ Complete rewrite from Flask to Go for better performance
• ✅ Native mobile app with React Native + Expo
• ✅ Guest mode for trying predictions without registration
• ✅ ONNX inference for faster ML predictions
• ✅ Real-time telemetry dashboard with live sensor data
• ✅ Offline support with request queuing
• ✅ Material Design 3 UI with dark mode
• ✅ Enhanced security with JWT refresh tokens
• ✅ Better error handling and user feedback
• ✅ Comprehensive API documentation

---

🖼️ Preview

---

⚙️ Tech Stack

Layer	Technologies
Mobile App	React Native, Expo, TypeScript, NativeWind
Backend	Go, Gin, MongoDB, ONNX Runtime
ML Models	Python, Scikit-learn, ONNX
IoT Hardware	ESP32, RS485 NPK sensor, analog sensors
Auth	JWT, bcrypt, guest tokens
State Mgmt	Zustand (mobile)
Styling	TailwindCSS (NativeWind), Material Design 3
API	REST, JSON, CORS

---

📬 Contact

Developed by Gagan Ahlawat and Team TerraDetect 🔗 https://terradetect.onrender.com

Contributions, issues, and feature requests are welcome!

---