AI-Based Road Debris and Garbage Detection System

Overview

This project implements an AI-powered deep learning solution to detect debris, garbage, and construction materials that are temporarily or permanently placed on roads in India. The system combines a Python-based computer vision backend with a modern React TypeScript frontend to provide real-time detection capabilities and an intuitive web interface for monitoring and alert management.

Problem Statement

In India, it's common for construction materials such as bricks, sand, soil, and stones to be temporarily stored on roadsides during building or wall construction projects. Additionally, bulk garbage is often disposed of or stored along roadways for extended periods. These obstructions:

• Create safety hazards and increase accident risks
• Make it difficult for pedestrians and vehicles to navigate roads
• Contribute to urban and rural cleanliness issues
• Require timely intervention from local authorities

Solution

Our AI-based detection system uses computer vision and deep learning to:

1. Automatically detect debris, garbage, and construction materials on roads
2. Classify different types of obstructions (construction debris, garbage, etc.)
3. Provide a web interface for real-time monitoring and visualization
4. Enable alert management through an intuitive dashboard
5. Support batch processing for large-scale monitoring

Project Structure

cd_ai/
├── backend/
│   ├── app.py                 # Main Flask/FastAPI application
│   ├── detector.py            # Core detection logic
│   ├── best.pt               # Trained model weights (YOLO)
│   ├── road.jpg              # Sample test image
│   ├── sample.mp4            # Sample test video
│   ├── 123.mp4               # Test video file
│   └── static/
│       └── uploads/          # Processed media files
│           ├── 123_detected.mp4
│           ├── sample_detected.mp4
│           └── ...
├── frontend/
│   ├── src/
│   │   ├── App.tsx           # Main React application
│   │   ├── main.tsx          # Application entry point
│   │   ├── index.css         # Global styles with Tailwind
│   │   ├── components/       # Reusable React components
│   │   │   ├── Loader.tsx
│   │   │   └── Navbar.tsx
│   │   └── pages/            # Application pages
│   ├── package.json          # Frontend dependencies
│   ├── vite.config.ts        # Vite configuration
│   ├── tailwind.config.js    # Tailwind CSS configuration
│   ├── tsconfig.json         # TypeScript configuration
│   └── index.html            # HTML template
└── README.md                 # This file

Features

• ✅ Real-time debris and garbage detection using YOLO model
• ✅ Multi-class object detection (construction materials, garbage)
• ✅ Custom dataset trained on Indian road conditions(rural and semi-urban areas)
• ✅ Modern React TypeScript frontend with Vite
• ✅ Responsive design with Tailwind CSS
• ✅ Video and image processing capabilities
• ✅ File upload and processing interface
• ✅ Real-time detection visualization
• ✅ Scalable architecture for deployment

Technical Stack

Backend

• Python with computer vision libraries
• YOLO model for object detection (best.pt)
• Flask/FastAPI for REST API
• OpenCV for image/video processing
• PyTorch for deep learning inference

Frontend

• React 18 with TypeScript
• Vite for fast development and building
• Tailwind CSS for responsive styling
• Lucide React for icons
• React Router DOM for navigation
• ESLint for code quality

Installation & Setup

Prerequisites

• Python 3.8+
• Node.js 16+
• GPU recommended for optimal performance

Backend Setup

# Clone the repository
git clone https://github.com/ranadebsaha/CB_AI
cd cd_ai/backend

# Install Python dependencies
pip install -r requirements.txt

# Run the backend server
python app.py

Frontend Setup

# Navigate to frontend directory
cd frontend

# Install dependencies
npm install

# Start development server
npm run dev

# Build for production
npm run build

Usage

Web Interface

1. Start both backend and frontend servers
2. Navigate to http://localhost:5173 (or configured port)
3. Use the web interface to:
   • Upload images or videos for detection
   • View real-time detection results
   • Monitor processed files
   • Access detection history

API Endpoints

# Upload and process image/video
POST /api/detect
Content-Type: multipart/form-data

# Get processed results
GET /api/results

# Health check
GET /api/health

Model Information

The system uses a custom-trained YOLO model (best.pt) specifically optimized for detecting:

• Construction debris: Bricks, sand, stones, building materials
• Bulk garbage: Waste materials, trash piles
• Mixed obstructions: Combined debris types

Model Performance

• Trained on Indian road conditions
• Optimized for various lighting conditions
• Real-time processing capabilities
• High accuracy for debris classification

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/improvement)
3. Make your changes
4. Test both frontend and backend
5. Commit changes (git commit -am 'Add feature')
6. Push to branch (git push origin feature/improvement)
7. Create Pull Request

Future Enhancements

Technical Improvements

• Mobile app development (React Native)
• Advanced analytics dashboard
• User authentication and roles
• Database integration for detection history
• GPS integration for location tracking

AI/ML Enhancements

• Model accuracy improvements
• Edge device deployment
• Multi-model ensemble
• Custom training pipeline
• Data augmentation techniques

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Roboflow for dataset management platform
• YOLO community for object detection framework
• React and Vite communities for frontend tools
• Tailwind CSS for styling framework
• Local authorities and communities for data collection support

---

Note: This project is developed to serve society by improving road safety conditions in India. We encourage responsible use and welcome community participation in making roads safer for everyone.