A real-time AI system that uses CCTV video streams to detect, classify, and estimate waste in public areas using deep learning and maps the locations for smart waste management.
Vision2Clean AI is developed as a project for the subject "Waste to Wealth to Wheels".
It uses YOLOv11 and computer vision to:
- Detect types of waste in real-time from mobile cameras acting as CCTV
- Estimate the waste quantity using bounding box size
- Tag waste with GPS (pre-mapped for each camera)
- Display detections on a live interactive map
- 🧠 YOLOv11-based waste detection
- 🗂️ Supports multiple waste categories:
- Plastic, Organic, Paper, Metal, E-waste, etc.
- 🛰️ Location tagging via static GPS coordinates
- 🗺️ Map visualization using Folium
- 💾 Logging to CSV / database
- 📲 Uses mobile phone cameras for live feed
| Component | Tool/Library |
|---|---|
| Detection Model | YOLOv11 (Ultralytics) |
| Live Feed | IP Webcam App / OpenCV |
| Map View | Folium (Python) |
| Data Storage | CSV / SQLite |
| Dashboard | React js |
git clone https://github.com/yourusername/vision2clean-ai.git
cd vision2clean-ai