Access the live project here:
🔗 https://savior-team.netlify.app/
This project presents an intelligent, real-time monitoring system designed for cross-border defense surveillance, focusing on the detection, comparison, and inventory management of weapons, vehicles, and other strategic assets. Using satellite imagery and field-level images, the system performs semantic segmentation and visual comparison to identify structural or inventory changes and updates them into a central database.
It integrates computer vision, cloud databases, and a user-friendly dashboard for defense analysts to receive alerts, perform image comparisons, and manage cross-border inventory data.
- 🔄 Image Comparison & Change Detection: Compare satellite and real-time field images using OpenCV + Deep Learning (UNet-based segmentation).
- 🚧 Object Detection & Inventory Classification: Detect and classify weapons, vehicles, or infrastructure elements from images.
- 🧠 Semantic Segmentation: Apply UNet++ models from the Jupyter notebook (
segmentation.ipynb) to extract masked segments of interest. - ☁️ Firebase Integration: Real-time database and user authentication via Firebase.
- 🌐 React.js Dashboard: Clean, modular UI for accessing detections and updating inventories.
- 🌍 MongoDB Atlas: Secure, cloud-native database to store and retrieve inventory entries and image metadata.
| Component | Technology |
|---|---|
| Frontend | React.js |
| Backend | Node.js (optional), Firebase |
| Image Analysis | OpenCV, UNet++ (PyTorch) |
| Realtime DB | Firebase Realtime Database |
| Cloud Storage | Firebase Storage / MongoDB Atlas |
| Model Training | Python, Jupyter Notebook |