Skip to content

Latest commit

 

History

History
142 lines (108 loc) · 5.43 KB

README.md

File metadata and controls

142 lines (108 loc) · 5.43 KB

🌪️ DisasterEye

AI-Powered Satellite Damage Assessment & Relief Prioritization

🏆 Winner of SRM Techno Hackathon 2026 🏆

DisasterEye is an intelligent web application that analyzes satellite imagery before and after disasters to assess damage, prioritize relief efforts, and estimate resource requirements. It features advanced image alignment, interactive geospatial mapping, and interactive analytics.

🎯 Key Features

  • Interactive Maps (Folium): Search locations to view dynamic geospatial maps of affected areas right in your browser.
  • AI-Aligned Image Processing: Uses OpenCV's ORB Feature Extractor to perfectly align and stitch misaligned satellite images before executing damage subtraction.
  • Damage Analysis Heatmaps: Visual representation of damage severity and structural changes based on AI comparison.
  • Data Analytics Dashboards (Plotly): Beautifully visualised pie charts and bar graphs for zone priority counts and resource distribution.
  • Zone-Based Priority Assessment: Divide affected areas into 4x4 zones to easily pinpoint the most critical evacuation regions.
  • Impact Estimation: Calculate affected population, injured people, and financial economic losses instantly.
  • Scenario Simulation: Adjust rainfall and severity factors on the fly—the app uses @st.cache_data to instantly process changes without reloading.
  • Report Generation: Automatically generate and download detailed assessment reports.

🛠️ Tech Stack

  • Frontend: Streamlit (Python web framework)
  • Computer Vision: OpenCV (cv2) with Headless mode for cloud readiness.
  • Geomatics & Maps: Folium, Streamlit-Folium, OpenStreetMap API
  • Analytics & Visuals: Plotly, Pandas, Numpy
  • ML/AI Integration: TensorFlow 2.14

📋 Requirements

streamlit==1.28.0
tensorflow==2.14.0
opencv-python-headless==4.8.1.78
pillow==10.1.0
numpy==1.24.3
pandas==2.1.1
matplotlib==3.8.0
seaborn==0.13.0
folium==0.15.0
streamlit-folium==0.15.0
plotly==5.17.0
scikit-learn==1.3.1

🚀 Installation & Setup

  1. Clone the repository

    git clone https://github.com/sidsahoo1203/DisasterEye.git
cd DisasterEye

  2. Create a virtual environment

    python -m venv venv
venv\Scripts\activate  # On Windows
# or: source venv/bin/activate  # On macOS/Linux

  3. Install dependencies

    pip install -r requirements.txt

  4. Run the application

    streamlit run app.py

  5. Open in browser

    http://localhost:8501

☁️ Deployment Guide (Streamlit Community Cloud)

This project is configured perfectly for Streamlit's free hosting model because opencv-python-headless is used in the requirements.txt.

  1. Push your latest code to your GitHub repo.
  2. Visit share.streamlit.io and authorize with your GitHub account.
  3. Click "New App".
  4. Select sidsahoo1203/DisasterEye from your repository list, leave the branch as main, and enter app.py as your main file path.
  5. Click Deploy! Your app will install and run securely on a custom .streamlit.app URL.

📁 Project Structure

DisasterEye/
├── app.py                  # Main Streamlit application and AI logic
├── requirements.txt        # Cloud-ready Python dependencies
├── README.md               # This file
├── data/
│   ├── before/            # Pre-disaster satellite images
│   ├── after/             # Post-disaster satellite images
│   └── dataset/           # Dataset folder for training data
├── models/                # ML models (for future AI integration)
└── outputs/               # Generated reports and heatmaps

💡 How It Works

  1. Upload Images: Users upload 'before' and 'after' satellite images securely in Tab 1.
  2. Alignment Calibration: AI utilizes ORB algorithms to detect structural key points in both images and match them for seamless subtraction logic.
  3. Damage Detection: Compare structural variations using image processing differences in Tab 2.
  4. Zone Mapping Analysis: Generates high-res visual maps detailing damage concentration and priority action levels.
  5. Dashboard Calculation: Converts pixel damage variations into critical socio-economic figures like injuries, budget losses, and resource estimates in Tab 3.

📊 Priority Levels & Analytics

Color Damage % Priority Action
🟢 Green < 20% Low Monitor
🟡 Yellow 20-40% Medium Prepare Resources
🟠 Orange 40-70% High Deploy Medical Teams
🔴 Red > 70% Critical Evacuate Immediately

🚀 Future Enhancements

  • Real satellite image API integration (Sentinel-2, Landsat)
  • Deeper TensorFlow models for automated damage classification (Siamese networks)
  • Historical disaster comparison analysis
  • Real-time data feeds from meteorological agencies
  • Mobile app version

👥 Use Cases

  • Emergency Management: Rapid disaster assessment and response planning
  • Insurance Claims: Objective damage quantification
  • Government Agencies: Relief distribution planning (NDRF, DMCC)
  • NGOs: Resource allocation for disaster relief

📝 License

This project is developed for SRM Techno Hackathon 2026.

👨‍💻 Author

Created by sidsahoo1203 as part of the SRM Techno Hackathon 2026

🌪️ Turning satellite data into life-saving decisions