SkyFusion - Satellite Object Detection Web Application

Features

• 🛰️ Satellite-Themed UI: Beautiful space-inspired interface with animations
• 🎯 Advanced Object Detection: Detects Aircraft, Ships, and Vehicles
• 🖱️ Drag & Drop Upload: Easy image upload with preview
• ⚙️ Adjustable Confidence: Real-time confidence threshold adjustment
• 📊 Detailed Results: Visual detection results with statistics
• 📱 Responsive Design: Works on desktop and mobile devices

Installation

1. Clone or download the project files
2. Install Python dependencies:
   pip install -r requirements.txt
3. Ensure the model file skyfusion_detector.pth is in the root directory

Usage

1. Start the Flask server:
   python app.py
2. Open your web browser and go to:
   http://localhost:5000
3. Upload a satellite image and analyze it:
   • Drag & drop an image or click to browse
   • Adjust confidence threshold (0.1 - 0.95)
   • Click "Analyze Image"
   • View results with bounding boxes and object counts

Model Information

• Architecture: Faster R-CNN with ResNet-50 backbone
• Classes: Aircraft, Ship, Vehicle (+ background)
• Input: Satellite/aerial imagery
• Output: Bounding boxes with class labels and confidence scores

File Structure

project/
├── app.py                    # Flask backend application
├── templates/
│   └── index.html           # Frontend interface
├── skyfusion_detector.pth   # Trained PyTorch model
├── requirements.txt         # Python dependencies
├── uploads/                 # Temporary upload directory
└── README.md               # This file

API Endpoints

• GET /: Main web interface
• POST /predict: Image analysis endpoint
• GET /health: Health check and model status

Technical Details

• Backend: Flask with PyTorch inference
• Frontend: HTML5, CSS3, JavaScript
• Model Format: PyTorch state dict with class information
• Image Processing: PIL for image handling, OpenCV for visualization
• Supported Formats: JPEG, PNG, JPG (max 16MB)

Browser Compatibility

• Chrome 60+
• Firefox 55+
• Safari 12+
• Edge 79+

Performance

• Inference Speed: Depends on hardware (CPU/GPU)
• Memory Usage: ~2GB RAM recommended
• File Size Limit: 16MB per image

Troubleshooting

1. Model not loading: Ensure skyfusion_detector.pth is in the correct directory
2. CUDA errors: The app automatically falls back to CPU if GPU is unavailable
3. Memory issues: Reduce image size or restart the application
4. Upload failures: Check file format and size (max 16MB)

Security Notes

• Files are temporarily stored during processing
• No permanent storage of uploaded images
• Input validation for file types and sizes
• CORS protection enabled

Development

1. Model Retraining: Update the model file and class names in app.py
2. UI Changes: Modify templates/index.html
3. API Extensions: Add new routes in app.py

License