Golf Ball Tracker

A golf ball tracking compass system that helps golfers locate their ball after hitting it. The system uses computer vision to track golf ball trajectories and provides directional guidance via a web interface.

Features

• Real-time golf ball detection and tracking
• Trajectory prediction for errant shots
• Web-based compass interface
• Works on Mac (prototyping) and Raspberry Pi (production)

Setup

Mac (Development)

1. Install dependencies:

pip install -r requirements.txt

2. Copy environment configuration:

cp .env.example .env

3. Adjust .env settings if needed (camera index, thresholds, etc.)

4. Run the application:

python src/main.py

5. Open your browser to http://localhost:5000

Raspberry Pi

1. Install system dependencies:

sudo apt-get update
sudo apt-get install -y python3-pip python3-opencv libopencv-dev

2. Install Python dependencies:

pip3 install -r requirements.txt

3. Configure .env for Raspberry Pi camera:

CAMERA_INDEX=0  # or use /dev/video0

4. Run the application:

python3 src/main.py

5. Access from your phone/tablet at http://<raspberry-pi-ip>:5000

Project Structure

golfeye/
├── src/
│   ├── detection/      # Ball detection modules
│   ├── tracking/       # Ball tracking and trajectory prediction
│   ├── compass/        # Direction calculation
│   ├── web/            # Web server and frontend
│   ├── camera/         # Camera abstraction
│   └── main.py         # Main application entry point
├── models/             # ML models (optional)
├── tests/              # Unit and integration tests
└── requirements.txt    # Python dependencies

Configuration

See .env.example for all configuration options. Key settings:

• CAMERA_INDEX: Camera device index (0 for default)
• FRAME_SKIP: Process every Nth frame (higher = better performance)
• LAUNCH_VELOCITY_THRESHOLD: Minimum velocity to detect ball launch
• WEB_PORT: Web server port

Usage

1. Position the camera on your golf cart to view the golfer
2. Start the application
3. Open the web interface on your phone/tablet
4. Hit your golf ball
5. The compass will guide you to the predicted landing location

Development

The system uses a hybrid detection approach:

1. Color-based detection (fast, good for white balls)
2. ML verification (optional, more accurate)
3. Kalman filtering for smooth tracking
4. Physics-based trajectory prediction

License

MIT