Badminton AI Vision

A computer vision system for analyzing badminton matches using AI-powered pose detection and court calibration to track player movements, distances, and speeds.

Features

• Pose Detection: High-precision skeletal tracking using YOLOv8/v11
• Court Calibration: Maps pixels to real-world coordinates for accurate measurements
• Distance Tracking: Calculates total movement distance for each player
• Speed Analysis: Real-time measurement of player instantaneous speed
• ROI Processing: Focus detection on relevant court areas
• Player Identification: Automatically distinguishes between players
• Visualization: Color-coded skeleton rendering with customizable display

Requirements

• Python 3.8+
• Dependencies listed in requirements.txt

Installation

# Clone the repository
git clone https://github.com/AnInsomniacy/badminton-ai-vision.git
cd badminton-ai-vision

# Create virtual environment
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Create directories for videos
mkdir -p raw_videos processed_videos

Note: You'll need to place a YOLOv8/v11 pose model in the models/ directory.

Usage

1. Court Calibration

Calibrate the court coordinates for accurate distance measurements:

python court_calibration/court_calibrator.py

2. Process Video

python main.py

3. Configuration

Edit the CONFIG dictionary in main.py to customize analysis settings:

CONFIG = {
    "input_path": "raw_videos/YourMatch.mp4",
    "model_path": "models/yolo11x-pose.mlpackage",
    "enable_distance_tracking": True,
    "enable_speed_tracking": True,
    "person_aliases": {
        "upper": "PLAYER 1",
        "lower": "PLAYER 2"
    }
}

Project Structure

badminton-ai-vision/
├── court_calibration/          # Court calibration tools
│   ├── court_calibrator.py
│   └── ROISelector.py
├── processor/                  # Core processing modules
│   ├── video_processor.py
│   ├── image_processor.py
│   ├── model_processor.py
│   ├── distance_processor.py
│   ├── speed_processor.py
│   └── mask_processor.py
├── models/                     # YOLO model storage
├── raw_videos/                 # Input videos
├── processed_videos/           # Output videos
├── main.py                     # Main application
├── renderer.py                 # Visualization utilities
└── requirements.txt

Key Configuration Options

Setting	Description
model_path	Path to YOLO pose model
confidence_threshold	Detection confidence (0.0-1.0)
input_path	Path to input video file
enable_distance_tracking	Enable distance measurement
enable_speed_tracking	Enable speed analysis
roi	Region of interest coordinates
person_aliases	Custom player names

Acknowledgments

• Ultralytics YOLO
• OpenCV