Sports_Performance_Analysis using OpenCV and Centroid Tracking

This project is a Python-based object tracking system that detects and tracks multiple moving people (or objects) in a video using OpenCV. It leverages a simple but effective Centroid Tracking Algorithm to assign unique IDs and compute trajectories and speeds for each detected object.

📽️ Demo

🚀 Features

• Background subtraction with MOG2
• Multi-object tracking with unique ID assignment
• Real-time trajectory plotting
• Speed estimation (in pixels/second)
• Automatic object removal if lost for too long

🧠 How It Works

1. Foreground Detection: Uses cv2.createBackgroundSubtractorMOG2 to detect moving objects.
2. Contour Detection: Filters out small blobs and calculates centroids.
3. Tracking: Maintains a mapping between object IDs and their centroids using a simple distance-based assignment.
4. Trajectory Drawing: Tracks the last 32 positions for each object and draws its path.
5. Speed Calculation: Computes object speed using frame rate and Euclidean distance between centroids.

🛠️ Requirements

• Python 3.6+
• OpenCV
• NumPy

Install dependencies

pip install opencv-python numpy