Project Overview

Objective: To detect and track players, referees, and footballs in video footage using advanced AI techniques.

Key Components

Object Detection and Tracking

• Utilize YOLO (You Only Look Once), a state-of-the-art object detection model, to identify and track players, referees, and footballs in video frames.
• Train the YOLO model to enhance detection accuracy and performance.

Player Team Assignment

• Implement K-means clustering for pixel segmentation to differentiate players based on the colors of their t-shirts.
• Use the segmented data to assign players to their respective teams.

Ball Acquisition Measurement

• Calculate each team’s ball acquisition percentage throughout the match using the player and ball tracking data.

Camera Movement Analysis

• Apply optical flow techniques to analyze and measure camera movement between frames for accurate player movement tracking.

Perspective Transformation

• Implement perspective transformation to accurately represent the scene’s depth and perspective.
• Convert player movement measurements from pixels to meters for a more realistic assessment.

Speed and Distance Calculation

• Determine players' speed and the distance covered based on the transformed perspective and movement data.

Requirements

To run this project, you need to have the following requirements installed:

• Python 3.x
• ultralytics
• supervision
• OpenCV
• NumPy
• Matplotlib
• Pandas