Soccerlytics-Advanced-Football-Video-Analysis-System

DESCRIPTION: Soccerlytics is an comprehensive football video analysis system that tracks and analyzes matches. It generates enhanced visualizations with player tracking, ball possession statistics, team assignments, and a unique 2D top-down view for tactical analysis.

🌟 Key Features

• *🔎 Player & Ball Tracking: Utilizes a state-of-the-art *YOLO model for high-accuracy detection of players, referees, and the ball. This is combined with the ByteTrack algorithm to assign unique IDs and track their movements seamlessly throughout the video.

• *🎨 Automatic Team Assignment: The system intelligently identifies team affiliations by analyzing player jersey colors. It uses *K-Means clustering on the pixel data from each player's torso to automatically group players into two distinct teams and color-code them.

• *⚽ Real-time Ball Possession: FootRec calculates ball possession by determining the closest player to the ball in every frame. This data is used to update a dynamic *possession bar on the screen, showing the percentage of control for each team as the match progresses.

• *🗺 2D Tactical View (Bird's-Eye View): A standout feature of the system. Using a *homography transformation, it maps the players' and ball's positions from the camera perspective onto a miniature 2D representation of the football pitch, offering a clear, tactical overview of formations and movements.

• *🎥 Camera Movement Compensation: To ensure positional data is accurate, the system employs the *Lucas-Kanade optical flow method. This estimates the camera's pan and tilt, allowing the system to differentiate between player movement and camera movement for more precise tracking.

• *⚡ Performance Metrics (Speed & Distance): The system estimates the *speed (in km/h) and total distance covered (in meters) for each player. These crucial performance statistics are calculated from the players' transformed positions and displayed in real-time.

• 💾 Data Caching for Efficiency: Computationally intensive tasks like object tracking and camera movement estimation can be saved to pickle files (stubs). This allows for rapid re-analysis and debugging without needing to re-process the entire video from scratch.

---

💻 Tech Stack

• Python
• OpenCV for video processing and computer vision tasks.
• Ultralytics YOLO for object detection.
• Supervision for tracking and annotation utilities.
• NumPy & Pandas for numerical operations and data handling.
• Scikit-learn for K-Means clustering.
• Matplotlib for visualization.

---

🚀 Getting Started

Follow these instructions to get a copy of the project up and running on your local machine.

Prerequisites

• Python 3.9+
• Git
• A trained YOLO model file (e.g., best.pt).
• A video of a football match (e.g., test1.mp4).

Installation

1. Clone the repository:

   git clone [https://github.com/your-username/FootRec.git](https://github.com/your-username/FootRec.git)
   cd FootRec
   
   

2. Create a virtual environment (recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   
   

3. Install the required dependencies: (You may need to create a requirements.txt file based on the imports in the script)

   pip install opencv-python numpy pandas ultralytics supervision scikit-learn matplotlib
   
   

4. Set up project structure:
   • Create a models/ directory and place your trained best.pt file inside it.
   • Create an invd/ directory and place your input video file (e.g., test1.mp4) inside it.

Usage

To run the analysis, simply execute the main script from the root directory:

python main_final.py