SkillAgentSearch skills...

SpinVision

SpinVision is a computer vision project that analyzes the motion of a cricket ball in video footage. It detects the ball's trajectory, predicts its post-bounce path assuming no spin, and visualizes the differences between the actual and predicted motions to estimate the effect of spin.

Install / Use

/learn @abhijitshukla/SpinVision
About this skill

Quality Score

0/100

Supported Platforms

Universal

README

SpinVision

A computer vision-based project that detects a cricket ball's motion in a video, predicts its post-bounce trajectory assuming no spin, and visualizes the difference between actual and predicted motion.

Output Preview


🔍 Features

  • Detects cricket ball position frame-by-frame using a YOLOv8 model.
  • Identifies the bounce point based on vertical motion.
  • Simulates post-bounce trajectory without spin.
  • Compares actual vs predicted motion to estimate spin deviation.
  • Generates an overlay video with trajectory visualization.

🗂️ Project Structure

.
├── coordinates/
│   ├── coordinates.txt               # Detected coordinates
│   └── coordinates_no_spin.txt       # Predicted trajectory
├── model/
│   └── cricket_ball_detector.pt      # YOLOv8 model
├── videos/
│   └── kohli.mp4                     # Input video
├── output/
│   ├── overlay_output.mp4            # Annotated output video
│   └── snapshot_with_spin_angle.png  # Snapshot with trajectory and angle
├── predict.py                        # Predict trajectory without spin
├── coord.py                          # Detect coordinates using YOLO
├── overlay.py                        # Visualize and compare trajectories
├── requirements.txt                  # Python dependencies
└── README.md                         # Project documentation

🚀 Getting Started

1. Install Requirements

pip install -r requirements.txt

2. Run Coordinate Detection

python coord.py

3. Run Trajectory Prediction

python predict.py

4. Generate Overlay Video

python overlay.py

📊 Output Example

The overlay video shows:

  • Green trajectory: Actual motion (with spin).
  • Red trajectory: Predicted motion (without spin).
  • Spin Angle Deviation label post-bounce.

📌 Notes

  • Make sure your YOLO model path and video file path are correctly set.
  • Modify frame dimensions if using a video of different resolution.

🔗 Model Download

Download the YOLOv8 cricket ball detector model from this link:
Download cricket_ball_detector.pt Place it in the model/


📄 License

MIT License

View on GitHub
GitHub Stars14
CategoryContent
Updated7mo ago
Forks4

Languages

Python

Security Score

82/100

Audited on Aug 21, 2025

No findings