Back to projects

Baseball Detection Using Computer Vision Techniques

  • Supervisor:Hee Choel Kim
  • Status: Completed

Baseball Detection Using Computer Vision Techniques, project leverages state-of-the-art deep learning models to achieve high-precision tracking and recognition of baseballs during gameplay. Utilizing the YOLO (You Only Look Once) version 11 framework, renowned for its speed and accuracy in object detection, our system achieves an impressive 95% accuracy in detecting baseballs in real-time. This project addresses challenges in fast-paced sports environments where accurate and rapid ball detection is critical for performance analysis, gameplay monitoring, and fan engagement. By integrating advanced computer vision techniques, our solution ensures consistent tracking even under varying lighting conditions, complex player movements, and occlusions.

Our model is designed to support applications such as automated umpiring, training analytics, and broadcast enhancements, providing valuable insights to players, coaches, and audiences. With the promise of scalability and adaptability, this system exemplifies the potential of AI-driven sports technologies to revolutionize traditional methodologies.

This was an 2 year real time collaborational project with industry.