제목 The Scheme of Detecting and Tracking Curling Stones for Curling Records Using Multiple Cameras
저자명

Junghu Kim, Jaehyun So, Sanghun Han and Youngjoon Han

초   록

 Recently, sports recording technologies have been improved through the use of information and communication technologies (ICT), making it possible for many sports to be more quickly and accurately recorded. Since curling, also known as “chess on ice,” is a highly strategic sport, it is appropriate for the application of ICT approaches. However, recent applications for curling records rarely use real-time recording technologies, and instead require manual recording of the positions and traces of stones. With this in mind, this paper proposes a new scheme for detecting and tracking curling stones. Our proposed method is able to automatically record curling play by processing images from multiple cameras. It uses a mean-shift algorithm and Kalman filter to quickly and correctly detect and track curling stones based on input images. The mean-shift algorithm uses sampled optimal colors to robustly track curling stone images. As a result, the average frame rate of the proposed method is 56.22fps, which is 12.42fps superior to existing methods. In addition to it, the proposed method is 1.97px more accurate.

Key Words curling, stone tracking, mean-shift tracking, optimal color information, multiple cameras
원문 수록처 Proceedings of the International Association of Computer Science in Sport Conference, pp 8-13. june 22-24 2014
자료유형 International Conference