An Environmentally Friendly Race Timing System Based on Cameras Only

Existing race timing solutions can be costly, time-consuming, or environmentally unfriendly. This paper proposes a novel video processing-based approach to race timing. A stereoscopic overhead camera is used to determine the moment of an athlete crossing the finish line using distance measurement. Ordinary RGB cameras are used to detect athletes and their racing bibs, and additional algorithms are employed for tracking and reidentification of athletes. The system achieves a high recall rate for athlete detection and matching detections to runners. However, the algorithm for athlete tracking based on path prediction needs improvement, as it currently isn’t well suitable for crowded areas and thus affects the identification result negatively. Further testing on the reidentification algorithm shows promising results. Overall, the proposed system has the potential to provide a reliable, environmentally friendly and cost-effective solution for race timing.