Automated Data and Video Processing in Track Cycling

• Autores: Thomas Vleeschouwers, Robbe Decorte, Maarten Slembrouck, Steven Verstockt
• Localización: Journal of Science & Cycling: Breakthroughs in Cycling & Triathlon Sciences, ISSN-e 2254-7053, Vol. 14, Nº. Extra 2, 2025 (Ejemplar dedicado a: Special Issue of the 2025 Science & Cycling Congress | Lille - June)
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • Recent advances in sports data analytics have driven the development of integrated performance monitoring systems. This abstract describes the implementation and uses cases of integrating a pan-tilt-zoom (PTZ) camera into the existing Wireless Cycling Network (WCN) deployed at the Wielercentrum Eddy Merckx in Gent, Belgium. WCN combines a precise timing system, ANT+ sensor data collection and an anaerobic capacity (W’) estimation. Automated PTZ camera tracking and recording enables immediate visual analysis by coaches. AI-computer vision models are used for detection and qualitative analysis of team pursuit changeovers. In addition to elite use cases, general engagement can be improved with event-specific real-time data display and personalized videos and session reports. These integrated capabilities and the modular design streamline coaching workflows and provide a platform for future applications across various track-based sports. Future work includes further development of analysis tools to support coaches, production deployment of the complete system and implementing it at additional tracks.