Multi-Sensor Based Analysis of the Changeovers in Team Pursuit

Authors

  • Robbe Decorte
  • Steven Verstockt Ghent University-imec, IDLab
  • Maarten Slembrouck

Keywords:

Team Pursuit, Track Cycling, Sports Data Science, Computer Vision, Machine Learning

Abstract

More and more data is getting collected in sports. Also on the cycling track, a lot of rider data is available to be analyzed, such as timing data from measurement loops, sensor data of the cyclists’ wearables and video data recorded by coaches. During BOSA 2024, the Becoming Outstanding in Sports Analytics winter school in Ghent, our participants were asked to develop a multimodal team pursuit analysis demo that can facilitate the work of the track cycling coaches of Cycling Vlaanderen. The proposed multimodal data analysis give coaches insights into how well the changeover between riders was performed by visualizing the key features of the change: the riding lines of the riders, the duration of the change, and the trends of the power data (which was collected using the WCN Wireless Cycling Network set-up). The video clips, in which the riding lines are visualized, are automatically generated using a change detection method that analyzes the Mylaps measurement loops data. Next, state-of-the-art object detection and tracking methods are used to show each rider’s trajectory. For each changeover, we visualize the results in an interactive dashboard. These visualizations should provide coaches the necessary insights to better coach and improve their athletes.

Downloads

Download data is not yet available.

Downloads

Published

2024-08-11

How to Cite

Decorte, R., Verstockt, S., & Slembrouck, M. (2024). Multi-Sensor Based Analysis of the Changeovers in Team Pursuit. Journal of Science and Cycling, 13(2), 33-38. https://jsc-journal.com/index.php/JSC/article/view/905

Most read articles by the same author(s)