Velocity and stroke rate reconstruction of canoe sprint team boats based on panned and zoomed video recordings

TL;DR

This paper develops an automated video analysis framework to accurately reconstruct velocity and stroke rate in canoe sprint, providing a GPS-level performance assessment without sensors.

Contribution

It introduces a novel, comprehensive video-based method utilizing YOLOv8, homography estimation, and pose analysis for performance metrics reconstruction in canoe sprint.

Findings

Achieved velocity MAPE of 0.011 and stroke rate MAPE of 0.009 compared to GPS data.

Demonstrated high correlation with GPS measurements (Spearman rho > 0.97).

Provided a fully automated, sensor-free analysis tool for coaches.

Abstract

Pacing strategies, defined by velocity and stroke rate profiles, are essential for peak performance in canoe sprint. While GPS is the gold standard for analysis, its limited availability necessitates automated video-based solutions. This paper presents an extended framework for reconstructing performance metrics from panned and zoomed video recordings across all sprint disciplines (K1-K4, C1-C2) and distances (200m-500m). Our method utilizes YOLOv8 for buoy and athlete detection, leveraging the known buoy grid to estimate homographies. We generalized the estimation of the boat position by means of learning a boat-specific athlete offset using a U-net based boat tip calibration. Further, we implement a robust tracking scheme using optical flow to adapt to multi-athlete boat types. Finally, we introduce methods to extract stroke rate information from either pose estimations or the athlete bounding boxes themselves. Evaluation against GPS data from elite competitions yields a velocity MAPE of 0.011 [0.008 0.014] (Spearman rho=0.974) and a stroke rate MAPE of 0.009 [0.006 0.013] (Spearman rho = 0.975). The methods provide coaches with highly accurate, automated feedback with minimal manual initialization work required, and without requiring sensors.