KInematikKanu

Project duration: 04/2021 – 03/2023

In elite sport, the continuous improvement of training and analysis methods is of enormous importance. The aim is, in particular, to make the best possible use of athletes’ individual potential to achieve optimal technical performance. To this end, video recordings are made in open water from a motorboat travelling alongside the athlete at a 90° angle, followed by a technical analysis of defined kinematic parameters (distances, angles). As part of these analyses, the first step is to select those images that correspond to specified phases within a paddle stroke cycle. Selected points on the athletes’ bodies and on the boat are then marked in these images, on the basis of which 2D kinematic parameters can be determined. This process requires considerable expertise as well as a high degree of manual intervention and is therefore extremely time-consuming.

Artificial intelligence (AI) methods offer enormous potential for automatically carrying out a wide variety of image and video analyses, thereby significantly speeding up the application of existing analysis methods on the one hand, and establishing entirely new methods on the other. In the KinematikKanu project, AI techniques are used to capture the expert knowledge of sports science researchers and coaches in machine-learning algorithms and, through integration into a software solution, make this available to a wide range of users for the individual analysis of sporting technique. As part of the project, the necessary algorithms are being developed, optimised and implemented. The aim is to establish state-of-the-art image processing technologies for training practice and, in future, to transfer these to elite youth sport.

This project is funded by research grants from the Federal Institute for Sports Science (BISp) following a resolution by the German Bundestag.

Federal Institute for Sports Science (BISp) | FKZ 2521BI1503

• German Canoe Federation e.V.
• Institute for Applied Training Sciences, Leipzig