Sensor-driven recording and annotation of dry slope jumps
(Sensorgesteuerte Aufzeichnung und Annotation von Trockenhangsprüngen)
INTRODUCTION: In order to facilitate the work of coaches and to ease the creation of personalized stories (that athletes can easily share on their social media accounts) we are developing a sensor-driven methodology to automatically record and annotate jumps at the dry slope of Sport Vlaanderen in Genk, Flanders, Belgium 1.
METHODS: ANT+ sensor data of the athletes are automatically recorded over the entire dry slope using WASP devices 2. Currently, data and signal strengths (RSSI) of a running pod sensor is analyzed to predict when an athlete is active (performing a jump) or non-active - other sensors will be evaluated too in future work. Person detection/tracking on synchronized video is used to validate and further optimize the timestamps of the athlete shots3. Based on the video data, the landing is also mapped and classified as good or bad (depending on the distance to the centerline from where the athlete landed). Furthermore, the number of rotations is counted by means of pose estimation analysis and is used to cluster/classify the jumps.
RESULTS: During our first tests, 3 different riders were asked to jump in random order. Their jumps were correctly detected based on the multimodal sensor data/features and automatically annotated with their sensorIDs. The classification (good/bad landing, number of rotations) was also tested. The accuracy of this classification step, however, still needs to be improved and will be part of future work.
DISCUSSION/CONCLUSION: First results show the feasibility of the proposed approach. However, some remaining issues still need to be resolved to end up with an accurate and fully automatic system. The valorization potential of the proposed set-up is high - it could also be used in other contexts (e.g. snowparks) and different sports.
REFERENCES:
1 https://www.sport.vlaanderen/international/international/sport-vlaanderen-genk/
2 https://npe-inc.com/wasp-3/
3 Verstockt, S. et al., 2022. 13th World Congress of Performance Analysis of Sport 2022.
© Copyright 2023 9th International Congress on Science and Skiing, March 18 - 22, 2023, Saalbach-Hinterglemm, Austria. Veröffentlicht von University of Salzburg. Alle Rechte vorbehalten.
| Schlagworte: | |
|---|---|
| Notationen: | Trainingswissenschaft |
| Tagging: | Datenanalyse Landtraining |
| Veröffentlicht in: | 9th International Congress on Science and Skiing, March 18 - 22, 2023, Saalbach-Hinterglemm, Austria |
| Sprache: | Englisch |
| Veröffentlicht: |
Salzburg
University of Salzburg
2023
|
| Online-Zugang: | https://ski-science.org/fileadmin/user_upload/ICSS_2023_Book_of_Abstracts.pdf |
| Seiten: | 63-64 |
| Dokumentenarten: | Kongressband, Tagungsbericht |
| Level: | hoch |