We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Motion trajectory tracking of athletes with improved depth information-based KCF tracking method.
- Authors
Zhang, Lina; Dai, Haidong
- Abstract
In this paper, a computer vision-based technique is proposed to track the motion trajectories of athletes to improve sport performance and deliver reliable physical data analysis. The moving target in the video was obtained through the Universal Background Eliminator (ViBe) moving target detection algorithm, and fused with the associated depth information based on the Kernerlized Correlation Filter (KCF) tracking algorithm. In order to better identify partly occluded objects, the depth information of the moving target and background segmentation threshold were acquired based on the depth distribution features within the original video tracking range. Meanwhile, the number of negative samples in the training sample set was reduced to prevent the model shifting problem caused by the background and obstacles. Experiments have been conducted on the dance trajectories of gymnasts as well as the motion trajectories of batting arms of badminton players. The results prove that the proposed algorithm can reduce the trajectory prediction error and improve the operating efficiency.
- Subjects
ARTIFICIAL satellite tracking; BADMINTON players; TRACKING algorithms; DANCE; TRACKING radar; GYMNASTS; DATA analysis
- Publication
Multimedia Tools & Applications, 2023, Vol 82, Issue 17, p26481
- ISSN
1380-7501
- Publication type
Article
- DOI
10.1007/s11042-023-14929-6