Tracking System for Ultimate Frisbee
- 1 Devlogs
- 7 Total hours
Tracks players using CV and hardware for advance analytics
Tracks players using CV and hardware for advance analytics
Primary use:
Currently for ultimate frisbee for specific real time player maps as well as post advance analytics.
Main systems:
Computer vision and mini wearbles
Computer vision currently uses a small trained YOLO11 based model to detect players.
Current dataset:
YOLO11 Player model + 800+ labelled specific
Currently working on:
Implementing custom ReID system using motion prediction and OSNet
Goal:
To create a single board computer that connects to a phone/ camera to provide real time analysis and analytics.
Attached:
A mini preview of the system, purely using the model and bytrack to map players.