This repository contains the detection part of the Smart Football Table project. The position of the ball gets detected and published to the mqtt broker.
Visit the documentation at the Smart Football Table website for more information.
Argument | Description | Sample Input | default |
---|---|---|---|
-v | path to an (optional) video file, overwrites camera | "-v path/to/video.avi" | empty |
-b | length of lightning trace | 64 | 200 |
-i | index of camera | 0 | 0 |
-c | color values (hsvmin,hsvmax) for object you want to detect (unneccessary for yolo) | 20,100,100,30,255,255 | 0,0,0,0,0,0 |
-r | recording output into given file name | "fileName" | empty |
env var = MODEL_PATH -e MODEL_PATH=...