This repository is a clone of Ultralytics' open-source research into future object detection methods, namely Yolo V5. It could be used for helmet detection on persons.
detect.py runs inference on a variety of sources, using the weights stored in weights/best.pt. The function detect within detect.py returns 3 outputs, <image details>, <detection details> and <time of inference>. The first two output parameters are strings and the last is a floating point number.
To see a running WebApp on Binder, click the following below and you can do helmet and person detection on your own images!
For more comprehensive object detection, head to the Yolo V5 repo from Ultralytics.