Personal Protective Equipment Detection for Grinding Machine Workers Based on Computer Vision

Abstract

Negligence in the use of Personal Protective Equipment for workers is one factor the occurrence of work accidents, especially in manufacturing industries such as grinding process. There has been research carried out to develop it Personal Protective Equipment detection system, but there is still no research specifically in the grinding machine area and there is no system that can take it. decisions from results that have been detected so that no action has been taken directly from worker negligence. In an effort to improve the safety of the workers and reduce the risk of work accidents, especially for workers grinding machine, the author in this final project created a Detection System Design Negligence in the Use of Personal Protective Equipment for Based Grinding Machine Workers Computer Vision using the You Only Look Once v5s detection model (YOLOv5s). This system is able to identify grinding machine workers wearing earmuffs, face shields, masks, gloves, and those who do not use or no earmuff, no face shield, no mask, and no gloves. From overall testing obtained (Frame Per Second) the highest FPS of 2.25 FPS from the image size 256, at an optimal distance of 3 meters with 88% accuracy and light intensity of 600 lux with 76% accuracy run using NVIDIA Jetson Nano 2GB and Logitech C270 webcam.