Computer Vision-Based Security System for Monitoring The Use of Personal Protective Equipment (PPE) In Workplace Involving Production Machinery

Abstract

Occupational safety is crucial as workers' lives are directly or indirectly influenced by safety policies in the workplace. One significant cause of workplace accidents is the failure of workers to wear personal protective equipment (PPE). Advancements in technology now enable PPE monitoring through computer vision. This research aims to develop a computer vision-based security system to address the negligence of wearing PPE in work environments with production machinery. If the system detects that a worker is not wearing the required PPE, it issues a warning and shuts down the production machine. The system employs computer vision algorithms, specifically SSD MobileNet V2 FPNLite 320x320 and SSD MobileNet V2 FPNLite 640x640 models, trained to identify individuals, helmets, safety suits, and gloves. The system successfully detects the specified PPE, issues warnings, and deactivates the production machine, achieving the best accuracy and F1 score values of 0.538 and 0.479, respectively.