Enhancing Personal Protective Equipment Compliance and Usage Accuracy on Construction Sites through YOLOv7 Detection

Frequent monitoring of adherence to personal protective equipment is crucial on construction sites to protect workers’ health and reduce potential risks in the workplace. Still, manually verifying the use of PPE during toolbox meetings is a lengthy process, complex, and susceptible to human errors. To enhance operational efficiency, mitigate risks, and ensure compliance with construction regulations, this study proposes utilizing artificial intelligence to detect PPE presence and its proper usage. The research focuses on developing a YOLOv7 model along with a web application for management monitoring and notification. The testing result shows that using the trained model with a training accuracy of $\mathbf{9 2 \%}$, the proposed system achieved high detection efficiency on personal protective equipment compliance and usage. Automating the detection and monitoring process allows the system to decrease the need for manual supervision, thus freeing resources for other tasks. Furthermore, it takes a proactive approach to addressing compliance issues, enhancing safety protocols, and reducing the chances of accidents and injuries.