<![CDATA[Occupational accidents represent a major challenge in the construction and manufacturing industries. This study aims to develop a deep learning model for real-time detection of personal protective equipment (PPE) usage using the YOLOv5 algorithm. Utilizing a dataset that includes four classes (hardhat, no hardhat, coverall, and no coverall), the model was trained and evaluated based on precision, recall, and mean Average Precision (mAP) metrics. The results demonstrated that the model achieved a high accuracy level with an mAP of 0.91 and stable performance. The model can also rapidly and effectively detect safety attributes even in complex work environments, such as varied lighting conditions and numerous background objects. Based on usability testing results of 85.35% and satisfactory black box testing, this study produced a prototype web-based application enabling efficient and effective PPE monitoring. The application is designed to support the improvement of workplace safety across various industrial sectors in a more practical and adaptive manner. It is expected to increase PPE compliance, reduce accident risks, and contribute significantly to workplace safety in the industry. The conclusion indicates that the YOLOv5 algorithm holds great potential for implementation in technology-based safety monitoring systems and supports the development of a safer and more modern industry.]]>
Copyrights © 2025
