Implementation of a convolutional network for detection of PPE in automotive repair services

This study presents a novel application of computer vision technology to enhance safety in the automotive service industry, with a specific focus on the real-time detection of Personal Protective Equipment (PPE) usage. Using a YOLOv8 convolutional neural network, this study analyzes approximately 2,000 high-resolution images from security cameras within automotive repair facilities to detect compliance with safety gear requirements. These environments are characterized by constant artificial lighting and frequent interactions among employees and customers, which introduces complexity in visual data processing. The YOLOv8 model was selected for its proficiency in real-time object detection, facilitating robust and efficient analysis while ensuring that only authorized personnel perform specific tasks. Initial results indicate successful uniform detection, though challenges remain in accurately identifying smaller items like safety glasses and gloves. This underscores the potential of advanced image recognition technologies to substantially enhance workplace safety and operational efficiency across various industries. This study not only demonstrates the system's effectiveness in PPE detection but also sets a foundation for future advancements in smart safety practices within the industrial sector.