Real-Time PPE Compliance Monitoring Using YOLOv8 with Integrated Video and Webcam Input

Workplace safety remains a major concern in industrial and construction zones, where ensuring that workers consistently wear Personal Protective Equipment (PPE) is essential to reduce the risk of injury. Manual monitoring methods are often inefficient, inconsistent, and prone to human error. In this work, we propose a lightweight, cost-effective PPE detection system that operates entirely on-premise, leveraging the capabilities of the YOLOv8 object detection model. The system is designed to function on commonly available computing devices such as personal laptops and desktop systems, without relying on expensive cloud-based solutions or specialized hardware like GPUs or edge accelerators. Through real-time processing of video streams—either live from a webcam or pre-recorded—the system identifies critical PPE elements, including helmets, vests, and gloves, using advanced computer vision and deep learning techniques. Unlike traditional cloud-deployed models that raise privacy and data security concerns, this approach ensures complete data residency and confidentiality, as all computations are performed locally. Additionally, the system maintains high detection accuracy and real-time performance even under varying lighting and background conditions commonly found in industrial environments. A modular software architecture ensures easy customization and scalability, making it suitable for small workshops as well as large-scale manufacturing units. Field evaluations confirm that the system reliably flags PPE violations and logs compliance data, allowing for automated reporting and actionable insights. This research underlines the potential of deploying AI-based safety systems on accessible hardware, democratizing safety compliance technologies for industries of all scales and budgets.