Industrial and construction sites are always prone to safety-related issues because of dangerous equipment, labor movement, and a lack of standardization in the use of personal protective equipment (PPE). In addition, workforce attendance monitoring is still carried out using traditional paper-based attendance systems or biometric systems that are prone to proxy attendance, system delays, and maintenance problems. The current system addresses safety monitoring and attendance monitoring as two separate issues, leading to disjointed workforce management and poor decision-making. This paper proposes a computer vision-based intelligent monitoring system that integrates real-time PPE use monitoring and automated workforce attendance monitoring using facial recognition. The proposed system uses the YOLOv8 deep learning-based object detection model for helmet and safety vest detection from CCTV cameras and facial embedding recognition models for worker identification verification. By addressing both issues in a single system, the proposed system allows automated workforce attendance monitoring and real-time safety usage monitoring without any additional hardware infrastructure. The system is intended to be implemented on top of the existing surveillance systems and has the capability to process live camera feeds, video files, and multi-camera systems. The experimental results carried out on the simulated industrial environment show high accuracy and real-time processing capabilities, making it suitable for Industry 4.0 environments. The proposed system encourages transparency in the workplace, minimizes safety incidents, prevents attendance fraud, and maximizes efficiency through scalability and best data practices.

Computer Vision Industrial Safety Monitoring PPE Detection YOLOv8 Facial Recognition Workforce Attendance Automation Deep Learning Smart Construction Sites Industry 4.0.