AI-Powered Workplace Safety: Helmet and Face Detection using YOLO for Access Control

The objective of this research is to create a computer vision-based system combining helmet wearing detection and worker identification for encouraging safety on construction and industrial workplaces. The existing method refers to CNN and sliding window techniques were utilized on a dataset of 5000 images by emphasizing precision, recall, and performance under controlled conditions. The proposed strategy refers to a method integrating YOLOv11 with Grassmann algorithms, emphasizing real-time performance and constraints such as occlusions, light conditions, and adaptability. The standard dataset permitted direct comparison, showcasing superior performance and capability in the method proposed. The outcome reflected the existing concept with an CNN of 89% accuracy was replaced with the proposed concept of 98% and placed at 0 .000 statistical significance level. It is observed that using YOLO v11 for facial recognition has significantly better gain than using CNN.