Combining Computer Vision and Drones for Proactive Construction Site Safety Monitoring

Worksite accidents have long been the leading cause of occupational injuries and fatalities worldwide, primarily due to two factors: the open and dynamic nature of the worksite environment and the inadequacy and incompetence of on-site safety managers. Recent advancements in deep learning (DL) and computer vision (CV) offer promising solutions to long-standing challenges in construction safety management. This paper proposes a proactive, real-time monitoring model for construction site safety, inspired by recent research integrating unmanned aerial vehicles (UAVs) with DL-based CV techniques. Specially designed data matrix (DM) tags were affixed to the safety helmets and vests of workers. The model captures DM-tagged images on-site and applies DL-based image recognition algorithms to assess individual risk levels, thereby enabling the implementation of preventive safety measures. Preliminary experimental results show that the model achieved a recall of 97.3% and a precision of 98.3% in worker identification. These findings highlight the practical potential of the proposed approach. The study concludes with a discussion on how the proposed approach could be applied to future advancements in construction safety management.