USE OF UNMANNED AERIAL VEHICLES AND COMPUTER VISION IN CONSTRUCTION SAFETY INSPECTIONS

Abstract

Construction is unarguably one of the most dangerous industries whereby many hazardous tasks and conditions exist, which may pose injuries, risks and fatalities to the workers. Hence, safety inspections are continuously carried out to maintain a safe environment. These typically involve safety officers who circulate around the construction site to detect unsafe working conditions, and ensure compliance with health and safety regulations. However, the task of efficiently supervising a large number of workers and consistently identifying all possible violations is still considered manual and tedious. Therefore, this paper takes the initial steps and presents work targeted at automating the safety inspection process using Unmanned Aerial Vehicles (UAVs). These are commonly known as drones and are small, aerial camera-equipped robots capable of rapidly visualizing spacious environments. More specifically, in this study, a UAV system is used to capture real-time videos from a construction site. The videos are then streamed to a central automated system and analyzed using digital image processing techniques to check whether construction workers are wearing personal protective equipment (PPE), in particular hard hats. The components of the proposed system were created and preliminary results highlighted the potential of using camera-equipped UAVs and computer vision to automate safety inspections in construction environments.