A Deep Learning Model for Detecting PPE to Minimize Risk at Construction Sites

The prominent reasons for accidents occurring on construction sites are traumatic brain injuries, collisions, being stuck by objects and sometimes these lead to death. Personal Protective Equipment (PPE) is one of the safety measures used to mollify those. In this paper, a deep learning model is developed for detection of PPE ignorance and the presence of unauthorized persons which will help security officers to take the right action to reduce the risk at the construction site. Also, this automated deep learning model alerts the security officers when workers or anyone entering the construction site without PPE with the help of face recognition technology. The experimental results showed 0.95 Mean Average Precision.