Automatic Site Inspection System in Construction Sites (ICI–Intelligent Camera Inspection)

Abstract

Current practices of predicting the performance of a construction team require inspections that are still mainly manual, time-consuming, and can contain errors. In this paper, we proposed an automated site inspection and violation detection system that can be used in construction sites. Our proposed system uses deep learning SqueezeNet pre-trained network to detect two main violations: i) health & safety and ii) electrical hazards. The accuracy of our proposed system is 91.67% in detecting health & safety violations related to none-wearing personal protective equipment (PPE) while the accuracy of our proposed system reached 92.86% in detecting electrical hazards.