Object-Detection applications For on-Site Construction Processes: A Review

Abstract

The advances in image capture technology have made it easier and more affordable to collect images and videos. Thus, most construction site are now equipped with surveillance cameras that continuously capture visual data which contain valuable information that can be used to assess the four main on-site construction processes: safety monitoring, progress monitoring, productivity monitoring and quality control. Computer vision algorithms and techniques are the most suitable tools to extract meanings (such as object location) from these data. The computer vision technique that has advanced the most and is the most used in the industry is object detection. This review aims to present the use of object detection to assess the four on-site construction processes and identify the challenges and opportunities that need to be addressed to correctly deploy and use this solution. Among the initial results there is a lack of research on object detection applied to quality control and the biggest challenges that object detection techniques face are the data acquisition, the construction environment and the limited hardware.