Real-time construction site layout and equipment monitoring

A construction site is a dynamically changing environment, whose layout continually changes in each process phase. Failing to consider the changes in site layout can cause serious safety issues, and result in extra cost and time for the project. Site layout planning (off-site) and site monitoring (on-site) are two major research topics that help address these problems. However, the gap between available data off-site and on-site must be bridged regularly to update and re-plan the site layout as project progresses. This paper presents a real-time construction site layout and equipment monitoring system using building information modeling (BIM). Site layout and equipment data are collected and synchronized with a BIM database, from which the layout planning is then conducted. In this paper, the first part of the system, an on-site monitoring system, is developed. This system can identify an equipment on a construction site, such as an excavator or a truck, with a camera system, and utilizes a vision-based pose estimation method to determine the position and the orientation of the object. The scale-invariant feature transform (SIFT) method is used to recognize the object on the construction site. Then, the point cloud is generated from object SIFT feature points. Finally, the position and the orientation of the object is estimated from point cloud by the viewpoint feature histogram (VFH) method. Real construction site images with an excavator and a truck are used in the validation experiments. The preliminary test results showed that this method is capable of performing recognition and pose estimation in real-time; however, the accuracy of determining the correct orientation was found to be inadequate for high-precision applications, motivating the need for future work in improving the introduced pose estimation method.