Automated Generation of Inspection Reports for Construction Operations

Abstract

This paper introduces a newly developed framework for the automated generation of inspection reports in construction. The method utilizes earlier published developments by the authors on object recognition and localization. Inspection reports in this paper are documents that monitor and record the progress of installed project components and their targeted locations. Components in this research are also referred to as objects, which include installed piping systems, tanks, and mechanical equipment. The reports identify the deviations of the installed components by comparing their as-built status to the asplanned ones. The framework integrates the data acquired by a Real-Time Location System (RTLS) and a computer vision-based LiDar technology to generate the reports using object identification and localization. While considerable work has been reported using such technologies, the contribution of the framework presented here lies in the efficient integration of these technologies to acquire as-built 3D coordinates of the installed components. The framework has been validated through laboratory experiments, demonstrating an accuracy of approximately 27 centimetres for the installations' coordinates in the inspection reports. The framework presented here can be used for project commissioning to ascertain the precise locations of the project's installations.