AI-based geometric quality inspection of construction projects with point clouds

The quality of construction projects directly impacts public safety. Poor quality leads to defects, failures, and risks, necessitating frequent and accurate quality control, especially for structural members of buildings. This thesis aims to propose an end-to-end geometric quality inspection framework that encompasses structural member detection within point clouds, compliance checks based on relevant regulations, and digitalized documentation using digital twin platforms. The proposed framework aims to be generalizable to different types of structural members, including beams, columns, and bracings, across a wide range of construction projects that minimize the need for manual interventions.