Automated recognition and measurement based on three-dimensional point clouds to connect precast concrete components

Abstract

The connection between precast concrete (PC) components is vital during the construction of prefabricated structures. However, PC components are generally inspected and measured manually, which is laborious and error-prone. Therefore, a novel approach was developed for automated recognition and measurement of rebars, concrete, and sleeves of PC components based on three-dimensional point clouds. The proposed approach recognizes rebars, concrete planes, and the edge points of sleeve ports using a one-class classifier, a clustering algorithm, and an alpha-shape algorithm, respectively. Furthermore, boundary estimation was performed to determine the dimensions and positions of the subcomponents. The obtained experimental results validate that an average distance of 3.0 mm exists between the center of rebar cross-sections and their corresponding sleeve ports. The proposed approach provides a novel solution for automatically recognizing rebars, concrete, and sleeves of PC components. Furthermore, it can potentially control the quality of PC components and reduce the cost of construction.