Plane-based registration of construction laser scans with 3D/4D building models

With the development of building information modelling (BIM) and terrestrial laser scanning (TLS) in the architecture, engineering, construction and facility management (AEC/FM) industry, the registration of site laser scans and project 3D (BIM) models in a common coordinate system is becoming critical to effective project control. The co-registration of 3D datasets is normally performed in two steps: coarse registration followed by fine registration. Focusing on the coarse registration, model-scan registration has been well investigated in the past, but it is shown in this article that the context of the AEC/FM industry presents specific (1) constraints that make fully-automated registration very complex and often ill-posed, and (2) advantages that can be leveraged to develop simpler yet effective registration methods.This paper thus presents a novel semi-automated plane-based registration system for coarse registration of laser scanned 3D point clouds with project 3D models in the context of the AEC/FM industry. The system is based on the extraction of planes from the laser scanned point cloud and project 3D/4D model. Planes are automatically extracted from the 3D/4D model. For the point cloud data, two methods are investigated. The first one is fully automated, and the second is a semi-automated but effective one-click RANSAC-supported extraction method. In both cases, planes are then manually but intuitively matched by the user. Experiments, which compare the proposed system to software packages commonly used in the AEC/FM industry, demonstrate that at least as good registration quality can be achieved by the proposed system, in a simpler and faster way. It is concluded that, in the AEC/FM context, the proposed plane-based registration system is a compelling alternative to standard point-based registration techniques.     

Applications of 3D point cloud data in the construction industry: A fifteen-year review from 2004 to 2018

3D point cloud data obtained from laser scans, images, and videos are able to provide accurate and fast records of the 3D geometries of construction-related objects. Thus, the construction industry has been using point cloud data for a variety of purposes including 3D model reconstruction, geometry quality inspection, construction progress tracking, etc. Although a number of studies have been reported on applying point cloud data for the construction industry in the recent decades, there has not been any systematic review that summaries these applications and points out the research gaps and future research directions. This paper, therefore, aims to provide a thorough review on the applications of 3D point cloud data in the construction industry and to provide recommendations on future research directions in this area. A total of 197 research papers were collected in this study through a two-fold literature search, which were published within a fifteen-year period from 2004 to 2018. Based on the collected papers, applications of 3D point cloud data in the construction industry are reviewed according to three categories including (1) 3D model reconstruction, (2) geometry quality inspection, and (3) other applications. Following the literature review, this paper discusses on the acquisition and processing of point cloud data, particularly focusing on how to properly perform data acquisition and processing to fulfill the needs of the intended construction applications. Specifically, the determination of required point cloud data quality and the determination of data acquisition parameters are discussed with regard to data acquisition, and the extraction and utilization of semantic information and the platforms for data visualization and processing are discussed with regard to data processing. Based on the review of applications and the following discussions, research gaps and future research directions are recommended including (1) application-oriented data acquisition, (2) semantic enrichment for as-is BIM, (3) geometry quality inspection in fabrication phase, and (4) real-time visualization and processing.     

A CNN-based 3D patch registration approach for integrating sequential models in support of progress monitoring

Significant advancements in three-dimensional (3D) imaging technologies have enabled the ability to effectively monitor and manage the progress of works in construction. Traditionally, 3D point clouds have been used in conjunction with building information models to visualize the progress of works. The discrepancies between ‘as-planned’ and ‘actual’ models are unable to be automatically identified using the existing approaches due the absence of an effective registration algorithm. To ensure the registration accuracy of multi-scanned point clouds, an automated method based on a data-driven Convolutional Neural Network (CNN) deep learning algorithm is proposed. In this instance, 3D Point cloud patches are aligned with spatial datasets that are scanned from different locations using range cameras. The registration results are used to automatically detect spatial changes when compared with different point clouds. The quantified changes are utilized to determine the percentage of work that has been completed at fixed intervals. The developed registration approach is tested and validated using a series of experiments. It is demonstrated that discrepancies between ‘as-planned’ and ‘actual’ models can be identified with a higher level of accuracy, which can enable the baseline for monitoring construction to be undertaken in real-time.