Han, S (2024) Understanding of AI-aided reality capture data for the development of BIM modeling for building façade inspections. PhD thesis, Marquette University, USA.
Abstract
The demand for building façade inspection continuously increases across the United States. Currently, an inspection is done manually during which an inspector must work in an unsafe environment. Recent advancements of drone, digital twin, and AI technologies make it possible to automate the building façade inspection operation and produce an editable as-is digital façade condition model for a better façade inspection as well as documentation.This dissertation presents the research results of the development of the Scan4Façade3DBIM, which is a fully automatic reality capture (RC) data analytics technology designed for as-is building façade condition and defect BIM modeling using drone photograph and photogrammetric point clouds. The workflow of Scan4Façade3DBIM starts with raw data preprocessing, followed by the cropped single building point cloud going through (building) Point Cloud Alignment, (wall plane) Point Cloud Separation, (wall plane) Point Cloud to (wall plane) Orthoimage, (wall plane section) Orthoimage Separation, (walls, elements and defections) Coordinate Conversion, and (walls, elements, and defections) 3D and BIM Modeling using the developed point cloud and image understanding algorithms, functions and tools. AI-aided image pixelwise segmentation is used to recognize the targeted façade elements and defections. The extracted walls, façade elements and defections, and their geometrical information are saved into a four-level data structure of Wall List, Wall Coordinates, Element Boxes and Defection Location and Dimension, and Irregular Element/Defection Counters in the optimized table-like formatted plain text files. In addition, Dynamo for Revit script is developed for automatic creating the building façades and defections as an editable Revit BIM model.Comprehensive case studies and experiments were conducted on different styles of buildings with 3D point clouds that were generated from different photogrammetry software with different drone-captured oblique photos. In addition, synthetic building points were used to analyze Scan4Façade3DBIM’s performance in terms of the point cloud quality (i.e., thickness). The results showed Scan4Façade3DBIM could accurately determine wall footprints even though the point cloud had a noise of 0.5 meters. The results of this research have advanced the knowledge of automatic generation of editable as-is façade+defect digital records from the RC data (i.e., 3D building point cloud and 2D defect oblique photos), which is a critical step to automate the building façade inspection using drone, digital twin, and AI technologies.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Thesis advisor: | Bai, Y |
| Uncontrolled Keywords: | noise; digital twin; United States; workflow; case study; façade; inspection; experiment |
| Index terms: | modelling, dissertation, dimension, inspection, conversion, case study, digital twin, data structure, workflow, script, point cloud, experiment, documentation, United States, thickness, drone |
| Subjects: | professional development, management, sociology, data collection methods, health monitoring assessment and metrics, data science, digital design, Geography, automation and robotics, manufacturing engineering, quality assurance, materials science, digital engineering, research dissemination and communication, analytical methods |
| Topics: | Quality Management, Health and Safety, Engineering Principles, Geographical Context, Digital Applications, Business Strategy, Information Management, Research Practice |
| Descriptive scope: | 3 PCE |
N.B. Descriptive scope is a count of how many of the five facets of empirical research are indicated by the words used in title, abstract and keywords. It is not intended as a judgement on the research; merely a count of the kind of word we would expect to indicate Phenomenon, Concepts, Theoretical framing, Empirical techniques, Analytical techniques. If all five are present, then a code of “5 PCTEA” will indicate this. If you feel the coding for this record is questionable, we welcome discussion around the terms we matched or the way we categorized them. The facet you would expect may not be coded, or a facet may be coded inappropriately. This can also bear on a larger question, of which facets should be treated as defining in construction management research. Please get in touch, and we will look at it. More details here