There is an increasing need to create accurate 3D as-built models of existing buildings for facility management and construction inspection purposes. To do so, photogrammetry and 3D laser scanning can be used to capture the geometric information and features of a building. However, the point cloud data generated from photogrammetry or laser scanning are heavy and noisy. Post-processing is needed and conversion to a lightweight 3D model is desirable. This project aims to develop methodology to clean and post-process point cloud data from photogrammetry and 3D laser scanning to reconstruct 3D BIM building models.
(1) Clean the point cloud data obtained from photogrammetry and/or 3D laser scanning. (2) Define the components to be modeled. (3) Extract features from the point cloud data using machine learning techniques. (4) Generate as-built 3D BIM models based on extracted feature values. (5) Evaluate the results.
(Students with computer programming background are preferred.)
(1) Students will be able to understand the state-of-the-art photogrammetry, 3D laser scanning, and BIM technologies. (2) Students will be able to process point cloud data using machine learning techniques. (3) Students will be able to compare and evaluate different 3D building model reconstruction techniques.