BIM enhancement using multi-modality imaging

Abstract

Scan-to-BIM process is currently a popular research area where a physical space is digitally captured using various imaging modality, and the resulting point cloud is converted into data-rich 3D BIM model representation. Currently, a lot of industrial application of Scan-to-BIM largely relies on the usage of Terrestrial Laser Scanner (TLS) to obtain a complete scan of the environment. However, this at times might be impossible due to insufficient space to setup the large and heavy TLS, or occlusions of certain areas which does not fall within its scanning line-of-sight. To contribute to the efforts of the Building Imaging team under SJ-NTU Corp lab to develop an intelligent Scan-to-BIM process, this FYP looks into the aspect of multi-modality imaging, particularly in the area of data fusion between TLS and handheld scanner. The first phase of this project focuses on the latest technology of handheld scanner available in the market that can fulfil the requirements for quality imaging, and to familiarize with the usage of handheld scanner. Then in the second phase, a coarse-to-fine procedure was developed to fuse data from both modality together. Successful results have shown that the proposed algorithms (4PCS and ICP) in the procedure have shown potential in merging point cloud data from different modalities together.