Development of as-built BIM presents an ongoing challenge for the global BIM and computing engineering communities. The manual process for constructing as-built BIMs is time-consuming and requires skilled workers, the modeling time can be several weeks or months for a normal building. Therefore, the development of automated asbuilt BIM from point cloud data is necessary and it´s the focus of current research. Point cloud refers to the type of data obtained through a 3D scanner. The scanned data is recorded in the form of points. Each point contains three-dimensional coordinates, color information and intensity. Due to these characteristics of point clouds, more and more researchers are considering applying point clouds in the field of asbuilt BIM 3D reconstruction. A very important part of the 3D scene reconstruction is to identify the specific CAD model in the point cloud. Semantic registration is an important approach to achieve this purpose. This thesis aimed to complete a workflow of point cloud registration with Point Cloud Library(PCL). We first reviewed the mature point cloud processing algorithms, and introduced the relevant knowledge background in our thesis. Then a variety of point cloud processing algorithms in PCL were used in the registration process, such as noise filter, keypoint detection, feature descriptor computation, correspondence grouping and so on. Through this workflow, the CAD model can be accurately identified from the point cloud. And in the experiment part, we input multiple point clouds in different indoor scene to verify the point cloud registration process and fully analysis the influence of some critical factors. Finally, combined with the analysis of the experimental results, we put forward some suggestions for future research in this direction.     «

Development of as-built BIM presents an ongoing challenge for the global BIM and computing engineering communities. The manual process for constructing as-built BIMs is time-consuming and requires skilled workers, the modeling time can be several weeks or months for a normal building. Therefore, the development of automated asbuilt BIM from point cloud data is necessary and it´s the focus of current research. Point cloud refers to the type of data obtained through a 3D scanner. The scanned data...     »