In the last decade, there has been a growing interest in the development of DT models for existing structures and infrastructures in the built environment. In this context, laser scanner technology emerges as a technology due to its ability to capture dense point clouds of the built environment with precise geometry and semantics. The creation of a digital model from PCD involves two main steps: firstly, semantic labeling of point cloud to separate objects, followed by the creation of the digital model. The conventional process of creating digital models is associated with challenges in processing point clouds and inferring the topological relation between elements. To address the challenges, the capabilities of Artificial Intelligence (AI) and Machine Learning (ML) capabilities in scene understanding are leveraged. The main objective is to develop an automated method for creating a digital model of staircase objects using PCD. To achieve this, a data-driven bottom-up approach is used to identify points corresponding to each instance of a staircase within the point cloud. Furthermore, an automated method is implemented to extract values of parameters required for creating the digital model. To create the digital model of the staircase, a library of parametric staircase models is collected. These models are defined by key parameters such as height, width, and length of each run within the staircase. Additionally, these models are made adaptable by defining the direction of the staircase and the number of landing treads. The extracted parameters from the point cloud data are used for the creation of the digital model based on the predefined staircase library. The result of testing the proposed approach on the dataset highlights its effectiveness with a mean error value of about 3.5cm in the estimation of the elements parameters.     «

In the last decade, there has been a growing interest in the development of DT models for existing structures and infrastructures in the built environment. In this context, laser scanner technology emerges as a technology due to its ability to capture dense point clouds of the built environment with precise geometry and semantics. The creation of a digital model from PCD involves two main steps: firstly, semantic labeling of point cloud to separate objects, followed by the creation of the digit...     »