Automated Geometric-Semantic Digital Twinning of Staircases from Dense Laser Scanner Point Clouds by Parametric Prototype

Abstract: Digital Twins (DTs) have become as valuable tools for management, and operation of assets. A main challenge, however, persists in the automated creation of high-quality buildings DTs, providing both precise geometry and semantics. This paper introduces a novel hybrid bottom-up, top-down approach for the automated creation of DT models of staircase structures using laser scanner point cloud. The proposed workflow involves separating inclined staircase points, designing parametric DT models, and model fitting through an optimization process. The results demonstrate the effectiveness of the proposed with an average accuracy of about 4 cm in determining the dimensions of the model’s elements.