Automatic extraction of building features for building façades based on laser scanning technology

Abstract

Feature extraction on façades from unstructured point clouds is a challenging work, especially in the presence of noise. Previous studies in reconstructing the building façade are limited to traditional urban buildings. However, facade with a big proportion of windows, like curtain wall, are less explored. This research aims to propose an automatic approach to extract building features of a facade with a big proportion of windows, such as curtain wall. This approach includes automated detection of the facades, openings detection and geometry information extraction. A real case is used for testing the framework and validating the effectiveness of the approach.