Automated detection of repeated structures in building facades

Abstract. Automatic identification of high-level repeated structures in 3D point clouds of building façades is crucial for applications like digitalization and building modelling. Indeed, in many architectural styles building façades are governed by arrangements of objects into repeated patterns. In particular, façades are generally designed as the repetition of some few basic objects organized into interlaced and\or concatenated grid structures. Starting from this key observation, this paper presents an algorithm for Repeated Structure Detection (RSD) in 3D point clouds of building façades. The presented methodology consists of three main phases. First, in the point cloud segmentation stage (i) the building façade is decomposed into planar patches which are classified by means of some weak prior knowledge of urban buildings formulated in a classification tree. Secondly (ii), in the element clustering phase detected patches are grouped together by means of a similarity function and pairwise transformations between patches are computed. Eventually (iii), in the structure regularity estimation step the parameters of repeated grid patterns are calculated by using a Least- Squares optimization. Workability of the presented approach is tested using some real data from urban scenes.