From point cloud to feature extraction on the façade of a historic building

Traditional visual surveying of Cultural Heritage can lack accuracy and objectivity, and be characterized by limited compatibility with other methods, reducing productivity and efficiency. However, recent advancements in reality capture technologies like terrestrial laser scanning are facilitating the acquisition of precise geometric and color-related data that can more effectively support surveying, maintenance, and repair works. The data can be evaluated for valuable information about each masonry unit, which can then be included in building maintenance management systems. To achieve accurate surveys, significant effort is required in the form of observation, data acquisition, and drawings. Consequently, there has been an increasing interest in the automation of extracting architectural features. This thesis explores how modern reality capture technologies can be used to automate the segmentation of construction units on building facades. The aim is to deliver time-saving methodologies for architects and engineers, addressing the growing demand for accurate building documentation. By utilizing techniques like terrestrial laser scanning, the study aims to improve the process of surveying, and preservation, and enhance building maintenance management systems. This thesis works on the experimental use of an innovative plugin for the open-source 3D data processing software 'Cloud Compare', which utilizes an algorithm based on the Continuous Wavelet Transform for semi-automated segmentation of 3D point clouds. The experiments carried out in this study consider two types of masonry to understand how existing tools respond to stone and brickworks. Overall, the thesis provides an overview of existing methods and offers directions for future works in the studies of automated processes for extracting structural building features.