A Mono-instrumental Approach to High-quality 3D Reality-based Semantic Models. Application on the Palladio Library

Abstract

The paper presents a new workflow to build geo-3D models able to populate a wide information system organizing and managing Palladio's works documentation. Main goal is to smooth the integration of data from different sources and promote a more profound understanding of Palladio's works. Starting from an integrated survey approach (laser scanner and Multi-Image Spherical Photogrammetry), reality-based 3D models are robustly constructed and semantically segmented from point cloud features based on Palladio treatise architectural semantics. The 3D models outputs are exported into a 3D Geographic Information System (GIS) to facilitate the visualization of Palladio's works with reference to topography. At the end, an evaluation mechanism is employed to control the quality of 3D models. The workflow could be easily generalized to other architects documentation systems both because Palladian case is a superset of most of other architect documentation dataset, allowing to focus and to solve all existing problems and because techniques and methods are general and appropriate for most part of historical architectures.