The recent years saw an extensive use of 3D point cloud data for heritage documentation, valorisation and visualisation. Although rich in metric quality, these 3D data lack structured information such as semantics and hierarchy between parts. In this context, the introduction of point cloud classification methods can play an essential role for better data usage, model definition, analysis and conservation. The paper aims to extend a machine learning (ML) classification method with a multi-level and multi-resolution (MLMR) approach. The proposed MLMR approach improves the learning process and optimises 3D classification results through a hierarchical concept. The MLMR procedure is tested and evaluated on two large-scale and complex datasets: the Pomposa Abbey (Italy) and the Milan Cathedral (Italy). Classification results show the reliability and replicability of the developed method, allowing the identification of the necessary architectural classes at each geometric resolution.

The paper presents the first part of a research project concerning the creation of 3D terrain models useful to understand landslide movements. Thus, it illustrates the creation process of a multi-source high-resolution Digital Terrain Model (DTM) in very dense vegetated areas obtained by integrating 3D data coming from three sources, starting from long and medium-range Terrestrial Laser Scanner up to a Backpack Indoor Mobile Mapping System. The point clouds are georeferenced by means of RKT GNSS points and automatically filtered using a Cloth Simulation Filter algorithm to separate points belonging to the ground. Those points are interpolated to produce the DTMs which are then mosaicked to obtain a unique multi-resolution DTM that plays a crucial role in the detection and identification of specific geological features otherwise visible. Standard deviation of residuals of the DTM varies from 0.105 m to 0.176 m for Z coordinate, from 0.065 m to 0.300 m for X and from 0.034 m to 0.175 m for Y. The area under investigation belongs to the Municipality of Piuro (SO) and includes both the town and surrounding valley. It was affected by a dramatic landslide in 1618 that destroyed the entire village. Numerous challenges have been faced, caused both by the characteristics of the area and the processed data. The complexity of the case study turns out to be an excellent test bench for the employed technologies, providing the opportunity to precisely identify the needed direction to obtain future promising results.

In recent years, the investigation and 3D documentation of architectural heritage has made an efficient digitalization process possible and allowed for artificial intelligence post-processing on point clouds. This article investigates the multilevel multiresolution methodology using machine learning classification algorithms on three point-cloud projects in China: Nanchan Ssu, Fokuang Ssu, and Kaiyuan Ssu. The performances obtained by extending the prediction to datasets other than those used to train the machine learning algorithm are compared against those obtained with a standard approach. Furthermore, the classification results obtained with an MLMR approach are compared against a standard single-pass classification. This work proves the reliability of the MLMR classification of heritage point clouds and its good generalizability across scenarios with similar geometrical characteristics. The pros and cons of the different approaches are highlighted.

In the field of geomatics, the development of 3D surveying techniques that can produce dense point clouds and high-resolution three-dimensional models has opened the possibility of using these techniques to generate digital replicas of a given asset to be used in contexts of cultural promotion or entertainment. These non-traditional applications have been made easier by the spreading of low-cost but fast surveying techniques, allowing them to be used in contexts or projects with fewer resources. Close-range photogrammetry is nowadays the cheapest, most flexible, and fastest technique that can be used to produce high-resolution and high-quality digital copies of reality. In this paper, we present an application of the Ant3D multi-camera system, a novel photogrammetric device designed for the rapid 3D surveying of sites characterized by poor accessibility and narrow spaces. The paper describes the three-dimensional acquisition of underground tunnels in a hazardous abandoned mining site. The digitalisation work aims to create 3D models to be experienced safely in a virtual- or augmented- reality immersive navigation of the site for cultural promotion. A workflow is proposed to turn raw reality-based 3D models into refined and efficient 3D assets that can be exploited in a game engine. Applications for mobile devices, desktop, and VR headsets, using the produced 3D assets are described.