Automatic detection of dampness phenomena on architectural elements by point cloud segmentation

In conservation analysis, a fundamental goal is the detection of decay patterns on architectural elements in the Cultural Heritage domain. The current research aims at supporting and facilitating the identification of dampness phenomena, through the creation of specific routines for the automatic segmentation and computation of dense point clouds, based representative colour ranges. The remote decay survey via reality capture and data processing enables cost-effective non-invasive and contactless methods. The proposed pipeline provides a geometric and spatial distribution of moisture evidences, without loosing information about the three-dimensional configuration. In reverse engineering, dense point clouds are representations of virtual 3D models, composed by huge quantities of data points, which enclose geometrical, colour and texture information. They are the results of 3D scans of real objects, carried out using techniques such as digital photogrammetry. The historical convent of “San Leonardo” (XVI century, Monopoli, Bari, Italy) was selected as test case for the experimentation and validation of the proposed procedure, which highlighted the presence of dampness phenomena consistently affecting the intrados of masonry cross vaults.