A Study of Denoising Algorithm on Point Clouds: Geometrical Effectiveness in Cultural Heritage Analysis

3D point clouds from photogrammetric and laser scanning surveys usually contain noise data due to different causes: non-cooperative material or surfaces, bad lighting, complex geometry, and low accuracy of the instruments utilized. Noise not only deforms the unstructured geometry of the point clouds, but also adds useless information and reduces the geometric accuracy of the mesh model obtained and, consequently, the results of any analysis performed on it. Hence, a cleaning of the raw data is mandatory. Point cloud denoising has become one of the hot topics of 3D geometric data processing. It helps in removing these noise data to recover the ground-truth point cloud, adding smoothing to the ideal surface. This paper uses a recent algorithm to investigate its usefulness for the creation of different cultural heritage models for structural analysis. The point clouds have been post-processed and compared with the raw data, extrapolating profiles, and subsequent meshestocalculate the real effectiveness of the algorithm on different geometries and case studies and to evaluate if this process can be utilized for the creation of more accurate polysurfaces (NURBS) from the 3D meshes for structural analysis.