Crowdsensing Close-Range Photogrammetry for Accurately Reconstructing a Digital Twin of a Cultural Heritage Building Using a Smartphone and a Compact Camera

The development of user-friendly and cost-effective survey technologies is critical in the protection and preservation of cultural heritage structures. Although they are commonly reconstructed using digital photogrammetry techniques, and with the integration of Terrestrial Laser scanners and Remotely Piloted Aircraft Systems, they are now increasingly being modeled by crowdsensed systems, which are easily accessible even to non-expert users. As such, the goal of this research is to evaluate the performance of a smartphone and a commercial compact camera in reconstructing a detailed and accurate digital twin of a cultural heritage object. In both cases, the close-range photogrammetric technique, based on the combination of Structure for Motion and Computer Vision approaches, was used. Those methods were tested on the Ognissanti Church, located in Valenzano, Italy. MicMac Graphic User Interface and CloudCompare, two open-source software and user-friendly interfaces, were used throughout the process. Thus, once camera calibration, preprocessing, and processing phases were completed, both collected databases were compared on the base of tie points image matching, average residuals in block-bundle adjustment resolution, and processing times. Despite variations in acquisition resolution and instrumental stability, and the fact that the point cloud from smartphone camera is 3 times less dense than that from compact camera, picture matching is equivalent after 15 to 20 orientation repetitions. The resulting two clouds were almost overlapping, with an average distance of 0.01 m. These findings also matched previous outcomes in the literature for small-volume reconstruction, substantiating the notion of performance independence from the size of the modified item.