The Integrated Survey of the Pergamum by Nicola Pisano in the Cathedral of Pisa

Abstract

(M.B.) The paper deals with the methodologies and development of digital survey protocols and uses the Pulpit of the Cathedral of Pisa as the case study. The documentation project of the pergamum has forced to face the morphological complexity of this sculptured architecture: the wealth of decorations, sculptures, panels, columns and caryatids, which make the monument unique, likewise increase the complexity of reproducing a 3D model that describes its morphological and material characteristics. A first database of point clouds had already been made in the past years Pancani (Piazza dei Miracoli a Pisa: il Battistero. Metodologie di rappresentazione e documentazione digitale 3D, 2016 [1]); these measurements were made using a terrestrial laser scanner Z + F 5010C with integrated camera: these measurements were made with a Z + F 5010C terrestrial laser scanner with an integrated camera, performing 15 high-resolution scans around the pergamum. The result obtained however was not excellent: the limit of this model is the lack of shooting points at different altitudes, which led to have the pulpit described in its main dimensions; however the static nature of the instrument has created a cloud of points with numerous areas of shade, especially in the portions at the top of the monument, which do not allow to appreciate its sculptural qualities. Recently, the problem of the documentation of the Pergamum has been tackled a second time, trying to analyze the errors of the past experience in order to obtain a reliable and complete 3D model. For the purposes of the research the problem was divided into two different points of view: the morphological and metrically more reliable survey with laser scanner instruments; the realization of a three-dimensional textured model obtained with SfM technologies; since both methodologies work with point clouds it is understandable that it is possible to dialogue between the two databases; in fact the laser data can be used to check the error of the photographic one. A new laser scanner measurement campaign was carried out to follow the methodological protocols developed in recent years of scientific activities by the research group coordinated by the writer Pancani (La Citta dei Guidi: Poppi. Il costruito del centro storico, rilievi e indagini diagostiche, 2017 [2]). In addition to controlling the quality of the data collected during the acquisition phase, particular attention was focused on the recording phase of the obtained points clouds. In the attempt to define procedures to guarantee the reliability of the general point cloud, the certification phase of the survey was carried out. The database constitutes the metric and morphological basis on which to base all the following phase of realization of the photographic model: the use of SfM technologies today allows to create 3D point cloud from the alignment of the various photographs, not creating a model in metric scale, but based on the size of the acquired pixels; this procedure allows in the acquisition phase to take frames from different points of view, easily and quickly succeeding in integrating the shaded parts otherwise present in the laser scanner survey; the development of the calculation algorithms of the programs has led today to have clouds of increasingly reliable points and comparable to laser instruments Gaiani (I portici di Bologna Architettura, Modelli 3D e ricerche tecnologiche, 2015 [3]). For the survey of the pulpit it was necessary to define reliability protocols during the phase of acquisition of the frames, so as to guarantee the correct definition of the final texture to be applied to the model, and to check the error of alignment of the photographs. Because of the complexity and size of the clouds of points obtained it was considered useful to manage the creation of the 3D model in programs for the management of mesh models, to be able to manage the large amount of morphological information, control the size and number of triangles to obtain and optimize the mesh eliminating any errors due to digital noise. Finally, the 3d model was texturized, obtaining a high definition mapping model of the whole object.