PSI - Issue 29

Nicola Cavalagli et al. / Procedia Structural Integrity 29 (2020) 165–174 Cavalagli et al./ Structural Integrity Procedia 00 (2019) 000 – 000

166

2

1. Introduction The preservationof cultural heritage is of primaryconcern especially in a ll those countries wherehistorydelivered significant architectural heritage, which is exposed to inherent degradation with time and to na tural hazards (e.g. seismic loads). Within this context it is crucia l to develop effective and integra ted methodologies a imed at characterizing the conservation state of a structure identifying and preventing potential vulnerabilities (Bitelli et a l., 2016; Breccolotti et a l., 2018; Pa lazzi et a l., 2019; Galassi et a l., 2020). Currently, visual inspections are the primary and most important practice for identifying the existenceof damage and to assess the hea lth sta te of historic structures. Actua lly, most of historic structures are characterized by a significant level of complexity in terms of constituent materials, existingdamage and geometry (Tempesta et a l., 2019; Zampieri et a l., 2019). Moreover it is often difficult and/or impossible to have direct access to the structural components(e.g. masonrybridges, towers andmore) (Gioffrè et al., 2017; Roselli et a l., 2018). In this framework, the analysis of damaged historic structures and the evaluation of the geometrical irregularities effects on the structural response is still a cha llenge (Cavalagli et a l., 2017; Severini et a l., 2018; Zampieri et a l. 2018); in this sense, the possibility of usingaccurate models which describe the actual geometry is advantageous for structural purpose. For this reason, in the recent few years, significant interest has been posed on emerging survey techniquesuch as Terrestria l Laser Scanner (TLS) and Unmanned Aeria l Vehicles (UAVs) based photogrammetric surveys, able to model the external envelope of the tested structural system through point clouds, providing geometric data quickly and safely, insteadof in-personvisual assessment. In the latter case, Structure fromMotion (SfM) procedures are used to achieve a 3Dgeometric model from its projection into a series of high-resolution images taken fromdifferent view points (Ippoliti et a l., 2015; Martinez-Carricondo et a l., 2020). TLS survey technique can be considered a “conventional” and reliable survey method, thanks to the advances in technology and computer graphics, and to its extensive use by practitioners and researchers (Olsen et a l., 2010; Guldur et a l., 2015). However, TLS technique requires specific work setting, amongwhich the accessibility of sites, expert users and, moreove r, the equipment can be considered rela tively expensive (Chen et a l., 2019). This maybe a significant drawback especially when, working on historic structures, site conditions and occlusions caused by different kind of objects (e.g. structural components causing self-shadowing) could make the application of this technique not able or not sufficient to perform an overall survey. As analternative, UAV-basedphotogrammetric techniquehas gained interest worldwide.With respect to TLS, using UAVs equipped with commercial cameras means to have lower cost of instruments, higher speed of data acquisition and, above all, a better representation of the 3D model through by using textures (Achille et a l., 2015; Kha loo et a l., 2018; Barba et a l., 2019). The aforementioned aspects are of crucial interest when information about historic buildings is required. In fact photogrammetrybased surveys are able to provide: 1. Virtua l 3D model to be used for historical documentation and geometrical information and to analyze the existence and the localization of significant cracks and/or lackingmateria l; 2. Photorealistic representation to be used for ma terial characterization and degradation analysis, comprising damage mapping andquantification of damage; 3. 3Dgeometric models to be used as basefor numerical analyses aimed a t assessing the structural performance. In this work, the accuracy of UAVs photogrammetric survey for the evaluation of historic masonry structural damages is discussed with reference to a full-sca le masonry arch bridge in the neighborhood of Todi, Central Italy (Pepi et a l., 2017). The bridge is characterized byan irregular segmental arch with a spananda rise of about 10 mand 4 m, respectively. The structure exhibits typicalmechanical deterioration phenomena consisting in cracks belowthe vault, material detachment in the side wa lls and the arches, and lack of material on theWest and East abutment. The damage level of the masonry is even worsened by vegetation grown between the cracks of the stones (Fig. 1). To achieve a reliable bridge inspection, close range digita l images were firstly acquired using UAVs equipped with cameras and then processedobtaininga texturizedpoint cloud. The accuracyof the obtained results were assessedby a direct comparisonof thepoint clouds obtained bySfMtechniqueandTLS survey. Finally, the texturizedpoint cloud modelwas used to build an accurate3Dmodel and to identifycracks evaluating the volumes of lackingmaterial.