Cultural heritage documentation using geomatic techniques;

Cultural heritage documentation using geomatic techniques; Case study: San Basilio s monastery - L Aquila Donatella Dominici, Elisa Rosciano, Maria Alicandro, Michail Elaiopoulos, Serena Trigliozzi and Vincenzo Massimi University of L'Aquila - Faculty of Engineering - Department Civil, Environmental Engineering and Architecture. Laboratory of Geomatics - Via Gronchi 18 - L Aquila - 67100 Italy donatella.dominici@univaq.it +39/ 0862 434118 Abstract Each community, passing through its history and its collective memories, has the duty to conserve, identify and properly manage its own heritage that, unfortunately, is often susceptible to transformations due to time, anthropogenic factors or damages caused by natural phenomena. For this reason, heritage artifacts are being extensively documented in order to be maintained in the very best way and bestowed for the benefit of future generations. Geomatic techniques (total stations, GNSS receivers and laser scanners), when integrated among them in an appropriate surveying methodology, data management and processing, can provide dedicated methodologies able to detect and survey the particular characteristics of such complicated structures. In this way, 3D geometric models of the building can be created and enriched with a wide range of information, results of the combination between the visual potential of the images with the high accuracy of the geometric survey. Thus, the use of geomatics represents an interesting tool both for the visualization and the metric evaluation of the structure itself providing an efficient support for professionals and public bodies. The survey of the old monastery of San Basilio, a 16th century building in L Aquila, central Italy, is presented in this paper. earthquake of April 6, 2009. The various geomatic techniques used to create the restitution of a 3D model and especially the laser scanning survey georeferenced with the use of the framing network will be presented too. In the first part of the paper, the framing networks, essential to ensure an adequate accuracy of the final three-dimensional model and their georeferencing, will be described. The second part will be devoted to the description of all operations that resulted in the three-dimensional model using the laser scanning techniques. II. THE HISTORY AND THE ACTUAL STATE OF THE BUILDING The first mention of S. Basilio s monastery dates back to 496 A.C. [1], according to them it has been founded by the monks of S. Equitius as a monastery dedicated to the Benedictine nuns of the cloister. In 1703, following a violent earthquake, the monastery s facade has been rebuilt and an arm was added to the west of the old monastery. Then, during the 17 th century the complex underwent architectural substantial changes, assuming the character that still distinguish it. Index Terms Laser Scanner, Heritage, Documentation, Total Station, GNSS, Geomatics I. INTRODUCTION The identification and the optimal management of cultural heritage represents an actual duty of each community as helps the preservation of the collective memories of the past that have to be bestowed to the future generations. Both monuments and historic buildings represent a large part of this collective memory as they are tangible physical elements present on the territory. This awareness gives rise to the need of creating a digital database that contains a detailed documentation, both qualitative and quantitative, of the architectural heritage. When all this information is gathered can easily be disseminated, on one hand, to the various professionals involved in rehabilitation and restoration and, on the other hand, to promote an easily accessible and innovative path of knowledge. Geomatic techniques, when used in a synergy with each other, can play a fundamental role in the process of creating this knowledge and certainly ease its comprehension and divulgation. This paper presents the 3D surveying for the state of the art evaluation of a historical building in L Aquila, central Italy, unfortunately damaged by the Fig. 1: A panoramic view of San Basilio Monastery In 1993, the University of L'Aquila has entered into an agreement with the Municipality of L'Aquila for the use of the structure for institutional purposes. The structure, as a result of work of architectural recovery and changes of the use destination, was finally completed in 2008 and dedicated to the memory of Prof. Luigi Zordan [2]. The earthquake of 2009 brought considerable damages to the building, making it unusable. The surveying operations and subsequent restoration were then commissioned at the University of L'Aquila. The structure consists of a ground floor of about 1300 square meters, a first floor of 500 square meters and last floor of 390 square meters. It presents a complex geometry derived from the overlapping architecture of different ages and stages, from 978-1-4799-3169-9/13/$31.00 2013 IEEE 211

abnormalities induced by structural damage and degradation of materials. III. PLANNING THE TOPOGRAPHIC NETWORKS For this case, both a framing and a control network have been materialized, both of them will be illustrated in the next paragraphs. III.II THE FINAL CONTROL NETWORK Having the framing network, the final network of control points has been surveyed using triangulations with a motorized robotic total station Leica TS30 (Figure 4). The instrument s precision is 0.5cc for angles and 0.5mm + 6 p.p.m for distances. In order to evaluate their quality an evaluation of their grade of precision and accuracy was needed. Thus, a rigorous method used for the design of geodetic networks has been used [3]. This methodology divides the problem in four grades with respect to the adjustment parameters of the least squares model: - ZERO ORDER DESIGN (ZOD) that regards the datum definition problem; - FIRST ORDER DESIGN (FOD) that regards the network s configurations; - SECOND ORDER DESIGN (SOD) that regards the weight to attribute on each observation; - THIRD ORDER DESIGN (TOD) that regards the network s densification. The classification of the various orders can be reviewed with respect to the least squares estimation procedure. In order to guarantee the objectives of reliability, accuracy and precision, the framing network has been created for the datum definition. In addition, a local network has been materialized and used for the direct georeferencing of all laser scanner measurements following, when possible, a high redundancy scheme. The design of a dense and redundant network was necessary mainly because of the structural complexity of this building. The careful planning phase, in addition to having secured a good configuration of all observations played a key role in obtaining the appropriate redundancy and reliability. III.I FRAMING NETWORK The framing network has been measured through GNSS techniques. The network s configuration (Figure 3) has been defined by keeping in mind the particular spatial distribution of the building that, indeed to the North, is adjacent to the old monumental walls of L Aquila. In addition, the monastery is located in an area characterized by tall vegetation and surrounding buildings, all of them causes of a poor signal acquisition (Figure 1). The framing network has been constituted by six points materialized and surveyed using the GNSS in static mode, with long sections of signal acquisition (30 min. minimum) for each vertex. The obtained data have been elaborated in post processing using the LGO software by Leica obtaining a fixed double difference solution. The final coordinates have been framed in the European reference system ETRF2000 (WGS84) using the Abruzzi s network of permanent stations. Fig. 3: GNSS framing network. Fig. 4: The final control network All measurements have been made using 4 iterations in a redundant network that permitted the least squares adjustment using dedicated adjustment software (Micro-survey Starnet). This elaboration computed the adjusted coordinates with their relative error ellipses that have been found to be smaller than 0.6mm, the whole adjustment has been controlled trough statistical tests (data snooping and Chi square) IV. LASER SCANNING SURVEY The terms "surveying" and "representation" are now accompanied by the adjective "digital" that has revolutionized the concept of knowledge and documentation of cultural heritage. Laser scanners represent an excellent surveying instrument that has been proven to be versatile and able to offer effective three-dimensional representations of the elements present in its field of action. Each set of measurement (point clouds) represents the surrounding space, up to a certain range. Thus, it is easy to understand that the difficulties connected with this type of survey increased proportionally with the complexity of the elements to be acquired. The advantage of using a device such as the laser scanner is that not only the coordinates of all these points are calculated but also information connected to their radiometric characteristics can be determined in a limited interval of time. In this case study, due to the complicated spatial distribution of the structure that had to be investigated, some precautions were needed among the various steps that are described as follows. IV.I POINT CLOUD ACQUISITION The number of point clouds has been decided considering the geometric characteristics of the building. Thus, the exteriors were detected using 10 point clouds with a high resolution (corresponding to a point grid of about 7 mm.). In adition, 10 paper targets and 6 natural points have been used for their registration and georeferensing. Regarding the building s interior about 100 point clouds have been made with a medium resolution of 15 mm. Their final distribution has been made as follows: 46 on the ground floor, 20 on the first floor and the rest to the last floor. Totally, 40 paper targets and 20 magnetic targets have been used to control the internal points clouds in all spaces. All gathered data were processed with the software Leica Cyclone 8.1, dedicated to the laser scanner Leica products. 212

IV.II DATA PROCESSING V. FINAL RESULTS Since every points cloud is related to an intrinsic reference system a merge between all point clouds have been made. The final point cloud was then georeferenced using the framing network illustrated above achieving an RMS value of 13mm for the exterior. To reach a higher accuracy, decreasing the weight of some measurements with high variance has become necessary. Thus, after various tests a reduction of 40% to the weight of certain groups of point clouds provided a final RMS value equal to 7 mm. Regarding the interior, the change of the use destination from convent to congress hall created numerous spaces separated by wall panels to accommodate different facilities; (i.e. audio/visual equipment / services). Besides increasing the number of the scans appropriately, the use of magnetic targets has been necessary (for their property to maintain fixed their center even if rotated) (Figure 5). The working conditions allowed, through the practice, finding deviation to the known strategies in order to achieve higher accuracy. The final RMS of the global model, taking the precautions listed, was found to be 9 mm. The final result is then the rasterized and georeferenced three dimensional model (Fig. 8). The obtained final 3D model represents a great tool through which geometric and structural characteristics of the building, as well as its structural components, their mechanical characteristics and state of conservation can be interrogated. Fig. 5: Magnetic target on the left and paper one on the right. The registration of the point clouds (not yet georeferenced) without this grouping technique presented a lower RMS (3 mm) with respect to the more laborious grouping strategy that presented an RMS of 7 mm. The table 7 illustrates the number of the created groups and their RMS value. TABLE 6. Grouping strategy and RMS Ground Floor 15-17-19 Bathroom 2 Magnetic target 26-27 Tunnel 2 Magnetic target 52/56 Bathroom 9 Magnetic target 80-81 Room 9 P.Tar. not visible First floor 89/94 Bathroom 2 Magnetic target 96-97 Tunnel / elevator 3 Magnetic target 100-101 Bathroom 2 Magnetic target Second floor 69/72 Bathroom 2 Magnetic target 73/75 Bathroom 2 Magnetic target 76/81 Tunnel 8 Obstacle 84/86 Bathroom 2 Magnetic target Exterior 6-57/62 Exterior 13 natural targets Figure 7 shows how the presence of an auditorium (on the right) has increased the number of finally made stations (in blue) with respect to the originally projected ones (in red) to cover the whole area. Even in this case a higher global RMS during the integration of all data has been obtained (10 mm). Fig. 7: Projected point clouds in red and final needed ones in blue. Besides this, accurate prospects, horizontal and vertical sections of the structure have been obtained (Figure 9). Fig. 8: Global three dimensional models The same vector products could also be created using known and tested traditional surveying techniques even in much longer time manner. In addition, during an emergency situation the presented laser scanning technique remains an efficient approach for a detailed documentation that among other advantages, present significantly lower risk for the operators. Certainly, the fact that laser scanning techniques can still not be completely used for monitoring purposes remains as many of its actual limitation are still under continues evolvement (ie the control points recognition in more epochs). 213

[6] Biagi L., 2006. I fondamentali del GPS. Geomatics Workbooks vol. 8, ISSN: 1591-092X. [7] Brigante R., Dominici D., Fastellini G., Radicioni F., Stoppini A. (2009): Confronto e integrazione fra tecniche geomatiche per la documentazione e il monitoraggio dei beni culturali. Proccedings of the XIII ASITA National conference in Bari, Dicembre 2009 ISBN 978-88-903132-2-6. [8] Caspary, W.F., 1987. Concepts of network and deformation analysis. Monograph 11, School of Surveying, The University of South Wales, Kensington, Australia. Fig. 9: The ground floor map and the main facade s prospect VI. CONCLUSIONS The present work demonstrates that the synergy between the use of various geomatic techniques allows efficient results of high precision and accuracy, even in a short time, even if needs special attention and evaluation especially in case of post-earthquake situations. Finally, this study s aim is to apply this methodology to the entire existing architecture heritage in order to create an updated database useful as excellent support for the elaboration and a correct management for the future reconstruction of these artifacts. The integration of geomatics in the field of historical and architectural heritage offers the precious opportunity to have a rich and up to date database, enriched with digital three-dimensional models of existing real heritage of considerable merit, easily disseminable, and extremely useful as excellent support for the future reconstruction or for those who are responsible for the promulgation of this knowledge. BOOK AND PAPER REFERENCES [3] Grafarend, E.W., Sanso, F., 1985.Optimization and Design of Geodetic Networks.Springer. [4] F. Guerra, C. Balletti, A. Adami, 2005 3D multiresolution representations in archaeological sites, Proceeding of CIPA 2005 XXInternationa al Symposium International cooperation to save the word s cultural heritage, Torino, 26 settembre 01 ottobre 2005. WEBSITE REFERENCES [1] Abruzzi Region http://www.regione.abruzzo.it/xcultura/index.asp?modello=%20baro ccoschedaaq&servizio=xlist&stilediv=monoleft&template=intind ex&b=menubaro2133&tom=133 [2] Univaq, http://www.univaq.it/section.php?id=1143 [5] Dominici, D., 1989.Tesi di dottorato: tecniche di analisi di reti tridimensionali d i controllo rilevate con metodi classici e GPS. 214