COSCH Final Conference
Guides to Good Practice in Documentation of Cultural Heritage Assets
Monday, 10th–Tuesday, 11th October 2016
i3mainz - Institute for Spatial Information and Surveying Technology – Mainz University of Applied Sciences,
Lucy-Hillebrand-Str. 2, 55128 Mainz (Germany)
PROGRAMME (PDF)SPEAKERS AND ABSTRACTS
Anna Bentkowska-Kafel, Multimodal Recording of Historic Coins. Worth the effort?
This presentation illustrates the final report on an exploratory study of Roman silver denarii conducted between 2014 and 2016 by a group of COSCH researchers working for scientific, academic and cultural institutions across Europe. A summary evaluation of approaches to multimodal recording will be offered. Scientific and museological questions to do with methods of technical examination, conservation, documentation and presentation of historic coins guided this research. An exhaustive list of questions can be found in the final report and on a dedicated website. Several non-invasive optical metrological, imaging and analytical techniques were applied to silver denarii of Diva Faustina from two different editions. The use of the same objects throughout the tests enabled meaningful comparison of geometrical measurements and other results that will be presented independently.
The study has been informed by earlier research, including: propositions for the development of relevant methods for recording historic coins in 3D (see Zambanini, M. Schlapke, M. Kampel and A. Müller, 2009, for an overview); applications of Reflectance Transformation Imaging (RTI) to numismatics (M. Mudge, C. Schroer et al., 2005 and 2012); metallographic examination of chemical composition of Roman coins using optical and scanning electron microscopy (SEM-EDS) (M. Ponting and K. Butcher, ongoing since 2014).
The question of adoption of advanced digital technologies in museum practice will be highlighted. Based on the input into this COSCH study from leading coin experts, namely numismatists and metal conservators, we ask whether modern optical recording technologies are ready to supersede traditional photography as the main tool for the study and documentation of historic coins. The identified value of the COSCH study will be presented, alongside its main issues, pointing towards possible solutions.
Irina Ciortan, Digitisation of cultural heritage at the National Museum of Romanian History.
The National Museum of Romanian History located in Bucharest holds a vast collection of Medieval manuscripts, Byzantine icons and archaeological objects of both Romanian and European cultural significance (made of ceramics, marble, stone, copper, gold and silver) covering historic periods dating from Prehistory to Modern Age. The museum has been under works of renovation for the last years, so most of the artifacts in its collection are not available for public visits.
The main purpose of the case study was to increase the collection of digitized objects of the museum and to enhance the possibilities of visualization in the virtual gallery, where the public could access objects that for the moment are not available on the site of the museum. A secondary objective of the proposed case study was to implement state-of-the-art non-invasive techniques for the image acquisition of the artifacts, using both spectral and spatial object documentation. Lastly, the final aim was to disseminate and inform about the digitization process and to provide guidelines and specifications of good practices that will serve benefit both the scientific and artistic community
involved in the field of cultural heritage.
Research is like an unpolished diamond, where new facets haven't been given the shine yet. So we can say that this Case Study has reached an end, but only from the COSCH framework point of view. At the COSCH Final Conference, there will be presented what has been achieved from the initial plan, what went right, what went half-right, what do the end-user think and what are future recommendations to improve similar image acquisition and processing pipelines for cultural heritage.
Amandine Colson, Case Study „Bremen Cog, when Science meets the public".
In summer 2014, we started our case study with a long list of objectives. The main idea was to test different measuring techniques in order to visualise, identify and analyse the on-going deformations processes on a PEG-conserved wooden medieval vessel, the Bremen Cog, exhibited at the German Maritime Museum in Bremerhaven.
The driving motivation was the possibility to use the same data to enhance public interest on current research questions and to inform them "live". The public learns about research projects usually much later: through exhibitions or publications, when results have already been achieved. Giving the opportunity to visitors to follow the process of research is part of Museums modern communication strategies.
After two years work with different specialists, we would like to present the work achieved and to have a critically look on the gathered experience.
The Bremen Cog has been successively measured with Photogrammetry, Terrestrial Laser Scanner, and Total Station. The data's geometry has been compared to validate the measurements. At that stage, some technical challenges occurred raising more questions. The initiative led to the opening of a PhD position at the German national Maritime Museum to carry on with this study. Recent contact with experts from the Vasa Museum gave another profundity of designing a satisfactory deformation monitoring system.
As a matter of fact the case study contributes not only to the technical aspects of imaging techniques, but also to a more global view. Such elaborated projects are to be managed and planed carefully and reflect the complexity of using imaging techniques in the cultural heritage field, where one deals with "unique" cases and not standardised objects. Once more it reveals the pertinence of interdisciplinary research.
Christian Degrigny, Authenticity of Germolles' mural decoration and painting technique used: the input of imaging techniques.
Situated in Burgundy, France, the Château de Germolles is the best preserved of the remaining residences of the Dukes of Burgundy. It was owned by Margaret of Flanders, wife of Philip the Bold, Duke of Burgundy and brother of Charles V, King of France. Built between 1380 and 1400, it shows the interest of French dukes in rural environments.
The noble floor is decorated with wall paintings that are a unique example of courtly love spirit that infused the princely courts of the time. After being concealed sometime in the 19th century, the wall paintings were rediscovered during the 2nd world war. They were non-professionally uncovered in the 1970s and then restored in 1989-1995.
Medieval accounting notes of the Château record the materials acquired to make the mural decoration, but they do not match the composition of the existing paintings. This interesting mismatch between the archival and material evidence, and the complexity of the painting technique used, were the reasons for setting up the Germolles case study.
The combined scientific and financial supports of COSCH Cost Action and DRAC (Regional Direction of Cultural Affairs) Burgundy enabled to study Germolles' wall paintings using some of the most innovative imaging techniques, alongside more traditional examination techniques, to address some of the questions raised, in particular the authenticity issue, the painting techniques used, the amount of original materials surviving and their conservation condition.
Best practices in documenting in an almost non-invasive way Germolles wall paintings were defined. Protocols of investigation included the observation of the paintings first globally and then locally using technical photography, colorimetry, IR thermography, hyperspectral imaging as well as highlight-reflectance transformation imaging and the elemental analysis of the decors with portable XRF (X-ray fluorescence) and micro-invasive LIBS (Laser Induced Breakdown Spectroscopy). Further investigation would require micro-sampling.
The investigations showed that the original decorations remain in traces and confirmed that the paintings were made with a sophisticated, complex technology which involved in particular the use of tin foils and gold leaves as suggested by the medieval records. The data collected is vast and varied and their management has been studied using basemaps such as orthophotographs on which macro and micro-photographs could be aligned.
Róbert Fröhlich, Levente Tamas, Alain Trémeau, Zoltan Kato, Region Based Fusion of 3D and 2D Visual Data for Cultural Heritage Objects.
There are many Cultural Heritage applications that use different types of 3D scanning techniques to capture a precise digital model of an object that needs to be documented. In many situations, for better results, the 3D data can even be combined with high resolution images captured with RGB cameras. These can also be color calibrated if a correct color representation is necessary. We propose a workflow for the fusion of these 3D and 2D visual data. Using the data acquired by cheap, standard devices, like a 3D scanner having a low quality 2D camera in it, and a DSLR camera with a high resolution sensor, one can produce high quality color calibrated 3D model for documenting purpose. The proposed processing workflow combines a novel region based calibration method with an ICP alignment step used for refining the relative pose of the camera to the 3D coordinate frame. It works on color corrected 2D camera images and 3D data, that doesn't necessarily contain intensity/color information, but can make use of it if it is available. These can be acquired with commercial 3D scanners and color cameras without any special constraint. In contrast with the typical solutions, the proposed method is not using any calibration patterns or markers. If a color calibrated model is to be produced, then color calibration of the camera has to be performed using a standard calibration pattern. Our workflow assumes that the input images are already color corrected. The efficiency and robustness of the proposed calibration method has been confirmed on synthetic data, while the complete workflow has also been verified on real data captured with different devices.
Piotr Kuroczyński, Virtual Research Environment for digital 3D reconstructions – Guides to good practice.
Since the 1990s the application of the digital 3D reconstruction and computer-based visualisation of cultural heritage increased. The virtual reconstruction and 3D visualisation revealed a new "glittering" research space for object-oriented disciplines such as archaeology, art history and architecture. Nevertheless the academics concerned with the uprising technology recognised early the lack of documentation standards in the 3D projects leading to the loss of information, findings and the fusion of knowledge behind the digital 3D representation. In the last decade guidelines on the scholarly approved computer-based visualisation and thoughts on 3D documentation (e.g. metadata) where introduced. The recent academic commitment to the design and application of so called Virtual Research Environments (VRE) for 3D research projects is promising. The investigation of the experiences with the VRE under construction and the establishment of the sustainable digital research infrastructure has still to be forced.
Based on the methodological fundamentals of the digital 3D reconstruction the potentials and challenges in the light of emerging Semantic Web and Web3D technologies will be introduced. The presentation subscribes a scientific methodology and a collaborative web-based research environment followed by crucial features for this kind of projects. As the groundwork a human- and machine-readable "language of objects" and the implementation of this semantic patterns for spatial research purposes on destroyed and/or never realised tangible cultural heritage (art&architecture) will be discussed. Using examples from the practice the presentation explains the requirements of the Semantic Web (Linked Data), the role of controlled vocabularies, the architecture of the VRE and the impact of a customised integration of interactive 3D models within the WebGL technology.
The presentation intends to showcase the state-of-the-art on the way to a digital research infrastructure. The focus lies on the introduction of scholarly approved and sustainable digital 3D reconstruction, complaint with recognised documentation standards and following the Linked Data requirements.
Tibor Lukic, Image processing techniques for crackle net analysis of old icons.
In easel painting (icons) mapping the status of conservation is a complex process, which function is to lay out detected damages or symptoms of decay. Detecting sharp discontinuities, such as crackles in the host medium is of significant importance. We present image processing techniques developed for crackle net structure segmentation of images obtained by different acquisition methods. Application examples on several images, presenting Orthodox icons, will be given.
Aurore Mathys, Guide to good practice of Reflectance Transformation Imaging: the Roman coins case study.
Many analysis and digitisation techniques have been applied to four Roman coins in the framework of the "Study of Roman silver coins using non-invasive spectroscopic and 3D imaging approaches". Reflectance Transformation imaging (RTI) and Polynomial Texture Mapping (PTM) are two of those techniques. They are usually generally regroup under the large appellation of RTI and are non invasive technique that enable to virtually re illuminate an object.
Four different RTI domes were used to image the Roman coins and the results were compared taking in consideration the differences in equipment.
This presentation will review the main differences between the domes and the impact of those differences, but also the differences between visualisation of differents algorithms and softwares. It will also draw some general conclusion on best practice for capture of RTI data.
The differences visually observed between the different imaged set will also highlight the challenges of comparing RTI dataset.
General conclusions for the coins case study state that all RTI/PTM data displayed a large amount of detail. RTI enables a realism and comprehension of the material surface that would otherwise only be possible by physical handling of the object. RTI is a very suitable technique to be applied to flat objects, but the size of the coins can be an issue for depth-of-field on high resolution images.
Vera Moitinho de Almeida, On the study of 3D digital models of historical silver coins: an integrated approach.
The silver Roman denarii used in the case study raise interesting questions concerning their provenance, authenticity, design and iconography, purpose of issue and historic usage. They also pose a considerable recording challenge due to their small size, particular material and surface properties, some of which may not be original, but a result of extensive cleaning.
Many relevant aspects for the analysis of historic coins have been traditionally described in rather ambiguous terms, without taking into account the advantages of quantitative measurements of, namely, shape/form and macro/micro topographic features. In the scope of this case study, a number of 3D systems and techniques (structured light scanner, laser scanner, Structure from Motion, photometric stereo) to record silver coins have also been applied by several European institutions. This presentation will summarize the results of two Short Term Scientific Missions (STSM), whose aims were manifold. Namely, to contribute to: (1) establish whether the selected recording 3D systems and techniques can support the comparison of certain features and properties; (2) a better knowledge about the surface characteristics of historic coins, by characterizing and analysing a number of important geometrical and topological features, using distinct metrological software, techniques and procedures; (3) the development of an efficient and comprehensive methodological framework for the 3D digital data capture, processing and analysis of historical silver coins; (4) the publication of a guide to good documentation practice, including relevant digital preservation guidelines.
Agnieszka Olejnik-Krugly, Przemyslaw Korytkowski, Color-Accurate Image Archiving of paintings with supporting spectrophotometry.
The usual practice in image archiving of paintings is to photograph a color target together with the painting. This allows creating and applying an ICC profile to the image for better color reproduction. Often only a few colors from the painting are represented in the target, resulting in a color profile that is not adequate. In this presentation we introduce a method for achieving better color accuracy by incorporating spectral measurements into the ICC profile. We combine spectral measurements of a selected spots on the painting with commonly used targets like: Color Checker Classic and Digital Color Checker SG. Our aim is to develop a methodology that enables obtaining a four-star class digital image according to the current FADGI "Technical Guidelines for Digitizing Cultural Heritage Materials" draft version from September 2015. This means that the mean color accuracy should be less than 2 dE2000 and the max color accuracy should be less than 10 dE2000. In our methodology we identify a set of potential points (with a consistent color of an area required for spectral measurements) in the painting from which we then select representative points (with characteristic colors), which will be measured spectrally. The data thus measured from the painting are merged with the color target data into a more comprehensive ICC profile.
M. Picollo, A. Casini, C. Cucci, L. Stefani (IFAC-CNR, IT); T. Vitorino (New University of Lisbon, PT, and IFAC-CNR, IT); E.K. Webb (Smithsonian's MCI, USA, and SEAHA CDT, University of Brighton, UK); S. Nascimento, J. Linhares, C. Montagner (University of Minho, PT); R. Pillay (C2RMF Musée du Louvre, FR); M. Hauta-Kasari, A. Gebejes, J. Hiltunen (University of Eastern Finland, FIN); J. Jussila (Specim, FIN); A. Jung (University of Leipzig, D); M. Vilaseca, F.J. Burgos, Xana Delpueyo (Technical University of Catalonia, ES); J.Y. Hardeberg, S. George, I.M. Ciortan (Norwegian University of Science and Technology, NO); I. Baarstad (Norsk Electro Optics, NO); J. del Hoyo Melendez, M. Obarzanowski (National Museum in Krakow, PL); L. MacDonald, M. Strlic, H. Mahgoub (UCL, UK); R. Schettini, S. Bianco (Università degli Studi di Milano-Bicocca, IT); A. Tremeau (University of Saint ‘Etienne, FR); E. Matoušková (Czech Technical University, CZ); J. Valach (Institute of Theoretical and Applied Mechanics of the Academy of Sciences, CZ); G. Trumpy (University of Zurich, formerly at the National Gallery of Art, USA); V. Moitinho de Almeida (formerly at The Cyprus Institute, CY),
Limits and advantages of different 2D VIS-NIR multi- and hyper-spectral imaging systems in addressing conservators' concerns.
Multi- and hyper-spectral imaging systems have been applied in museums, galleries, and conservation laboratories as an in situ technique for the study and accurate digital documentation of artworks, mainly paintings. In the past decades an increasing number of devices has been developed for such purposes.
Multi-spectral imaging (MSI) devices are constructed to acquire a sequence of reflectance images on a number of non-contiguous spectral bands, which are usually selected by means of a suitable set of filters. Hyper-spectral imaging (HSI) systems, on the other hand, are designed to acquire a fine spectral selection of almost contiguous spectral-bands in the investigated spectral region by means of dispersive optical modules (gratings). MSI devices are intended to image an area in some tens of spectral bands, while the HSI systems are capable of imaging an area in hundreds of bands simultaneously. This distinction leads to the conclusion that generally MSI approaches are not the most effective for identifying chemical compounds of artists' materials.
MSI and HSI results span from enhancing and revealing underdrawings, identifying past conservation treatments and measuring colour to extracting reflectance spectra for the identification of materials and producing maps of the spatial distribution of materials over an entire object.
However, the large number of available systems emphasises the importance of obtaining a better understanding of the elements that are part of the data acquisition and processing for different instruments and methodologies to provide accurate and precise answers to museum professionals' questions and concerns.
This presentation will report the experience gained within the Colour and Space in Cultural Heritage (COSCH) COST Action in facing constructive discussions with painting conservators on the use of MSI and HSI techniques.
Ruven Pillay, Acquisition, Calibration and the Processing of Hyperspectral Data: Lessons from the Round-Robin Test.
The Round-Robin test conducted by Working Group 1 included a wide range of participants, spectral imaging equipment and methodologies. The results obtained revealed a high level of variability and a number of problems related to system calibration and data processing. Incorrectly calibrated or overly noisy data make tasks such as materials classification, change detection or colorimetric analysis unreliable and prone to error. The Round-Robin Test has made clear the need for better guidance, training and software that is adapted to the needs of spectral imaging users in the cultural heritage community.
In this presentation, therefore, we will look specifically at hyperspectral imaging and address how the equipment, the set-up, how the system is used and how data post-processing can affect the resulting data. We will look in detail at each step in the acquisition and calibration pipeline and use the results of the Round-Robin Test to highlight pitfalls, the limitations of existing spectral imaging systems and the various sources of error and of instabilities. We will also use the results to highlight best-practice and show how cultural heritage users can take simple steps that can dramatically improve the quality of their hyperspectral acquisition and data.
The White Bastion fortress which has been standing at the outskirts of Sarajevo, Bosnia and Hercegovina since medieval period, has been virtually revived through a combination of interactive 3D virtual environments and digital stories. Remains of the fortress from medieval, Ottoman and Austro-Hungarian period have been digitized and 3D reconstructed, and then inserted in 6 virtual models of the assumed appearance of the object through time. All digital reconstructions are interconnected through digital stories in which the eternal soldier of Sarajevo is talking about the times and characters from the fortresses history.
Through user evaluation of this project we have defined answers to research questions set up in front of us at the beginning of this COSCH Case Study. The results of this evaluation are presented in form of a guide to good practice, so they can be useful for the future projects.
Alain Trémeau, Zoltan Kato, Levente Tamas, Maciej Karaszewski, Frank Boochs, Recommendations and guidelines for the use or the implementation of processing chains of colour/spectral and 3D data in cultural heritage.
Despoina Tsiafaki, A. Koutsoudis, F. Arnaoutoglou, A. Michailidou, Kantharos: the visualisation of an ancient Greek vase.
A key issue in the study of cultural assets is their often fragmentary condition. That causes serious problems and questions regarding their study as well as their presentation. Pottery fragments are the most numerous findings in every excavation and archaeological site. Furthermore, pottery plays an essential role for the reconstruction of the past, since it provides information for all the aspects of life (private, public, religion, death, economy, society, trade etc). Therefore a thorough study and presentation of the pottery fragments can lead to a better knowledge of the Past.
The focus of this work is the visualisation of an ancient Greek drinking vessel, a Kantharos, that was unearthed during the excavations at the settlement of Karabournaki (ancient Therme) in the area of Thessaloniki (Greece). The vase dates in the Archaic period (7th-6th c. B.C.) and it was found in fragments among the settlement's architectural remains. The vessel is of great archaeological significance due to its peculiarities in terms of shape, decoration and function. Therefore its digital completion and 3D reconstruction would contribute to its better study and publication.
We discuss on the 3D digitization of the Kantharos fragments that was based on the Structure-From-Motion / Multiple View Stereovision (SFM/MVS) and a custom automated data collection system. A detailed description of the digitisation pipeline is given along with details related to the quality of the 3D digital replicas of the shreds. Furthermore, we present our manual virtual reassembly and reconstruction pipeline of the Kantharos by describing the challenges, issues and ambiguities discovered while analysing the geometrical features of each shred. A number of photorealistic reconstruction visualisations of the artefact are presented in order to question the applicability of the solution for the actual reconstruction.
Emmanuelle Vila, Shadi Shabo, George Mouamar, Emmanuelle Régagnon, Remy Crassard, Alain Trémeau, "ArchéObjets3d" project – Experimental use of hand held 3D scanners for recording archaeological artefacts on the field.
3D digital modelling has become a widely adopted method in archaeology. It provides another useful approach to record, to display objects, and to virtually preserve ancient cultural artefacts. Until recently, most of the 3D imaging equipments were expensive and stationary (CTscan, MRI, THz). Thanks to the development of portable laser scanning devices, many scanners of reasonable quality that can create good resolution images are now available and easy to operate.
Archéorient laboratory in Lyon (CNRS, Université Lyon 2, France) has a long history of fieldwork in the Near East, the Arabian Peninsula, and the southern Caucasus and therefore has been working on large assemblages of archaeological artefacts from some of the most important prehistoric and early historic sites in these areas. Particularly in those regions and following excavation, the study, the preservation and the display of an archaeological artefact are challenging. Post excavation processing differs through countries, according to the impact of heritage culture, the available scientific expertise and the local conservation and storage resources. On the field, the required conditions for analyses are not always available and exportation is rarely possible. Our aim is to integrate the use of 3D scanning technology in our research for archaeological purposes and heritage preservation.
In summer 2015, a regional program Palse IPEm (Programme avenir Lyon Saint-Etienne) funded the Archéorient Laboratory in order to purchase two 3D hand-held scanners (Archéobjets3D project). This grant offered innovative research opportunity for the laboratory members. One of the scanners is used to scan large sized objects and the other enables 3D recording of small sized artefacts with high resolution precision (3D resolution, up to 0.1 mm; 3D point accuracy, up to 0.05 mm; Texture resolution 1.3 mp; Colors 24 bpp).
Our project was to evaluate the application of those hand held scanners for practical use on the archaeological field, to test the technical characteristics –accuracy of the obtained image, quality of recorded details of artefact's surface such as the texture (colour) and geometric patterns– and their adequacy with scientific exploitation of virtual collections.
The results of our experiments and the challenges encountered during the use of the scanners will be exposed, both in technical and archaeological terms. Practical examples will be presented and discussed based on different archaeological artefacts (ceramic, bone, lithic). Conclusive guidelines will address the perspectives of our approach.
Keats Webb, COSCH STSM Research: Summarizing coordinated and collaborative research on hyperspectral imaging devices and methods for cultural heritage documentation.
In recent years the application of non-invasive imaging spectroscopy techniques, specifically hyperspectral imaging, in the visible (Vis) and near infrared (NIR) has increased in the art conservation field. These techniques have expanded capabilities over that of spot (1D) spectroscopic techniques and have the ability to map the spatial distribution of materials, to extract reflectance spectra for material identification, to enhance and reveal underdrawings, to identify past treatments, and to measure colour. The increased use of these techniques, correlating with the availability of a larger number and variety of instruments, has introduced a diversity of users and usages that require reproducibility and comparability of the resulting data. There is a need for a better understanding of the spectroscopic instruments, the elements of data acquisition, and the accuracy and reliability of the data from different instruments and institutions, in addition to minimizing the knowledge gaps of imaging spectroscopy users. Working Group 1 of the Colour and Space in Cultural Heritage (COSCH) COST Action initiated a round robin test (RRT) to work towards standard methodologies and best practices for cultural heritage spectral imaging. The presentation will summarize the RRT initiative and data received, explore the variations observed with the resulting datasets, and present findings from the coordinated research initiative.
Further information about the venue and directions:
Arrival at Frankfurt airport (FRA):
All trains to Mainz are scheduled to depart from track 2 or 3 at Frankfurt Airport regional train station. Both tracks are located on the same platform. Ticket vending machines (green/turquoise) are located directly on the platform and also one level above the platform. One way ticket costs 4,65 EUR.
- S8 (with final destination Wiesbaden): every 30 minutes (29 respectively 59 minutes after each full hour)
- Regional train (final destination Saarbruecken): every two hours 8:37, 10:37, 12:37, 14:37, 16:37, 18:37, 20:37, 22:37
- Regional train (final destination Koblenz): every two hours: 9:24, 11:24, 13:24, 15:24, 17:25, 19:24, 21:24, 23:24
Additionally on weekdays:
- Regional train (with final destination Alzey): 16:09, 17:09, 18:09
- Regional train (with final destination Saarbruecken): 7:37, 15:38, 17:38, 19:38
Arrival at Frankfurt-Hahn airport (HHN):
Please check the following website: http://hahn.orn-online.de/busfahrplan_mainz.htm
How to find the Hochschule Mainz?
The easiest way to come to the Hochschule Mainz is to take one of the following bus lines from Mainz Main Station (Hauptbahnhof):
- Line 68 (direction Lerchenberg / Hindemithstraße)
- Line 69 (direction Hochschule Mainz / Campus Linie)
They start at Hauptbahnhof on platform G or F
You can buy a ticket directly from the bus driver or at a ticket vending machine. Information on public transport within and around Mainz is provided by