Electromagnetic waves from microwave to X-ray have been used in heritage science research filed. We have examined masterpieces of modern paintings among collections of the Ikeda Museum of 20th Century Art in Japan, by using THz pulsed imaging, near-infrared camera, and X-ray fluorescence, depending on the points to be discussed. "Venus and sailor (Homage to Salvat-Pasasseit)" by Salvador Dalí (1904-1989) was exhibited when he put on his first solo exhibition in 1925, and according to the artist himself, it was painted on another old painting acquired by his father, at least four times, before painting a Venus. Several intervention treatments had been carried out in the 20th century. The backside of the original canvas was lined with a thick linen cloth, and was glued to the aluminium panel tightly with wax. Thus, it is worth measuring by THz pulsed imaging to observe internal structure. Experimental results proved the multi-layer structure of the painting, and the thickness of each layer can be estimated. "Parade "by Marc Chagall (1887-1985) is a gouache painting on paper and we observed this painting by a common near-infrared (InGaAs) camera of which wavelength is around 1200 nm. Many lines of under-drawings, which are not visible by naked eyes, appeared in the infrared image. "Figure in front of the sea" by Joan Miró (1893-1983) is an oil painting, of which white parts give two different ultraviolet fluorescent characteristics. The element detection by using XRF indicated that two types of white pigments are used. One includes lead, and another was calcium based composite, which is unusual oil paintings. By using these examples, we discuss the role of electromagnetic waves for modern artworks.

Limoges enamels are a type of painted enamel, in which layers of enamel are applied to a metal underlay, most commonly copper. These types of enamelled objects are very reactive to the relative humidity (RH) and the presence of organic acids in the environment. Fluctuations in RH levels can cause weeping, crizzling, cracking and even spalling, while organic acids can result in the formation of salts on the surface of the enamels. Limoges enamel panels from the British Museum collection were examined visually and some panels were found to exhibit degradation in localised areas and colours of the enamel. An investigation combining both spectral imaging in the visible/near infrared and Optical Coherence Tomography (OCT) at various wavelengths, was conducted on a collection of Limoges enamels. Spectral imaging was employed to obtain the spectral reflectance of the enamel in order to identify the colorants of the various coloured glazes and the degree of degradation. OCT images at multiple wavelengths provide the virtual cross-sectional images of the microstructure and condition of the enamel below the surface. This showed insight into whether there were any spectral features correlating to the level of degradation on and below the surface. The investigation showed interesting correlation between the spectral features at longer shortwave infrared wavelength regime of the different coloured enamel and the virtual cross-sectional images produced by the OCT at several wavelengths.

Cleaning of easel paintings includes the removal of pollutants from the surface of varnishes or paint layers, the partial or complete removal of aged varnishes, and the removal of unwanted overpaintings. These operations aim at restoring the readability of the artwork. Conservators generally use chemical or mechanical actions to eliminate these materials. When confronted with poorly soluble varnishes or binders, conservators are forced to use aggressive chemical solutions. Their penetration degree is hard to assess and they can be harmful for the operator’s health. In the case of very fragile artworks, any mechanical action can induce irreversible damage to the original material. In these cases, control and innocuity of the cleaning operation can be challenging. Laser-matter interaction processes are an interesting alternative to traditional methods in such very demanding cases. It generally involves two distinct methodologies : the first one consists in directly ablating the layer with a pulsed ultraviolet laser [1] while the second implies laser softening of the material which can then be removed with a mild chemical product [2]. In this study, we propose a parametric comparison of these two laser-matter interaction processes for the removal of natural varnish (dammar) from well-known photosensitive tempera painting [3] (minium and lead white). For the ablation experiment, we used the fourth harmonic of a nanosecond Neodymium:YAG laser (Quantel CFR 10 ns 50 mJ 266 nm). For the hybrid approach, a commercial microsecond Erbium:YAG laser (El.En. “Light Brush 2” 2940 nm) was used, combined with a mild solvent. Layer thickness measurements prior and after cleaning was performed by Spectral Domain-Optical Coherence Tomography (Thorlabs GAN220-OCT Base Unit + Thorlabs OCTP900M Scanner) to compare cleaning resolutions with both techniques. Treatment innocuity was controlled by X-ray diffraction and pulsed Raman spectroscopy to detect any pigment modification. Finally, given that cleaning of easel paintings can be very time consuming, a comparison of the duration of the operation is proposed. [1] C. Fotakis, “Lasers for art’s sake,” Optics and Photonics News, vol. 6, no. 5, pp. 30–35, 1995. [2] A. De Cruz, M. L. Wolbarsht and S. A. Hauger, “Laser removal of contaminants from painted surfaces,” Journal of Cultural Heritage, vol. 1, no. 2, pp. S173–S180, 2000 [3] P. Pouli and D. C. Emmony, “The effect of Nd: YAG laser radiation on medieval pigments,” Journal of Cultural Heritage, vol. 1, no. 2, pp. S181–S188, 2000

In the conservation of oil paintings, the migration and eruption of metal soaps can lead to damage, but few non-invasive methods for their early detection are available. Eruptions from metal soap aggregations are generally observed only in their advanced status as protrusions and efflorescence at the paint surfaces, after they have seriously affected the appearance of the painting. The mechanisms of metal soap formation have been investigated extensively [1]: metal soaps are generated when metal-based pigments are dispersed in a drying oil as a consequence of the leaching of metal ions from the pigment and their following bonding to fatty acids. This gives rise to the formation of metal carboxylate or soap complexes that can further agglomerate in crystalline clusters and aggregates. The metal soap aggregates can increase in size, up to hundreds of micrometer in diameter, and can finally migrate through the paint stratigraphy at the painting surface [2]. The most successful method for determining the metal soap formation is through the identification of molecular vibrations in Fourier Transform Infrared Spectroscopy (FTIR) [3]. While reflectance FTIR can provide information on the surface of paintings, surface measurements cannot reveal what is happening within a paint layer, and cannot provide information on the size and distribution of aggregates in multi-layer systems. A cross-section sample is necessary to understand the chemistry within a paint, and Attenuated Total Reflection-Fourier Transform Infrared (ATR-FTIR) spectroscopy, employed for the analysis of paint cross-sections, provides detection of metal soaps through specific chemical fingerprint of crystalline carboxylates [2,4]. Metal soaps have also specific optical features with respect to the surrounding pictorial layers. In particular, they are highly translucent under visible light. This observation inspired the present research, where we propose the use of Optical Coherence Tomography (OCT) for the non-invasive detection of metal soaps and their distribution within paint layers. In fact, OCT is capable of probing subtle changes in the refraction index of a complex stratified material. The method is non-invasive and can be applied directly on top of a painting surface, providing information up to hundreds of microns below the surface depending on the paint mixture. We present results of exploratory research on the use of OCT for metal soap detection. First, we have employed OCT on paint cross-sections containing metal soap agglomerates already evidenced by ATR-FTIR mapping and scanning electron microscopy (SEM), with the aim of understanding the optical response provided by the method when probing metal soaps. The approach is then extended to the non-invasive investigation of aged paint mock-ups, with analysis applied directly on top of the painted surface. Finally, we present the first attempt of non-invasive identification of oil painting degradation on real work of art, which represents the final scope of the research. References [1] Hermans, J. J. et al, Phys. Chem. Chem. Phys. 2016, 18: 10896-10905. [2] Gabrieli F. et al., Anal. Chem. 2017, 89: 1283-1289. [3] Robinet L. et al, Studies in Conservation 2003, 48: 23-40. [4] Spring M. et al., Anal. Bioanal. Chem. 2008, 392:37−45.

Cadmium yellow, based on cadmium sulfide (CdS), was popular with artists beginning in the mid-19th century. This pigment may discolour or darken and/or exhibit loss of adhesion and formation of white globules on the painting surface. Its degradation involves photooxidation of CdS and the formation of degradation products (sulfates, oxalates, and carbonates) [1-4]. However, not all historic paints containing cadmium yellow degrade. It appears that CdS paints produced between the late 19th and early 20th centuries are particularly prone to degradation [4]. It has been speculated that the degradation of cadmium yellow may be related to imperfect synthesis methods, resulting in the formation of a more reactive form of CdS. However, the link between the presence of this reactive pigment and the tendency of the paint to deteriorate is not well understood. CdS, a IIb-IVa semiconductor, exhibits optical emission from intra-bandgap trap states (TS) [5-6]. The TS emission is related to CdS crystal defects and can provide useful information about surface states, as is routinely performed in nanoscience research [7]. Here, photoluminescence imaging, microscopy, and spectroscopy – complemented by other X-ray based spectroscopies and microscopies – have been used to examine at first historical and modern manufactured CdS-based paints following artificial aging and lastly the degradation of cadmium yellow paints in the painting Femme (Époque des “Demoiselles d’Avignon”) (1907) by Pablo Picasso. This study found that in degraded CdS paints the TS emission is much higher in intensity and shifted to shorter wavelengths with respect to preserved paint layers. This observation indicates a higher density of TS in the degraded paint, which promotes the surface reactivity of CdS particles and the subsequent paint degradation. In general, results indicated that the analysis of TS emission is a highly sensitive method for identifying early signs of degradation in CdS paints. [1] Van der Snickt, G. et al., Anal. Chem. 2009, 81 (7): 2600–2610 [2] Mass, J. et al., Analyst 2013, 138: 6032–6043 [3] Anaf, W. et al., Dyes and Pigments 2015, 113: 409-415 [4] Monico, L. et al., Chemistry–A European Journal 2018, 24.45 [5] Cesaratto, A. et al., Analytical Methods 2014, 6.1: 130-138 [6] Rosi, F. et al., Microchem. J. 2016, 124: 856-867 [7] Krause, M. M. et al., Phys Chem Chem Phys 2015, 17:18882-18894.

Optical pump-probe microscopy can provide contrast in highly heterogeneous media and a wide range of applications has emerged, primarily in biology, medicine, and more recently in cultural heritage. The localized nature of nonlinear interactions leads to high spatial resolution, optical sectioning, and large possible imaging depth in scattering media. Using pump-probe microscopy, we have studied paint samples and historic artworks in order to provide high-resolution, depth-resolved images of pigments and their degradation products. Here we demonstrate the use of this imaging method for high-resolution studies of Vermilion degradation. Vermilion (mostly cinnabar (alpha-HgS)), an important red pigment used in historical paintings, blackens over time, and metallic Hg and beta-HgS (metacinnabar) have been implicated as possible degradation products. Conventional elemental analysis techniques have trouble differentiating alpha- and beta-HgS; X-ray spectroscopic techniques can differentiate between the two phases, but obtaining sufficiently high three-dimensional spatial resolution is challenging. Pump-probe microscopy can differentiate metallic mercury, alpha-HgS and beta-HgS and map each distribution on the microscopic scale. We studied artificial degradation of alpha-HgS; femtosecond pulsed laser irradiation induces an irreversible phase shift of alpha- to beta-HgS, in which the initial presence of beta-HgS grains can increase the rate of conversion in their vicinity. Continuous ultraviolet exposure instead generates both liquid Hg and beta-HgS, with a conversion rate that increases with elevated temperatures. Finally, we reveal the presence of beta-HgS as a natural degradation product in a cross-section of discolored Vermilion in a 14th century Italian painting.

Non-linear optical microscopy (NLOM) techniques have been recently introduced for the study of cultural heritage objects due to its non-invasive nature and imaging capabilities with 3D micrometric resolution. For paintings, it has been recently reported that valuable information about composition, layer thickness and state of conservation can be obtained by NLOM [1-3]. Ensuring correct analytical protocols for the application of this novel technique requires the determination of the laser power thresholds that allow measurements under safe conditions, an aspect especially important when studying sensitive materials such as paintings. In this work, we present a novel methodology to determine the laser power thresholds for safe analyses of painting layers by the NLOM modality of multiphoton excitation fluorescence (MPEF). We also present the results obtained in a set of acrylic paints, extensively used by artists over the past century thanks to their properties and low cost of manufacture. To that purpose, samples were prepared as thin layers over a glass substrate and MPEF signals were induced with two different femtosecond laser sources: a Ti:Sapphire laser with wavelength of 800 nm, repetition rate of 80 MHz and pulses of 70 femtoseconds; an optical parametric oscillator pumped by a Yb-based laser with repetition rate of 80 MHz and dual output: at 800 nm with pulses of 100 fs and at 1040 nm with pulses of 140 fs. The excitation wavelength affects the determined thresholds and the results obtained show a strong dependence on the light absorption properties and chemical composition of the painting material. [1] Oujja M., Psilodimitrakopoulos S., Carrasco E., Sanz M., Philippidis A., Selimis A., Pouli P., Filippidis G., Castillejo M. (2017) Phys. Chem. Chem. Phys. 19, 22836-22843. [2] Liang H., Mari M., Cheung C.S., Kogou S., Johnson P., Filippidis G., (2017) Opt. Express 25, 19640–19653. [3] Dal Fovo A., Oujja M., Sanz M., Martínez-Hernández A., Cañamares M.V., Castillejo M., Fontana R. (2019) Spectrochim. Acta A 208, 262-270.

Parchments are made from an untanned animal skin, which is preserved by liming, scraping and drying under tension. Parchments are very sensitive to water, which causes in extreme case the denaturation of collagen, its main constituent, to gelatin. We implement polarization-resolved Second Harmonic Generation microscopy (P-SHG) and demonstrate that this quantitative method allows probing the early stages of collagen degradation within parchments in a non-invasive way. Nonlinear optical (NLO) microscopy advantageously provides non-invasive three-dimensional (3D) multimodal imaging of scattering samples with micrometer-scale resolution. Among the collected signals, SHG signals are specific for dense non-centrosymmetric materials and therefore provide a unique structural probe of fibrillar collagen at the micrometer scale. P-SHG allows quantitative determination of the fibrillar collagen organization: main orientation of the collagen fibrils within the focal volume and local disorder. This modality is used for the characterization of modern parchments artificially altered with dry heat. The degradation states of these samples are also determined by using the gold-standard technique: the measurement of the shrinkage temperature by differential scanning calorimetry (DSC). P-SHG appears as a quantitative tool for the determination of the collagen degradation state in a noninvasive way. Based on these results, this approach is used for the investigation of historical parchments from the Chartres’ library, that were exposed to fire and then water as a result of a bombing at the end of the 2nd World War. The visual states of the manuscripts are heterogeneous and P-SHG shows its interest for the local characterization of their conservation states. Moreover, these analyses were also performed to compare restored and unrestored parchments in this collection and the measurements revealed that the restoration performed did not alter the conservation state of the fibrillar collagen within the parchment.

Ancient Chinese glazes are colorful and mysterious, but it has been puzzling for the researchers to illustrate the coloring mechanism of glaze scientifically and characterize glaze colors specifically. In this study, the macro-spectrum measurement system and spectral measurement specification of ancient ceramics are built up. The location and distribution of the color of the specimens of each kiln are marked on the color chart. A quantitative optical evaluation method is established for the description of the color and texture of the ancient ceramic glaze. The color and texture of ancient glaze is determined by its chemical composition and microstructure. Factors affecting glaze color include chemical factor and physical factor. Fe2+/Fe3+ ions determine the color of transparent glaze, and the body color beneath the transparent glaze also modify the visual effect of the porcelain. While the coloring mechanism of opaque and translucent glaze is more complicated. The study shows that, the mild blue colors of Jun glaze and Ru glaze mainly result from the amorphous photonic structures in the glazes, which is an important breakthrough in understanding the coloring mechanism of ancient Chinese glazes. Computer simulation of glaze structure and simulative calculation of optical properties is carried out to establish the corresponding relationship between structure and reflective spectrum. Coloring mechanism for the glazes from the same kiln site, however, is not invariable, highly related with glaze composition and firing conditions. The analysis reveals that ‘oil spot’ patterns of the Jian bowl of the Song Dynasty contain large quantities of highly pure epsilon-phase iron oxide. It also has a special 2D intricate microstructure with interesting optical properties, which are responsible for the remarkable silvery look of the ‘Oil spots.’ Using plant ash as the main raw material, composite decorative glazes and high-iron crystallization glazes are designed and prepared. The spectral database of ancient Chinese glazes can also provide important reference for the non-destructive classification and authentication of ancient ceramics from different kilns.

We introduce a compact, low-cost hyperspectral camera using the Fourier-transform approach, which is based on an ultrastable birefringent interferometer, a camera lens and a 10-bits silicon monochrome CMOS camera. The spectral accuracy, resolution and sensitivity of the camera are characterized by imaging a test object made of Spectralon color materials with calibrated spectral reflectivity. At its largest lens aperture, the camera demonstrated short acquisition times, interferometric contrast remarkably higher than 84%, high spectral accuracy, and background spectral level of only 1/1000. As an application, we will report the hyperspectral measurement of important works of art, such as an Egyptian cartonnage dated between Ptolemaic and Roman Periods. The sensitivity of our camera allowed to perform HSI both of reflectance and of fluorescence. In particular, one of the ancient pigments of the cartonnage is Egyptian Blue, which exhibits unique near-infrared luminescence emission. To measure the fluorescence of the sample, we illuminated the cartonnage with a broadband LED at wavelength of 617 nm. The emission spectrum measured by our camera is in excellent agreement with the fluorescence reported in the literature. The camera allowed to map the Egyptian Blue pigment in many of the decorations. These results show the ability of our HSI system to perform fast and high-quality spectral imaging, demonstrating its potential for applications, among others, to conservation science.

The use of non-invasive hyperspectral imaging techniques has become standard practice around the world in the materials analysis and study of precious cultural heritage objects such as drawings, paintings, murals and more. However, the non-linear mixing of spectral signatures from complex and heterogenous objects with multiple colorants present below the resolution limits of the camera can complicate material identification. Consequently, ground truth measurements are still usually obtained from microscopic samples removed and embedded to expose stratigraphy and obtain sub-surface information about the artist’s material choices and technique. A microscopic hyperspectral imaging technique would map molecular information in these samples at a high spatial and spectral resolution while avoiding some of the challenges of complimentary techniques, such as swamping fluorescence in Raman spectroscopy or long integration times using FT-IR spectroscopy. This work examines the construction of a dark field hyperspectral microscope for cultural heritage samples. Using a tunable light source to illuminate monochromatically from the visible to near infrared wavelengths, diffusely reflected light was collected with a long working distance, 20x objective. We sought to de-couple the illumination and detection arms of the constructed microscope in order to fully utilize the power of the tunable light source across its whole spectrum. By mounting the optical train on a rotating arm, we can achieve multiple angles of illumination and optimize lighting conditions. The sample is also rotated in order to reconstruct an even distribution of light across the field of view. This multi-axis movement capability also provides exciting opportunities to leverage more than simple spectral information from an image series such as surface topography and differential phase contrast information. The developed microscope was used to characterize the pigment distribution in the stratigraphy of painting cross sections as well as other materials systems such as opaque glass.

UV-Vis fluorescence spectroscopy is a convenient non-invasive tool to study polychrome historical objects. It can help to identify the nature of certain materials when they present specific fluorescent properties. However, given the complexity of such stratified and heterogeneous materials, the attribution of the detected fluorescence to a specific constituent (e.g. a pigment or a binder composing a paint layer) is not straight forward. Moreover, the experimental data need to be corrected for a number of effects that can influence the recorded spectral distribution. The application of the self-absorption correction procedure (that require the simultaneous recording of reflectance profiles) to experimental fluorescence data gathered on the polychromatic surface of the Codex Borbonicus, a 16th century Aztec manuscript, is described. The results are confronted to a different methodology that is based on the hypothesis of transparent non-scattering paint layers. This second approach allows to establish more clearly the material origin of the detected emission and to discriminate apparent fluorescence (emitted by the substrate and transmitted through the paint layers) from intrinsic emission generated by the coloring materials under study. As a matter of fact, most of the various emission profiles measured in the paint layers of the studied manuscript actually originate from the substrate, and should not be used to characterize the coloring materials. This study demonstrates how carefully fluorescence spectroscopic and imaging data recorded on complex and stratified materials should be interpreted.

Pigment identification and mapping gives us insight into an artists' material use, allows us to measure slow chemical changes in painted surfaces, and allows us to detect anachronistic uses of materials that can be associated with either forgeries or past restorations. Earlier work has demonstrated the potential of a dictionary-based reflectance approach for pigment classification. This technique identifies pigments by searching for the pigment combinations that best reproduce the measured reflectance curve. The prospect of pigment classification through modeling is attractive because it can be extended to a layered medium -- potentially opening a route to a depth-resolved pigment classification method. In this work, we investigate a layered pigment classification technique with a fused deep learning and optimization-based Kubelka-Munk framework. First, we discuss the efficacy of the algorithm in a thick, single-layer system. Specifically, we consider the impacts of layer thickness, total pigment concentration, and spectrally similar pigment combinations. Following a thorough discussion of the single layer problem, the system is generalized to multiple layers. Finally, as a concrete example, we use the two-layered system to demonstrate both the impacts of layer thickness and dictionary content on paint localization within the painting. Results of the algorithm are then shown for mock-up paintings for which the ground truth is known.

Among metal alloys, bronze has been used for millennia to produce many kind of objects, from decorative and religious pieces to sculptures. Although many bronze objects have been well preserved to this day, copper and bronze can be affected by severe degradation involving the formation of copper chloride products, known as “bronze disease” [1]. These chloride products are considered responsible for the rapid decay of bronze, which can ultimately cause the loss of unique and historical artworks. However, copper chlorides constitute only a small fraction of the vast and complex class of corrosion products which can form on historical bronze and copper. It is essential to be able to differentiate between stable from more dangerous unstable corrosion products. A major challenge of bronze conservation is to non-invasively identify the unstable chloride corrosion throughout the whole artefact. This information is critical for decision-making in conservation process. Different analytical approaches are available to identify these corrosion products. However, most conventional methods require sampling which is not only destructive, but also usually unrepresentative of the object as a whole. Spectral imaging constitutes a class of non-invasive techniques which have been largely used for the identification and mapping of pigments in paintings [2]. However, it is more challenging to apply spectral imaging to 3D objects. We have developed a hyperspectral imaging set-up in the visible and near infrared spectral range for 3D objects to rapidly identify and map the corrosion products non-invasively [3]. A preliminary application has been done to study ancient bronze artworks from the National Museum of China. These results are further supported by analysis carried out on selected samples of the same collection with Fibre Optic Reflectance Spectroscopy, Attenuated Total Reflection (ATR)-FTIR spectroscopy, X-rays diffraction, Raman spectroscopy and SEM-EDX. The possibility of successfully performing hyperspectral imaging for the study of corrosion products would represent a completely novel approach in the analysis of ancient bronze and copper objects. [1] Scott, A. D. A Review of Copper Chlorides and Related Salts in Bronze Corrosion and as Painting Pigments. Studies in Conservation, 45: 39-53 (2000) [2] Liang, H. Advances in multispectral and hyperspectral imaging for archaeology and art conservation. Applied Physics A, 106 (2), 309-323 (2012) [3] Liang, H. et al. Remote hyperspectral imaging with simultaneous 3D texture mapping. Optics for Arts, Architecture, and Archaeology, 10331-25 (2017)

The exact measurement of positions is of fundamental importance in a multitude of image-sensor based optical measurement systems. A new method for enhancing the accuracy using cameras is the multipoint technique. The presentation will be focusing on the application of the multipoint technique for deformation measurement of large buildings. The goal is to reduce the overall measurement time (consisting of image acquisition and processing time) while obtaining high accuracy of position measurement. However, pixelated sensors suffer from problems which lead to inaccuracies. Among these problems are a limited pixel well capacity as well as noise and discretization. Discretization can be for example lessened by averaging the calculated spot position in time. Yet, “multiplexing in time” is not admissible for real-time applications. Multipoint techniques addresses discretization by spot-replication which is the equivalent of “spatial multiplexing”. The working principle simply upgrades a camera vision system with a computer-generated hologram (CGH) before the imaging lens. The CGH, similar to a diffraction grating, replicates the number of seen emitter positions in object space into a multiple spot pattern in image space. Together with the application of several light emitters onto the building, the CGH replicates the spots captured by the sensor to a predefined pattern. Finally, the position is reconstructed by averaging over each replicated spot pattern, where each pattern corresponds the position of a light emitter. Keeping in mind spatial multiplexing, this technique can theoretically reduce discretization and noise errors by √N, where N is the number of replicated spots. Applying this technique to the deformation measurement of large buildings can improve accuracy in time and position while using cost effective hardware. A big advantage of camera systems is that they can be installed after final construction of the building. This means that it can be used to monitor the state of antique buildings, even buildings belonging to cultural heritage. Experimental results showed an improved positioning accuracy by a factor of ~3 (where N = 16 spots).

The study and analysis of the defects of wall paintings is possible by stimulated infrared thermography, this approach is of great interest to the scientific community in charge of the conservation and restoration of these works of art. The work presented consists in identifying the optimal way to stimulate and analyze the signals measured on these works of art in order to reveal defects invisible to the naked eye such as internal decohesions. For this we compared two types Stressing of an academic sample, excitation by means of halogen sources emitting in the visible and near infrared and an excitation source emitting in the middle infrared. We compared the two modes of excitation and then applied to these two experimental approaches a post-treatment: the SVD. The results presented allowed to observe the interest of the infrared sources and the additional contribution made by a post-treatment of the SVD type.

Ancient mosaic decorations can be exposed to uncontrolled environmental conditions. The main mosaic conservation problems, induced by external factors, are related to temperature and humidity variations, which may result in moisture, salt crystallization and accelerated material deterioration within the underlying layers. Though such deterioration in the subsurface is not immediately visible, it can strongly affect the preservation of the entire structure and the decoration layer in the short or long-term future. In this paper, we present and discuss laboratory experimental results of Digital Holographic Speckle Pattern Interferometry (DHSPI), applied to the subsurface diagnostic of a mosaic model a) after artificially induced thermal change and b) under simulated environmental conditions in climate chamber. The two methods are used in alternative way first laboratory conditions with induced thermal wave by thermal lamps directly to the sample and after relaxation the sample is put inside the climate chamber in which the relative humidity follows a cycle of increase and decrease between two values. The reactions of the sample are recorded in a semi-real stepwise mode every a sufficient time period of few minutes. In both excitations the sample responds immediately and follows the change revealing inhomogeneities in construction and hidden defects within the bulk of the mosaic. The z-sensitivety of the optical geometry allows to record the shrinking and swelling displacement witnessed by interference fringes. The DHSPI system is a portable, remote-control and non-destructive interference method, already used in artwork diagnostics research. It provides in visual form immediate visualization of surface reactions and micro-deformation after induced stress, revealing subsurface discontinuities in complex objects like mosaic and different types of artworks. The tests were performed on a custom-built mosaic model with known construction and known defects, which are simulating voids, detachments and deteriorated, restored or reshuffled areas. Two types of induced stress were used: 1) low thermal excitation emitted by IR lamps directly to the surface; 2) simulated environmental conditions inside the climate chamber. In regards of the in-situ diagnostic investigation any structural alteration occurs as a natural reaction to the surrounding environment and monitoring of surface reactions and defect detection under natural fluctuations it could be considered as a very important aspect of implementation to optimise range of applicability for non-destructive remote portable instrumentation. The obtained DHSPI results on the detection of known defects in the mosaic under artificially induced thermal alteration and under simulated environmental conditions are compared in this study.

X-ray imaging is one of the oldest technic used in art analysis. X-Radiography is now completed with tomography. Used on bronze sculptures, it gives unrivalled information on state of conservation or fabrication process. The treasure of Bavay, discovered in 1969 in the north of France, is a hoard constituted by 371 bronze objects dated from the 1st to 3rd century AD. Among objects of very diverse shape and function, an exceptional set of Roman statuettes was discovered. Two Mercury statuettes were studied and, apart that they are hollow, their characteristics are very different. The metal walls are thin and even for the Lysippean Mercury and could only be done by the indirect lost wax casting process. On other hand, Indigenous Mercury still own its inner refractory core made of clay, whose shape is a very simple evocation of the statue. Here the direct lost-wax process was used, with first the core shaping, on which wax was directly carved. In some cases, spatial information is difficult to read on flat 2D radiographies, so complementary images are made thanks to 3D tomography. Reconstruction algorithms generate a 3D object issued from 720X-Ray images, showing the internal and the external surfaces. Virtual cutting can show any part of the object, allowing a detailed visualization of the inner parts. For example, Jupiter statue in tomography showed on the inner surfaces metal infiltration and a secondary casting of a lead-tin. Other analysis can be added to tomographic model: the surface 3D acquired with a scanner can add more precision on external surface representation: it was made for the Jupiter statue. In some cases also, photography for surface color analysis, or X-ray fluorescence chemical maps could also be placed on 3D models for a better materials localization.

3-D laser scanning is a novel measurement technique, which is frequently used for documentation of artworks. In the process of 3-D scanning one can obtain computing 3D models of artworks to be documented. It allows one to produce detailed digitized archives of important Cultural Heritage objects. Moreover, it makes possible digital reconstruction of sculptures and architectural fragments. One more important application is concerned with manufacturing moulds and replicas for replacements of sculptures in case their preservation requirements do not allow them to remain in their original place. This paper describes several most important case studies of the use of 3D laser scanning for replication and digital reconstruction of out-door sculptures in St. Petersburg city. One of projects was the replication of exterior marble sculpture “Primavera” (18th century, Italy) from the collection of the State museum-preserve “Tsarskoye Selo”, which is located in a small town Pushkin near St.Petersburg. It was a first case study in Russia concerned with combined use of 3-D laser scanning and CNC milling for replication of stone sculptures. Another case study was connected with creation of precise copy of marble bust of Peter the Great from the collection of the State museum-preserve “Peterhof”. Russia has donated a statue of Peter the Great to the French Academy of Sciences to mark 300 years since the Russian czar visited Paris in a pivotal moment in his crusade to open his country toward the West. The bust was placed in the building of the French Academy of Sciences in Paris, because Peter the Great was its honored member. The scanning of both original marble busts was carried out by means of a triangulation based scanner made by Konica Minolta (model Vi9i). Post-processing of the scanning data resulted in creation of STL files. Machining was done by using a 7-axis CNC milling machine, which created precise copy from a new block of Carrara white marble. When machining was complete, a small amount of hand finishing was done for removal of small, localised, drill markings. Results of one more project also will be presented. This project did deal with reconstruction of sculpture, which was destroyed during WWII. In the framework of this project digital reconstruction and physical replication of zync sculpture "Eve at the Fountain" were carried out. This artwork was the XIX-century copy of marble sculpture by English sculptor Edward Hodges Baily. Mentioned zync sculpture was part of the art collection of Maximilian Joseph Eugene Auguste Napoleon de Beauharnais, 3rd Duke of Leuchtenberg, who had country estate Sergievka near St.Petersburg city. This collection was almost completely lost during WWII as it was a place of fierce fighting between Soviet and German troops. Fragments of sculpture "Eve” were found in the earth in the process of construction works in 2007, but they represent only 7% of total square of surface of the sculpture. It does not allow to use conventional techniques of reconstruction of damaged sculptures, which are usually used in practice of restoration of stone and metal scupptural monuments. 3D model of this sculpture was created by means of both scanning of found fragments and computing “sculpturing” procedure carried out using software RapidForm, Zbrush and KeyShot. Latter did allow to create photo-realistic images of reconstructed sculpture that made possible to get the virtual replica of lost sculpture, which can be used for its display. In addition, using 3D printing (Fused deposition modeling technology) we created replica of the statue. Then its surface has been artificially decorated to imitate a historical patina. I gave an idea about original appearance of this sculpture.

OCT is used often in the field of cultural heritage studies for stratigraphy analysis of translucent layered objects and varnish layers. However, in these cases only a small (typically 1 cm2) area of the sample is imaged revealing only details on the sampled area. We increase the scanned OCT area from the cm2 to the m2 scale while still maintaining micrometer axial and lateral resolution. We accomplish this by scanning adjacent volumes with an OCT system probe mounted on large scale stages. We segment the OCT volumes to identify the primary air/painting interface. Based on the segmented surface, the volumes are stitched together in height and in angle to provide a stitch less image fusion of multiple thousands of OCT image volumes. Our technique is implemented on the “Girl with the pearl earring” painting of Vermeer. We image an area covering 0.4x0.4 meters. From the OCT data we determine the (sub) surface topology, scattering strength, and layer thickness all fused together on this large scale. In our sub meter scale imaging we identify large scale effects such as brush strokes, threads of the canvas, and long crack patterns. We observe that strong scattering regions in the OCT data correspond to high reflectivity areas of the photographic imaging. In our medium scale OCT images, we identify painting restorations since these restorations, although they superficially resemble the paintings surface properties such as height and reflectivity, they fail to reproduce the depth dependent scattering strength of the original painting. Our small scale imaging gives insight in the crack formation process. Moreover, it identifies the paintings’ glace layer and subsurface structure which indicates the way the painting artist created the strong ‘spotlight’ effect that gives this painting such a strong esthetic appeal.

Accurate measurements of geometric shape and deep internal structural details of cultural heritage artifacts are nearly always informative. However, complex but delicate heterogeneous materials often prevent or disrupt easy characterization, yielding poor interior measurements of these objects. Optical coherence tomography (OCT), a non-contact and non-destructive imaging technique based on low-coherence interferometry, shows new promise for cultural objects analysis. In OCT, the interference pattern between the backscattering light from the sample arm and the reference light from the reference arm is measured to recover the depth-resolved images of opaque materials. In this paper, we present a time-domain OCT (TDOCT) system for the paint imaging using an InGaAs amplified photodetector (PDA10DT, Thorlabs) and a commercially available 400 nm broadband supercontinuum laser with a central wavelength of 2 um. This long wavelength allows high penetration depth in high scattering material of paint samples. By using a broadband light source, the theoretical axial resolution within our system is as high as ~8.8 um. While the dynamic range of FDOCT is determined solely by the number of pixels in the detector array, our TDOCT system can achieve a much larger axial dynamic range simply by acquiring more samples per A-scan. We characterize the system performance using phantoms and mock-ups representative of real works of art. For reconstruction, we utilize regularized deconvolution techniques on the captured data by measuring the Point Spread Function (PSF) of the system. This deconvolution process can increase the depth resolution and provides better image quality than standard OCT alone. Our system provides an economical and practical way to achieve high performance imaging for cultural heritage applications in terms of penetration, resolution, and dynamic range.

Optical coherence tomography (OCT) is a well-known and in-demand technique for non-destructive testing and inspection of art and cultural objects. However, OCT has several restrictions on the applications. All the commercial available OCT systems operate in the near-infrared and visible spectral ranges. Due to the Mie regime of scattering, OCT faces physical limitations for the investigation of highly turbid materials such as paintings and ceramics. Several attempts were done to overcome the barrier using intelligent optics approaches and post-processing; meanwhile the intense developments of novel, bright and spatially coherent mid-infrared supercontinuum sources allowed to shift OCT spectral window towards longer wavelengths where the scattering is in minor effect. Nevertheless, there are still only a few solutions that are not popularized due to challenges of this range, such as a lack or high prices of detectors, optical components or a lag in the specification. Besides the reduced scattering, the mid-infrared range provides valuable information about the chemical compositions due to the fundamental vibrational absorption lines in this spectral region. That enables to open new possible applications of the system combing OCT with mid-infrared spectroscopy modality to get chemical as well as structural information about the object. In our work, we developed the first Fourier-domain multi-modal mid-infrared OCT that operates in the wavelength range around 4 um and performs co-registered spatially resolved spectroscopic measurements in the range of 3.1 um – 4.4 um. The system is based on a low-cost thermal-based pyroelectric detector and supercontinuum source covering the spectral region from 1 um to 4.5 um. Applying advanced post-processing and detection approaches, we demonstrate the sensitivity of the system around 82 dB, the lateral resolution of 30 um and axial resolution around 35 um. The specifications allow us to achieve the enhanced penetration depth into the test samples: oil paint objects and ceramics, while the registered spectra allowed to differentiate the types and constitution of the paints. Therefore, the system has shown a high potential of a new spectral window for art diagnosis and presents the low-cost and available solution.

We describe a novel technique for simultaneous measurement of refractive index and dispersion using optical coherence tomography (OCT). The technique is applicable to in situ non-invasive refractive index and dispersion measurements of all transparent or translucent materials with flat surfaces (e.g. cubes, wedges etc.) through a multi-angle approach giving typical precisions of 0.01 for the refractive index. The method can be applied to the restoration of modern sculptures made from plastics where it is important to measure the refractive index of an object using a non-contact method, so that a resin-based adhesive with a matching refractive index could be found to restore a scratched or broken sculpture [1,2]. A seamless restoration depends on the matching of the refractive indices between the object and the adhesive in the visible part of the spectrum, since scattering at the interface decreases with increased match in refractive indices. Previous studies found that the match in refractive index needs to be better than 0.02 for adhering broken pieces and better than 0.04 for superficial scratches [1,2]. Most OCT operate in the near infrared, however, for the best visually seamless restoration, we need to ensure a match in the refractive indices in the visible part of the spectrum centred at ~550 mm. We present the results using an in-house developed 550 nm OCT with axial resolution of ~0.8 microns and transverse resolution of 5 microns. [1] A. Lagana and T. B. van Oosten, “Back to transparency, back to life: research into the restoration of broken transparent unsaturated polyester and poly(methyl methacrylate) works of art,” in ICOM-CC 16th triennial conference Lisbon 19-23 September 2011: preprints, (2011). [2] A. Lagana, R. Rivenc, J. Langenbacher, J. Griswold, and T. Learner, “Looking through Plastics: Investigating Options for the Treatment of Scratches, Abrasions, and Losses in Cast Unsaturated Polyester Works of Art,” in ICOM-CC 17th Triennial Conference, (2014).

Typically, varnish coatings were applied to Old Master paintings to improve their visual appearance, saturating the colours. However, over time, varnishes deteriorate and discolour, obscuring or altering the appearance of the artwork, necessitating their removal (usually through the use of a solvent) and replacement. Varnish removal presents a degree of risk to the paint surface and may result in loss of original materials or other damage. Thus, diagnostic methods that can support conservators during varnish removal are very valuable. Optical Coherence Tomography (OCT), a white light interferometry technique based on the Michelson interferometer, has been proposed as an ideal tool for monitoring the cleaning and treatment of artworks and historical objects. This is owing to its ability to provide non-invasive and accurate measurements across the entire surface of an artwork of the stratigraphy of transparent and turbid multi-layered structures. In many cases, however, difficulties may arise when attempting to distinguish degraded glaze layers on paintings (coloured, translucent layers containing pigments with refractive indices very close to the binding medium) from layers of aged varnish due to similarities in their appearance and scattering properties in OCT scans. In order to mitigate the possibility of damage through the accidental removal of glaze layers, we propose a simple and entirely non-invasive technique to identify the spectral features of a semi-transparent layer using a combination of OCT and spectral imaging in the visible range. Building on preliminary studies 1,2, the technique utilises an algorithm to automatically obtain the thickness distribution of the surface layers from the OCT volume of a region with relatively uniform pigmentation and colour. This information is then coupled with the spectral data from the same region to derive the spectral characteristics of each translucent surface layer. Such characteristics may then be used to the discriminate between aged varnishes and glaze layers. The potential of this approach has been demonstrated on mock-up samples and paintings from the National Gallery in London. [1] Lange, R., Liang, H., Howard, H. and Spooner, J., “Optical coherence tomography and spectral imaging of a wall painting,” SPIE Newsroom (2011). [2] Liang, H., Lange, R., Howard, H. and Spooner, J., “Non-invasive investigations of a wall painting using optical coherence tomography and hyperspectral imaging,” Proc. SPIE 8084, 80840F (2011).

The analysis of painting materials and techniques provides important information to history and archaeology research, as well as to scientists and conservators involved in the conservation of objects of cultural heritage. However, the examination of statistically significant number of objects is required to understand the material use typical of a class of objects (e.g. paintings from a certain geographic region, in a certain period, for a certain purpose). This necessitates efficient data collection and the development of methods for their effective analysis. The spectral imaging system developed in our group, PRISMS, enables automated high resolution spectral imaging in the visible/near infrared regime of paintings of any size. The automatic clustering of the spectral reflectance data at the pixel-level can be used for the initial classification of the areas according to their reflectance spectra. In this study, a new clustering algorithm, based on self-organised mapping (SOM), for the sequential automatic analysis of large spectral datasets is presented. The development of such methods makes the analytical procedure more efficient by reducing the number of areas that need to be further examined based on their unique reflectance spectra. The preliminary clustering of the spectral imaging data is supported by the high-resolution x-ray fluorescence (XRF) maps. The multimodal spectral characterisation of the painting materials is completed with the application of high spectral resolution Fibre Optic Reflectance Spectroscopy (FORS) and Raman spectroscopy on each cluster area. Optical coherence tomography (OCT) is used to examine the 3D microstructure of the substrates as well as the painting technique through the examination of the stratigraphy. Examples based on data from collections of Peruvian paintings on paper substrate, Chinese export paintings and Peruvian paintings made in the style of Chinese export painting will be presented.

This study presents a novel multivariate strategy that can be used for automatic on-site processing of mid-FTIR total reflection macro mapping data obtained from investigation of artworks. Mid-FTIR total reflection macro mapping has been scarcely used in the conservation field and to the authors’ knowledge, this is the first time that the technique was performed to characterize metal substrates. The chemometric strategy developed, based on the integration of PCA, brushing and clustering approaches, allows to automatically identify the regions of interest (ROIs) and to extract the related average spectrum. Thanks to the automatic data management, HSI (hyper spectral imaging)-based analyses may be exploited by personnel without specific advanced chemometric knowledges, such as conservation scientists or conservators with a scientific background. The analytical protocol was tested as a rapid way to evaluate the conservation state and the performance of cleaning methods on bronze objects. Both activities are commonly planned during the restoration campaign and require an on-site analytical procedure for a quick and easy diagnosis. The performances of the method were evaluated on aged standard samples (bronzes with a layer of green basic hydroxysulphate treated with different organic coatings) and on site on areas of the Neptune Fountain (Bologna, Italy, 16th century).

Scanning macro‐X‐ray fluorescence (XRF) spectroscopy on works of art provides researchers with rich data sets containing information about material composition and technique of material use in a compelling visual format in the form of element‐specific distribution maps. Collection these maps, however, typically assumes a two dimensional object, and few works of art are truly flat. Small nuances in elemental intensity may be introduced into element distribution maps by surface topography. The inability to confidently ascribe a change in signal intensity to actual elemental composition versus topographically-induced variance, therefore, presents a challenge, particularly when attempting to identify markers of artists’ techniques, compare several objects, or overlay/register images from scanning XRF with those from other imaging modalities. This paper introduces a new methodology for post-processing scanning XRF data sets to correct for elemental intensity variations as a function of topography. The method augments the acquired XRF data based on a three-dimensional reconstruction of an object and a set of elemental intensity/distance response functions. These response functions act as a calibrated guide for modifying the intensity map based on depth variation. The geometry-based parameters of local surface shape (curvature), distance of the XRF detector from the surface, region of intersection of the incident fluorescence beam with the surface, and the orientation of the incident beam with respect to the surface normal, are each accounted for in the calibration phase as a large set of pre-acquired examples. This provides a mechanism for capturing and understanding the anticipated variations in the macro-XRF data, interpolating the examples in order to smoothly estimate variations, and applying those variations as corrections to macro-XRF data collected on non-planar surfaces. Macro-XRF variation as a function of the geometry is explained, with an emphasis on understanding the parameters that induce the most severe errors in the XRF estimates. The representational framework for collecting, storing, and summarizing calibration data over a large number of scans is discussed, followed by several proof of concept examples.

Infrared thermography is a non-destructive testing technique that affects many areas. This technique of analysis is, for example, very interesting in the field of restoration and conservation of heritage works. The possibilities of active thermography can help in the early detection of defects in works of art and their characterization. In this work we will demonstrate that it is possible to detect old restorations in frescoes and murals by flash method. This new possibility offered by thermography will make it possible, for example, to verify the compatibility of a restoration with the original work, which can prevent the appearance of a defect and may allow the conservator to follow up on restoration. In first we will demonstrate the feasibility of this approach on a laboratory specimen containing different types of clogging materials and then present results of analyzes carried out in situ at one restoration sites that demonstrate the possibility of locating these restorations but also to characterize the pictorial technique used in this one.

Israel in Egypt is an oil on canvas painting by Sir Edward Poynter, in 1868 the painting is depicted in the Illustrated London News in which there are reports of changes made to the composition of the painting after its first exhibition. Visible and infrared imaging techniques are used to determine whether additions to the initial composition can be identified from underdrawings. The painting measures 137 cm x 317.5 cm and was not able to be relocated for the study, therefore portable imaging equipment was used throughout. A Canon 700D DSLR camera was modified to allow near-infrared imaging when combined with a set of longpass filters at 720 nm, 850 nm and 950 nm. An Osiris infrared reflectography camera was also used to look further into the infrared with a sensitivity range of 900 nm – 1700 nm. To obtain high-resolution images with the modified DSLR, a 100 mm lens was used from a distance of 6 metres. In both visible and near-infrared, eight images were taken across the surface of the painting and these images have been combined into high-resolution visible and near-infrared panoramas. Images from the Osiris infrared camera have been processed in Matlab to create a mosaic from the overview image with high-resolution regions of interest. All processed images were registered in Matlab along with the woodcut engraving of the painting shown in the Illustrated London News. Machine learning was then investigated as a technique to increase the resolution of the Osiris infrared overview image, using the high-resolution DSLR mosaics as inputs and an area of the Osiris infrared mosaic for a training area. An interactive web-browser viewer was created to enable display and comparison of the registered high-resolution images, allowing users to explore and zoom in to specific areas of interest across the four high-resolution images simultaneously. Conservators and art historians can utilise the resulting images combined with the image viewer to analyse the painting and potentially develop a new interpretation of the composition.

The digital image processing allows respecting principles of correct restoration theory, as it is possible to obtain different kinds of re-integration such as sottotono, tratteggio, and puntinato or even a chromatic reconstruction suitable for archaeological restoration. The main purpose of this work was to analyse the different methodological and practical aspects of Tattoo Wall® technique, especially the possible changes due to ageing artificially induced by extreme humidity and/or solar irradiation. For the experimental tests, we chose to work on a colour scale as wide as possible, to test each single colour, and on different materials. In particular, two kinds of mortars were chosen as support: a traditional one, made of lime and marble powder, and a hydraulic mortar, containing marble powder combined with Ledan C30, particularly suitable for restoration in environments with high relative humidity (RH%). The different samples produced were aged under UV irradiation and in a close box with 92% of RH. To obtain the UV ageing, samples were placed in a Solar Box chamber under the following conditions: temperature 55 °C, irradiation power 550 W/m², and UV filter at 280 nm. Six cycles of 168 hours were performed. Moisture ageing was obtained by placing the sample in a box with RH% fixed to 92% thanks to the presence of salts (sodium sulphate deca-hydrated) for compressively two years. Reflectance spectrophotometry for colour measurements, Fourier Transform Infrared spectroscopy (FT-IR) and hyperspectral imaging (HSI) were used to assess artificial sunlight influence. The experimental data were statistically treated in order to evaluate their significance. Testing enabled us to verify the high stability and durability of Tattoo Wall® under extreme solar irradiation and high relative humidity, with almost no chromatic alterations.

Stone monuments are a high part of our Cultural Heritage and their preservation is of crucial importance because of their historic and cultural significance. Weathering affects naturally and fatally stones which made buildings, they combined several effects that lead to mechanical, chemical and biological deteriorations. Life on stones as microbial communities colonized on first substrates freshly exposed to the atmosphere where climatic conditions as sunlight and water availability, temperature are the key factors to microbial development. Microbes are the most important pioneering colonizers to settle on stone. Microbial attachment on surfaces with mineral dissolution is a rapid process and provides easily available nutrient source. The formation of biofilms embedding microcolonies in an organic polymer matrix, favours the settlement of other micro-organisms and the deterioration of the substrate that varies from the discoloration, production of organic acids, salting that induces physical damage and helps to the development of macroscopic vegetation. Detecting earlier the pioneer development of biofilms before the greening of stone and its visual damage is an important key in the conservation of Cultural Heritage to take action at the beginning of the development and limit the cost of monument restoration. For this purpose, IRT has been used as a non-destructive technique to detect pioneer biofilms on stone. Experiment has been displayed on two limestones, Courville and Savonnieres stones used in major buildings in eastern France and the surroundings of Paris. Stone samples have been exposed in an outdoor test to favour the natural colonisation by bacteria. Then three samples for each stone have been collected after few months and compared to three non-colonised sterile stones throughout IRT measurements of stone surfaces stimulated by a flux of photons. The absorption of photons increases the temperature of samples surface that is recorded by a bolometer camera (FLIR SC655). Then mathematical post-processings has been carried out in the aim at highlighting thermal variations. As a result, biofilm on stone has been detected by many post-processings like SVD, PPT and temporal statistics. Results varied from internal properties of both stones. Post-processings detected better biofilm on Savonnieres which is a macroporous stone, than Courville stone which is a microporous stone.

The basic material of library collections is paper. There occur also other kinds of materials in libraries, especially collagen materials – parchments and leather, but also textiles, and in modern collections also plastics. Project "Advanced techniques of cleaning of books and manuscripts" is supported by the Ministry of Culture of the Czech Republic in 2018-2022. The main goal of the project is to develop techniques suitable for cleaning of the mentioned library materials by laser and by carbon dioxide snow. Cleaning will be carried out on model samples at first, then on real library objects. To optimize parameters of both cleaning techniques, it is fundamental to select method of evaluation of their effectiveness in removing dirt and dust, and at the same time the method of observation, whether undesirable changes in cleaned materials do not occur during cleaning, i.e. their degradation. Measurements of changes in color using spectrophotometer Konica Minolta CM-700 d in L*a*b*color space, in particular evaluation of parameter L*, serve for the first orientation evaluation of effectiveness of cleaning. Multifunction microscope Hirox RH 2000 provides 2D and 3D pictures of surface of cleaned materials. It also serves for measurement of changes in roughness and wetting power of material surfaces. Use of scanning electron microscope imaging is of advantage for finding the degree of deterioration of integrity of collagen fibres. Degradation of leather and parchments will also be monitored by measurement of changes in shrinkage temperature using hot table microscopic technique. After finding the most suitable cleaning technique and its parameters for individual library materials, a variable line will set up for cleaning the library collections, which is the planned actual result of the project. This work was supported by the Ministry of Culture of the Czech Republic under the grant Advanced techniques of cleaning of books and manuscripts, identification code DG18P02OVV048.

The Agris Helmet is a masterpiece of the Celtic art made in the middle of the 4th century A.D. It has been found near Agris, France, in 1981 during an archeological dig under the supervision of José Gomez-de-Soto. This composite object is made of an iron cap covered with bands of bronze. The bronze is itself covered with pure gold leaf, with embedded coral decorations attached using silver rivets. The helmet, discovered in pieces, was restored in Mainz during the 1980th and is now held by the Musée d'Angoulême in Angoulême (France). The present project led jointly by the AOROC department of the ENS and the C2RMF and supported by PSL, was an opportunity to do complementary work on the Helmet to document it thoroughly. Our aim was threefold. First, we wanted to document the Helmet in such a way that future studies would be doable even without physical access to the original piece. Then, we were willing to search the best 3D technique and practice for this complex piece. In addition to these technics, we wanted to enhance the knowledge about the object (material and technical identification) and provide new data for archaeology and art history. We began the study by doing a complete photographic coverage of the Helmet thus highlighting every single visible details on its surface. In order to go further concerning the structure of the object itself, we chose to use X-Ray technology. This technology allowed us to see behind the surface and understand the organization and construction of the Helmet. Furthermore, multiple techniques of photogrammetric imaging and 3D digitalization (with differents sets of parameters) were used to obtain the best set of data. For the study of the Helmet, the high-resolution models provided information about tools marks and other traces of the production techniques. The results of those studies were particularly instructive for the historians of the Celtic period and gave us new possibilities to the researches for the work on this precious object.

The complementary use of X-ray fluorescence (XRF) mapping, spectral imaging, and Raman mapping, allows for the analysis and identification of important artistic materials used in the production and illustration of illuminated manuscripts. This project uses combined non-invasive imaging techniques to analyse 17th – 19th century manuscripts from the British Library’s Southeast Asia Collections so that more can be understood about the adoption and evolution of artistic materials and techniques used in Maritime Southeast Asia. Using multiple different imaging techniques has shown to provide positive results, however, a consequence of this is the collection of large amounts of data, necessitating the automatic and unsupervised analytical techniques used in machine learning. Data collected in-situ at the British Library using macro-XRF mapping, macro-Raman mapping, and Spectral Imaging, will be analysed using a range of machine learning techniques to cluster pixel information representing materials used in southeast Asian manuscripts.

The work presents preliminary results of the multianalytical investigation for one of the frescoes of Raphael’s workshop, “Venus Tying a Sandal”, from the State Hermitage Museum collection. The frescoes were overpainted and restored several times. Previous restoration and research of the fresco were carried out at the Hermitage Museum in 1975-1978. Microchemical methods, IR spectroscopy, and TLC were used to determine the composition of pigments, binding media, and conservation materials. In the present study non-destructive and non-invasive techniques were applied: X-ray radiography, UV luminescence, IR reflectography, and XRF analysis. Using IR reflectography allowed us to find the alteration of some details of composition (for example, the drawing line on fingers of Venus's right foot). Samples taken during previous and present investigations were studied by means of optical and polarized microscopy, molecular spectroscopy, and SEM–EDX. It came as a surprise to conservators that painting contained ground layer consisting of mix of calcium carbonate, gypsum, different types of silicates, and particles of charcoal. This was uncommon from the fresco technique standpoint. Most probably, this is due to the fact that the painting was transferred to a canvas and conserved multiple times. In the course of the study of the paint layers some conservation materials containing calcium carbonate, gypsum, kaolinite, zinc white, barium white, lead white, and Prussian blue were identified. The author's materials included calcium carbonate, gypsum, ochre, celadonite, cinnabar, smalt, and Naple yellow. Data from previous restorations of the frescoes together with new research data, acquired with the help of modern analytical methods, enable us to understand the condition of the investigated painting more precisely and get more insights into applied conservation techniques and materials used by the respective authors.

During the David Lynch’s exhibition Silence and Dynamism displayed at the Centre of Contemporary Art Toruń paintings of the cycle with the artist’s alter ego images of the figure “Bob” from the period 1996-2001 were investigated in order to propose proper conservation treatment. Investigation began with the thorough observation in the ViS and UV light and portable microscope observation. Non invasive elemental analyses were conducted using portable XRF microscope revealing some of traditional pigments like cadmium red, iron black, ivory black as well as titanium white. Significant amounts of silica were identified in the industrial adhesive used as paint. During continued research artist delivered company names of products, gave permission for sampling the objects and delivered model materials of the adhesive and of the carpenter glue. In the next step SEM-EDX analyses as well as FTIR conducted on samples enabled further recognition. The main painting material - Taylor 2080 Adhesive - was identified as chlorinated polyethylene close to the industrial product Agitan P 813. McCloskey (Man O’War Marine Spar Varnish) used instead of typical artists varnishes proved to be alkyd resin what is in accordance to the producer description. The amorphic silica indicated by the producer was identified and visualized in the SEM-EDX. In carpenter’s glue - Titebond III Ultimate Wood used by artist for “consolidation” repairs polyvinyl chloride based glue was detected as the main component. In the primer acrylic medium with chalk and titanium white was used. Investigation revealed wide use of non-artistic materials by David Lynch. Recognized materials enabled to propose modified buffered solution of controllable pH 5 and conductivity 5,5 mS resulting in successful cleaning of the works.

The bronze dagger in the richly decorated scabbard was recently discovered in the process of archaeological research in the burial which belongs to the Sakas culture, and dates to VIII-VII centuries BC. The scabbard have been decorated with granulation and incrustation. Incrustation were in bad preservation: stones inserts were lost, and only traces of it are preserved. In addition, some preserved minerals in the process of degradation dramatically decayed, changed surface color and now look brown-gray. In the process of conservation, were taken the samples of deep blue, blue-green and gray colors minerals and mastic adhesive. Microparticles of minerals were selected at the tips of sharp wooden sticks with polymer adhesive. But the polymer makes strong luminescence when samples studied by Raman. It consisted of two stages. At the first stage, a luminescent map was recorded under excitation at 488 and 532 nm. Then the heterogeneities on it were determined. At the next stage, the Raman spectra with a 785 nm laser were recorded with these inhomogeneities. The deep blue mineral was identified as azurite. Natural lapis lazuli was identified on the gray inserts. This mineral decayed and changed color. Turquoise was identified on the blue-green inserts. The samples of mastic adhesive which attach stones were additionally investigated by IR spectroscopy. The research results provide data for better understanding of the original image of ancient scabbard richly decorated with inlays. The location of the lost incrustation was recognized by the smallest particles that preserved. We emphasize that natural lapis lazuli and azurite have been used for incrustation. We can argue whether this was due to the original artistic design in the manufacture of the scabbard, the intentional replacement of expensive lapis lazuli with azurite, or the repair of inserts that had lost in ancient times.

The detection of archaeological sites has in recent decades largely relied upon methods in either aerial photography, or satellite imagery. Whilst aerial photography has enabled a much-improved understanding of the distribution and density of archaeological sites, they provide limited information both in terms of the material composition and if soil/crop conditions do not happen to favour site detection. Multi-spectral and hyperspectral imaging can enable and enhance the detection of otherwise hidden archaeological sites, as well as enable condition monitoring of inaccessible historical and architectural exteriors, but with such data being typically selected from satellite images, with limited spatial resolution, this potential has not yet been fully exploited. The Imaging & Sensing for Archaeology, Art history and Conservation (ISAAC) Lab at Nottingham Trent University has, in recent years, developed and improved a spectral imaging system (PRISMS) which can remotely image historical objects and interiors, such as large wall paintings, at long range (tens of metres). Such a system provides a reflectance spectrum per pixel, ranging from the visible to the near-infrared, from which materials can be identified using automated data processing software and machine learning approaches. A drone-based miniaturised adaptation of this current system is being developed for long-range spectral imaging of large-scale archaeological sites and architectural exteriors. The development of automated software using machine learning techniques, optimised specifically for outdoor data capture and site feature detection including the detection of specific archaeological deposits, is also underway. We present current progress in the development of this drone-based spectral imaging system for archaeological applications.

Microscopic mid-infrared imaging spectroscopy (mid-IR, 4,000-350 cm-1) is used for the mapping of chemical functional groups associated with artists’ materials in paint cross-sections using an ATR arrangement. There is interest to apply this technique on a larger scale in order to map artists’ materials over the surface of a painting. The limitation to imaging on the macroscale is that no contact can be made with the artwork. To work in non-contact mode, reflectance mid-IR spectroscopy is often done using a thermal source having a temperature above ambient in the laboratory. This ensures the mid-IR radiation reflected off the object under study is greater than the emitted radiation by the object under study or the thermal background of the laboratory. Macroscopic mid-IR imaging spectroscopy using reflectance mode has been shown to allow the collection of usable mid-IR image cubes from paintings, although the quality of the mid-IR spectra acquired is often hindered from the need to limit heating of the artwork by the thermal source to < 4 C during the data acquisition. In mid-IR remote sensing imaging spectroscopy of the earth, the dominant source of thermal radiation is the emitted radiation from the object under study, rather than the reflected radiation off of it which comes from the cold overhead sky. Here we show by placing a highly reflective tube between the painting and camera and introducing a low temperature source to simulate the sky radiance, the thermal radiation from the room can be reduced, allowing the IR radiation emitted by the painting to dominate. Thus, emissivity spectra of the object can be recovered. Using this technique, mid-IR emissivity image cubes of paintings were collected at high collection rates with a low noise, line-scanning imaging spectrometer which allowed pigments and paint binders to be identified and mapped.

We present the developments of a mobile remote spectroscopy system, consisting of Raman spectroscopy and laser-induced breakdown spectroscopy (LIBS) that work in the range from 3 to 15 m, for wall paintings and architectural interiors. Remote Raman spectroscopy identifies molecular structure for wall paintings at stand-off distances and inaccessible heights (e.g. churches, caves and tombs). Complemented with LIBS, which reveals the elemental composition of the materials, this remote spectroscopy system is able to provide complementary information for material identification. Careful considerations have been given to the design of the system to meet the special needs for the study of wall paintings: 1) efficient Raman measurements at low intensity to avoid laser induced damage; 2) able to operate indoors under natural light. The system operates with a continuous wave (CW) laser at 780 nm for Raman and a pulsed laser for LIBS. A telescope is used for the collection of both Raman and LIBS signals. In comparison with conventional benchtop micro-Raman instruments, similar signal to noise ratios (SNR) and sensitivities are achieved with the remote Raman system at 4 m. With a large laser spot at sample (4 mm at 4 m) and therefore much weaker laser intensity (5-6 orders of magnitude lower than those of typical micro-Raman systems), the risk of laser induced degradation is significantly reduced. For Raman, a 780 nm CW laser is employed for excitation, which is less prone to fluorescence. For rapid response to the changing daylight, an automated online daylight subtraction program is developed. Compared with typical ns-pulsed lasers with repetition rate of 10-100 Hz, a CW laser could have a measurement advantage of 7-8 orders of magnitude in efficiency, if the damage threshold in intensity is comparable to that of a typical CW laser in remote operation. Assisted with a computer-controlled telescope mount, small area remote Raman spectroscopic mapping (2D scanning) is also realised to complement long range remote visible and near infrared spectral imaging for material identification in our mobile lab campaigns.

A mid-infrared spectroscopic imager needs to be portable and tough for the identification of dyestuffs used for murals in ancient tombs during archaeological on-site analysis. Meanwhile, an extremely compact and tough hyperspectral camera with mass less than 2 kg is required for mounting on drones to observe nutritive components like nitrogen and phosphorus. However, a conventional Fourier-transform infrared (FTIR) spectrometer is large and expensive (e.g., HyperCam, made by Telops Co., has a mass of 30 kg and costs 1.5 million USD). Because the Michelson interferometer is not robust against mechanical vibration, a mechanical antivibration system is inevitably needed. We proposed a near-common-path wavefront-division phase-shift interferometer as an imaging-type two-dimensional Fourier spectrometer [1]. Because the proposed interferometer has strong robustness against mechanical vibrations, a palm-sized Fourier spectroscopic imager can be realized without an anti-mechanical vibration system. We developed a palm-sized (80-mm cube weighing 0.5 kg) and tough hyperspectral camera for mid-infrared light (wavelength of 8–14 µm) that can be operated using only a notebook personal computer. Furthermore, the field of view of a conventional hyperspectral camera is narrow (e.g., 6.4 deg × 5.1 deg). However, employing a proposed field angle correlation method [2] and using a fisheye lens as the objective lens, the field of view can be expanded to 180 deg. The total price of the mid-infrared two-dimensional spectroscopic imager is no more than several thousand USD because a low-price microbolometer (Vision Sensing, VIM-80G2, wavelength range: 8–14 µm, 80  80 pixels, price: 300 USD) is used. Additionally, a long-stroke (10 mm) and high-resolution (Optical encoder resolution: 100 nm) impact-drive actuator (Technohands XCWT70-10 weighing 30 g) is introduced as a low-price (1000 USD) and tough phase-shift stage with cross-roller linear-motion guides.

This paper investigates the capability of optical coherence tomography (OCT) as an advanced, non-invasive method for 2D and 3D tomographic imaging of the surface and stratigraphy of glass cultural heritage using a commercial ThorLabs Ganymede II Fourier domain, 930 center wavelength system with axial resolution of 4-6 µm. Results from historical glass, in conjunction with model alkali silicate glasses subjected to artificial aging at 90º C and 90% RH, represent aged surfaces from common unstable glass formulations used the 19th century for objects such as glass musical instruments and photographic substrates and encasement materials. Results demonstrate that OCT is a powerful technique for characterization of hydration layers, as well as surface and subsurface cracking, which commonly develop in alkali silicate glass upon natural aging in humid environments. The images also contain phantom features, however, that are the product of either multiple reflections or image rendering, and thus must be interpreted with care, as discussed. In order to better assess the potential of OCT for the analysis historical glass, results are also compared to other methods, including destructive analysis by scanning electron microscopy with backscattered electron imaging (SEM-BSE) of cross sections, and non-invasive light microscopy and NIR fiber optic reflectance spectrometry (FORS). Shape and depth of the glass hydration layers as determined by OCT are compared to SEM-BSE information from the metal-leached layer. Results also show how light microscopy aids OCT volume and surface rendering interpretation, while OCT helps ascertain subsurface cracking and the depth of FORS detection, thus greatly enhancing our ability to detect early stages and progression of glass deterioration. Given the ease of use and affordability of commercial OCT systems, the technique provides an important means with which to analyze glass and make preservation decisions, such as environmental design and potential treatments.

Imaging spectroscopy (IS) is the collection of spectral signatures across a sensor’s field of regard from which baseline chemical composition of materials can be obtained without the need for sampling. This makes IS a powerful chemical sensing tool for the study of paintings of archaeological and historic importance. In this research, two imaging spectroscopy sensors were used to analyze ancient and Byzantine Cypriot wall paintings of funerary, civic, and religious context within the district of Paphos in the southwest region of Cyprus: (1) a custom single pixel scanning reflectance spectroscopy (SRS) system assembled from a UV/Vis/SWIR (350-2500nm) ASD FieldSpec3 fiber optic spectroradiometer coupled to a low cost 2D mechanical scanning arm, and (2) a highly portable Vis/NIR (400-1000nm) hyperspectral imaging spectroscopy camera (HSI) – the IQ by Specim. Both systems were used in situ to collect 3D spatial-spectral cubes of monumental paintings in reflectance and luminescence imaging modes. Data analyses on wall paintings from two important sites, the Roman tomb 3882 and the church of Agia Paraskevi, highlight the advantages of these field-deployable imaging spectroscopy systems for material characterization and interpretation. The SRS system produced high spectral and low spatial resolution 3D data cubes from the Vis to SWIR and 2D chemical maps of features present in the scanned regions of the paintings. The Specim IQ also produced 3D spectral cubes within the sensor’s field of regard with higher spatial resolution and demonstrated a significant advantage with its ease-of use field applications. Additionally, the IQ system acquired NIR luminescence data sets to identify and map the ancient synthetic pigment, Egyptian blue. The systems used in tandem revealed the palette of ancient and Byzantine artists and provided visualization of pigment applications, lending insight into artistic choices and painting techniques within their respective historical contexts.

In this work, case studies from a number of works of art demonstrate the ability of scanning macro X-ray fluorescence (MA-XRF) spectroscopy for the non-invasive examination of the physical characteristics and elemental composition of historic gold leaf in works of art. Macro-XRF scanning has opened up new avenues of research by providing insight into how these micron-thin sheets of gold were manipulated and applied in works of art. Besides elemental composition, data provided by MA-XRF includes visualizing the shape, size, and application techniques of individual gold leaves. For example, measurements of the dimensions of individual gold leaves obtained directly from XRF map data of 15th century panel paintings reveals differences in the dimensions of gold leaf across different administrative regions throughout the Italian peninsula. This work suggests a deterministic system of leaf production, which varied between the city-states. In addition, a comparison of the degree to which gold leaves were overlapped during gilding reveals clues about the artistic hand of individual artists. Taken together, these measurements provide a material understanding of artistic practice. Our knowledge of how gold leaf was manufactured is based on historical treatises rather than material analysis. The ability to confidently detect variations in minor and trace elements in historical gold leaf may provide new tools to better understand dating, location of production, and trades. The suitability and inherent limitations of MA-XRF for the semi-quantitative analysis of the gold leaves, based on a feasibility study using a set of modern gold leaf samples will also be discussed. A ground truth for the concentrations of these, and other trace, elements was validated using inductively coupled plasma-mass spectrometry (ICP-MS). Linear regressions for gold, silver, and copper provide a model for relating XRF intensity to concentration that can be tested on the XRF map data obtained from the historic artworks.

Items made of glass beads had been the important part of humanity everyday life, culture, and global trade. The beginning of industrial-scale production of glass beads in the main glass-making centres led to the beadwork art explosion in the mid 18th century. This craft was extremely popular in Europe and Russia. But the glass is often unstable material, so the problem of historical beadworks preservation from museum collections is an actual issue for curators and conservators. Preliminary the samples of historical glass beads of different colours and conservation states were studied by means of XRF, XRD, HR TEM, SEM/EDX, EBSD, CL and FTIR microspectroscopy. Only one kind of beads—potassium-sodium-lead glass beads containing large crystals of KSbOSiO4— were found to be subject to more intense destruction than others. The Raman spectroscopy study of turquoise glass beads is presented. Raman spectroscopy is a suitable and fast non-destructive technique for obtaining qualitative information about glass samples. The Raman spectra from the glass regions of various conservation states were obtained in the backscattering geometry using a LabRam HR800 Raman spectrometer (Horiba Jobin-Yvon). The 488 nm line of argon ion laser was used to excite the Raman spectra and simultaneously record Stokes part of laser induced fluorescence up to 1μm. Analysis of the Raman spectra showed that the defective glass areas, regardless of the crystallites formed, are characterized by an increase of non-bridging oxygen bonds. Transitions from Q3 to Q2 units occur at these domains. Also the SEM/EDX analysis showed that the formation of extensive network of cracks is observed and elements are redistributed, including Na and especially K. The investigation of the processes occurring in the intermediate region allowed us to propose a non-destructive method for monitoring the state of the glass beads preservation based on the spectral composition of its luminescence.

The present work is dedicated to a study of several ancient belt buckles decorated with polychromic incrustation. Notable nomads jewelry was found in South Siberia on the bank of the Yenisei River during archaeological excavations in 2015-2017. The artifacts were found in burials which are dated 1 century BC and related to the Xiongnu culture. It’s made of black material and decorated with geometrically arranged blue-green, orange, red and white inserts. Previously various opinions were expressed in the scientific literature considering the material of the buckles base. It was supposed to be dark bone or antler, slate, coal or jet. The multi-analytical study was carried out by OM, FTIR and Raman spectroscopy, XRF, SEM-EDX and XRD methods. SEM reveals the sample’s morphology, the plant fibers and mineral inclusions in the buckle plate. The FTIR results there are several mineral compounds in the buckles plate: kaolinite, dolomite and quartz in addition to the absorbance bands of organic components. Basing on obtained results in accordance with XRD study it could be assumed that the buckle body material is fossil coal. Another aspect of the study was the identification of mineral composition of incrustation materials. Among the blue-green inlays were identified two different minerals: turquoise and chromian muscovite (fuchsite) mineral . While all red inserts belong to the silicate mineral group carnelian, according to the Raman spectroscopy data the red inlay hue depends on the relative content of its components (hematite and moganite). The white imbedded material was identified as aragonite. We suppose that these are the mother of pearl inserts. Visually similar materials turned out to be different minerals. Our findings emphasize the question: what was sacred in the culture of ancient nomads: particular material or special color?

Reflectance Transformation Imaging is a technique that provides a digital and useful representation of the surface of an object (photometric and geometric attributes) created from a sequence of images taken from fixed observation position while varying the position of the light source around the object at each shot. Thanks to a further reconstruction processing, the continuous angular reflectance for each pixel can be computed from the set of discrete acquisitions and rendered interactively. Current mathematical functions that allow reconstruction of the continuous angular reflectance from the discrete acquisitions are PTM, HSH and the DMD. For these three approaches, a uniform spatial distribution of light sources is an implicit hypothesis. Protocols to perform RTI acquisitions are either dome-base or free-form systems. Regardless the acquisition method it is almost impossible to achieve an acquisition under ideal conditions that satisfies the hypothesis of light’s uniform distribution over the hemisphere. This can alter the reconstruction processing and affect its visual quality. In the context of cultural heritage and given the size of objects as the in-situ acquisition conditions, adopted systems are generally free-form, which tends to increase the level of error in the condition of uniformity of lights distribution. To address this issue, we propose a methodology consisting in the estimation of the real spatial distribution of the lighting directions used during acquisition that we use to generate a weight for each light position enabling to correct its contribution on the Least Square Regression. The proposed method is evaluated on two datasets involving big CH objects. Proposed method shows significant improvement of the reconstruction and rendering. Moreover, the calculated difference maps show significant changes confirming the need to consider the spatial non-uniformity of acquisitions by performing a weighted regression based on the proposed directional density method to generate the contributions for the WLS regression.

In our paper we will present ongoing project results connected with monitoring conditions of 12 fragments of the Museum of King Jan III’s Palace at Wilanów (Warsaw, Poland) building façade by 3D structured light scanning method. During 30 months, 7 series of three-dimensional measurements are planned. Each of series assumes the accurate 3D-measurement of the same twelve elevation fragments with dimensions of 120 mm x 120 mm with 2500 points/mm2 resolution, indicated by the conservation and architecture departments as particularly interesting. This article describes the methodology of measurement process, 3D scanner head and other devices developed especially for this action as well as data processing path and analysis algorithms. In addition to having such accurate measurement data, we must still be able to match the measurements carried out in the same place at intervals of several months. For this purpose the measurement fields were marked with special aluminium targets, topped with three intersecting planes. The algorithm of their detection, analysis and use for aligning data from subsequent measurement series is discussed. A portable SLS 3D-measurement head with two cameras, integrated with linear drive has been developed for scanning purposes and adopted to use in outdoor condition. Developed 3D scanner with such high resolution has measurement volume which is limited to 70x50x10 mm for single scan. In less than 25 minutes, 40 measurements are acquired at various positions covering the entire area with support of controlled linear stage stand. Individual scans are pre-aligned with limited accuracy and has to be fitted using Iterative Closest Point algorithm. The final representation of each fragment is a colorful cloud of points with more than 200 million 3D points. We present results of 3D measurements with initial monitoring procedure for assessing changes of 3D surfaces over time.

Microclimate-induced deformations of artifacts are potentially harmful phenomena for cultural heritage collections. In particular, when artifacts are made of hygroscopic materials, variations of temperature, pressure and humidity of the environment can trigger cycles of deformations able to eventually damage irreversibly the work of art. The deformations may be subtle in amplitude but, at the same time, affect a large area of the object. Hence conservators need a tool able to quantitatively evaluate and record such deformations, providing accurate and high-resolution data from large areas. The aim of this work is to investigate the feasibility of using optical micro-profilometry technique as diagnostic tool in microclimate studies. The monitoring of the shape variations on artifacts is carried out in a non-invasive way by means of a scanning system based on a conoscopic holography sensor and surface data analysis procedures that have been adapted for the task. A procedure that uses reference standards, insensitive to thermal and hygrometric variations, has been implemented, with the scope to assess short and long-term drifts of the instrumental set-up. As it results from previous studies, optical microprofilometry offers a number of advantages if applied to the investigation of the state of conservation of works of art: it is a fully non-contact technique, it allows to reach high spatial resolution, it can operate in non-laboratory conditions. In this work, a quantitative measurement of such components was attempted by analyzing the shape response of various hygroscopic materials including a wooden icon, datable to the XVIII century, to controlled variations of the main environmental parameters (temperature, pressure, humidity). Proper control and recording procedures of such parameters were implemented too. This type of investigation could help to keep under control the conservation state of artifacts subject to microclimate variations, eventually leading to stresses and structural failure in the long term.

The château de Selles (dating back to the 14th century) in Cambrai is one of the most important monuments in France for medieval graffiti. Today the some parts of the engraved walls are endangered by a complicated weathering process. The present project lead by archaeologists in collaboration with universities (history, archives) and companies which are specialized in 3D surface scanning and 3D modelling, and funded by the Ministry of Culture of France since 2011 aims to apply archaeological methods and new optic technologies in order to preserve the endangered engravings and to study the graffiti. For documentation, color photography (fontal and side light) is used for the survey, archive and deciphering by tracings of the engravings. Other photographic technologies as RTI and photogrammetry have been tested to get the fine engravings more legible. 3D high resolution (micronscale) scanning of the engraved parts of the walls have been used to get a digital print of most of the traces on the wall in order to be able to work on the 3D models. The 3D models of the engraved are used by archaeologists and paleographers to explore precisely the engravings and decipher objectively the motif and texts. They offer a complete new view on these witnesses of history. The repeated 3D scanning of some parts of the castle (especially the most endangered) have led to a 6 year monitoring of the weathering processes. A complete photogrammetric survey was made on the outside and inside of the castle in order to build up a virtual visit to the castle thanks to a virtual reality headset that helps the visitor to discover the engravings in an immersive experience. The photogrammetric data produced for the virtual reality are also used to build a dynamic database including spatial information. The aim is to make the data available and offer the possibility to make the databases evolve in time.

Precise and robust 3D model reconstruction is required in various outdoor scenarios such as aircraft inspection, rapid prototyping, and documentation of an archaeological site. While laser scanning provides a robust solution for online 3D scene reconstruction, the resulting model may be incomplete if the scene is cluttered. Structure-from-Motion (SfM) technique provides flexibility concerning the scene structure. However, its accuracy may be limited especially for textureless surfaces. This paper is focused on the development of a mobile structured light 3D scanner for online reconstruction. Our approach combines photogrammetric methods and deep learning. We use a structured light projection for fast data acquisition. We use deep learning based pattern matching to improve accuracy up to 0.1 pixels. Our FringeMatchNet is based on the U-Net architecture. The network produces an estimated shift heatmap with subpixel accuracy. Each pixel (x,y) of the heatmap represents the probability that the shift between the two image patches is equal to (x,y). We generated a large dataset that combines synthetic and real image patches to train our FringeMatchNet. We use a sequence of light codes to find the corresponding patches at the coarse scale. We evaluate our scanner in different scenarios such as the inspection of an aircraft wing, rapid prototyping, and online reconstruction of an archaeological site. We use ground truth 3D models and etalon objects to estimate the accuracy of our scanner. We compare the accuracy of our FringeMatchNet with other stereo matching algorithms both hand-crafted (SGM, LSM) and modern deep learning-based. The evaluation proves that our network outperforms hand-crafted methods and competes with modern state-of-the-art deep learning based algorithms. Our scanner provides the measuring volume of 400×300×200 mm for an object distance of 600–700 mm. It combines portability with an object point resolution between 0.2 and 0.1 mm. Online evaluation of our mobile scanner in the outdoor environment proves that it is robust against complex scene configuration, low textured objects, and semigloss surfaces. Deep learning-based stereo matching using our FringeMatchNet facilitates subpixel registration and allows our scanner to achieve sub-millimeter accuracy in the object space.

A large amount of iron and steel artifacts produced in Europe between the 2nd and 14th centuries AD were constituted from pattern-welded composites, phosphoric-iron, and steel components. Phosphoric iron is a substitutional alloy, which is obtained by using iron ore from swamps in which decomposition of dead organisms enriches the iron rich soil with phosphorous, and this is a typical condition of large plains of Central Europe. This alloy provides no particular improvements in mechanical performances of iron, but on the surface of object it can be optically distinctly distinguished from non-phosphor components. Thus, identification of phosphoric iron in archaeological iron objects is important for determining the manner of their production and technological characteristics [1]. At present, successful identification of phosphoric iron can be invasively achieved by metallography (use of specific etchants such as Oberhoffer reagent and measuring hardness of the metal matrix). Another technique commonly used to characterize the bulk structure of metal ancient artifacts is neutron tomography, using cold neutrons. It provides image-data enhancing differences between the phosphorous rich iron areas and the standard iron and steel ones into the object. Neutron imaging is an effective and well-established method but it is complex to implement because of the high costs and the limited accessibility. Therefore, the exploration of new techniques capable of providing additional data on the nature of alloys would be highly needed. In this work, a pilot study addressed to test the applicability of reflectance hyperspectral imaging to the investigation of ancient metallographic artifacts is presented. So far this technique has been successfully used for the non-invasive diagnostics of polychrome surfaces in several types of artworks [2], but has never been applied to investigate corroded metallic surfaces. Hyperspectral imaging measurements in the VIS-NIR range (400-1700 nm) were performed on the surface of a pattern welded phosphoric iron artifact, made on purpose as a replica of an historical object from the archaeological site of Kobilic (Croatia) [3]. The same artifact was analyzed also using neutron tomography, so that the phosphorous rich areas were univocally identified throughout the entire object by using the latter method. The comparison of the reflectance hyperspectral data with the near surface neutron tomographic slice provided prominent similarities. These preliminary results encourage further investigations on merging these two imaging techniques belonging to diverse applicative areas. References [1] Stewart, J.W., Charles, J.A., Wallach, E.R. "Iron–phosphorus–carbon system: Part 3 – Metallography of low carbon iron–phosphorus alloys." Materials Science and Technology 16(3), 291-303 (2000). [2] Cucci, C., Delaney, J. K., Picollo, M. "Reflectance hyperspectral imaging for investigation of works of art: old master paintings and illuminated manuscripts." Accounts of chemical research 49(10), 2070-2079 (2016). [3] Thiele, A. "The manufacturing technology of a pattern-welded knife from Kobilić (Republic of Croatia)." Archeologické Rozhledy LXX, 457-467 (2018).

Terahertz time-domain imaging (THz-TDI) and Spectral domain optical coherence tomography (SD-OCT) are two investigation methods capable of providing 3D datasets from which depth profiles, cross-sectional images, plan-type images and volume rendering of an object can be derived. These novel photonic technologies are extremely relevant to the field of heritage science, for which the comprehension of the stratigraphic structure of a cultural heritage object may help in the understanding of its manufacturing technology and state of preservation. Different imaging depths, fields of view and axial/lateral resolutions characterize the two imaging technologies, which thus provide different but complementary information of the same scanned object. In this study we show how the THz-TDI and SD-OCT scans performed on a fragment from the underside of an Egyptian coffin (Theban area, 21st/22nd Dynasties, Egyptian antiquities department of the Louvre museum – Paris, France) gave precious insights on its full structure and stratigraphy. The artwork consists of a wooden panel covered by a preparation layer on which typical ancient Egyptian pigments have been applied to depict Egyptian otherworldly figures and scenes. The varnish and most of surface paint layers have been efficiently detected and imaged by means of OCT, while the preparation layer and the support have been imaged by means of THz-TDI. A complete view of the depth profile of the investigated archaeological remains has, therefore, been made possible through the acquisition by the two different recording techniques followed by careful analysis and comparison of the results. In addition, the THz-TDI and SD-OCT cross-sectional images have been presented in a unique image grid and space by processing the acquired signals so that the different datasets are dimensionally consistent. This is possible through the multimodal image fusion method that we recently developed for merging the outputs provided by the THz and OCT analytical systems available in our laboratory.

The Integrated Platform for the European Research Infrastructure ON Cultural Heritage (IPERION-CH) is a consortium of 24 partners that aims at establishing a unique European research infrastructure for restoration and conservation of CH. Within the framework of IPERION-CH, French institutions, in collaboration with other European institutions, are involved in the development of a Laser-Induced Breakdown Spectroscopy-Laser-Induced Fluorescence-Raman Spectroscopy (LIBS-LIF-Raman) portable instrument. Such a multi-analytical prototype instrument is able to combine these three laser-based spectroscopic techniques to simultaneously provide complementary elemental and molecular analytical information from the same analysis point. To that purpose, laser sources, appropriate optics and detection modules have to be integrated on a mobile platform capable of supporting conservation campaigns when extensive measurements and on-site decision-making are needed. LIF and Raman are generally implemented in continuous mode, although the pulsed laser mode is also recommended in order to improve the ratio between Raman and fluorescence background signals ratio, and to open up the possibility to work in ambient light. Therefore, operating in pulsed mode provides the opportunity to use only one laser source to carry out the analyses with the three spectroscopic techniques. This work focuses on the system design, the development of measurement protocols and on the data treatment and interpretation. The wavelength of 355 nm has been chosen as the excitation source after testing the different wavelengths available of the harmonic outputs of the nanosecond Nd:YAG laser. Signal collection is set as a 4-f imaging system with a lens of high numerical aperture. Due to the different spectral range, spectral resolution and detection time windows required to perform LIBS, LIF and Raman spectroscopies, the configuration of spectrometers should be considered respectively. The next step in the instrument design aimed at minimizing the weight and dimensions of the prototype. Results obtained during the development of the prototype will be presented and discussed.

Taj Mahal, the 17th century architectural wonder is known worldwide for its visual appeal in white marble with beautiful inlay works in the style of ‘Pietra Dura’ (pietra means stone and dura means hard) from Florence, Italy. It has recently come under scrutiny owing to the long-drawn restoration process for structural damage and surface discoloration. The conventional means of structural survey in heritage architectures, like the Taj Mahal, is based mostly on visual assessments by experts which is highly time-consuming and subjective; and in most cases can be done only after the damages become apparent. Consequently, the restoration involves exhaustive repair procedures. In this work, we have introduced a combination of optical characterization techniques, namely, µ-Raman Spectroscopy comprising variable power 532 nm CW laser (Laser Quantum UK) and single-grating Spectrometer (Andor UK); and a Terahertz Time Domain Spectroscopic Imaging (THz-TDS imaging system from Toptica Photonics) covering the range of 0 – 6 THz, as an alternative approach for both surface and sub-surface damage detection in Marble structures. In our work, we have simulated some of the most common structural and surface damages in heritage Marble architectures found worldwide to some selected representative models. These model slabs bearing similar ornamental colourful motifs like the Taj Mahal were probed by the above two techniques, which showed tremendous capability to sense very fine cracks which are visually impossible to detect. Also, minute changes in surface quality with the introduction of organic materials were mapped with high accuracy. This paves the way for an efficient and trusted survey platform that could be deployed in future which would be able to provide a fast and reliable structural assessment of our cultural heritage architectures for the next generation.