Essay: Developing spectral imaging: A standardized tool for collaborative conservation support

Multispectral imaging offers potential to support both conservation and preservation studies of manuscripts, books and other two-dimensional objects.  This includes images that offer insight into the unseen, as well as data for scientific studies and comparisons against standard samples and other manuscripts.  During the past 18 years, narrowband multispectral imaging systems have advanced with the latest equipment, software and work processes needed to provide the conservation community with a standardized research tool. Advanced spectral imaging equipment, techniques and work processes have become standardized tools to support conservation and preservation studies of manuscripts and objects.  This allows studies of both the overall object and specific key areas of interest by both preservation specialists and scholars.  The ability to transfer the resulting digital data to preservation and conservation specialists – as well as curators and scholars – allows further evaluation and collaborative study of the object.  Digitally processing and combining the resulting image set can reveal additional important features on the objects that are not visible to the eye in natural light.
Over almost two decades, cross-disciplinary teams have developed narrowband multispectral imaging system technologies as an important tool for conservation.  Narrowband multispectral imaging uses narrow spectral bands of light to illuminate an object, instead of putting broad bands of light on the object and filtering out the unwanted bands.  While different terms tend to be used interchangibly – “spectral,” “multispectral,” hyperspectral” – in practice multispectral imaging refers to images taken with three to twenty wavelengths of light with a good spatial field of view; hyperspectral imaging refers to images taken with large numbers of very narrow wavelengths of light (sometimes hundreds) and tends to have a smaller spatial field of view; and spectral imaging refers to any imaging that takes an image using any number of wavelengths of light (and can be used for either multi- or hyperspectral). In narrowband multispectral imaging, a medium-format monochrome camera takes a series of high-quality digital images, each illuminated by a specific narrow wavelength of light from low-heat LED light sources. Following image collection, processing of spectral images provides additional data on the materials, colorants and other object components. This could include data on the spectral performance of encasement materials, especially glass and quartz, to minimize future deterioration.
Narrowband multispectral imaging systems benefit from not only the latest generation of equipment and software, but also work processes developed in collaboration with the conservation and information management communities to provide a standardized research tool. Effective spectral imaging requires not just collection of quality images, but the ability to manage and exploit large amounts of integrated data and metadata.  Data management, operation, training, information storage and access are required for the collaborative analysis of the images and image products from the spectral imaging system.  Data collected in standard formats can be made available for access and sharing for further analysis.  With common standards and techniques, this can include collaboration with other institutions’ spectral imaging studies and data. Integration is also possible with other standard digital images or data collected with other instruments, including X-ray fluorescence (XRF) and microscope equipped Fourier transform-infrared (FT-IR) spectrometers.
Ultimately, multispectral imaging technologies and techniques must meet the needs of a range of users in the conservation community.  The captured data must be able to stand up to any review and scrutiny.  This will includes not the on-site multispectral imaging, but also advanced image processing of the images to provide better visualization of spectral responses.  The product needs to include standardized multispectral data with the data structure and management to support digital preservation of and access to all the spectral images on institutional information systems.  With effective digital preservation, the data will support potential follow-up scientific and/or imaging studies with additional technologies, as with the mummy mask imaging project.
The product from multispectral imaging systems has the potential to support both scholarship and conservation by providing more insights into the history and construction techniques of manuscripts.   In addition to digital images, the spectral data can also be used to scientifically analyze the ink and parchment, potentially in comparison with other inks and parchment samples.  This could be of particular relevance in comparing manuscript leaves from one collection with similar manuscripts or “missing” leaves from the manuscript. a leaf of the Syriac Galen Palimpsest in 2015).  Technology has had a major impact on cultural heritage studies of manuscripts and artifacts over the past few decades.  From the dawn of writing through the 20th Century, researchers needed physical access to the object to conduct their studies.  With the advent of digital imaging and data sharing technology, researchers can now digitally access manuscripts and objects without regard for geographic distance.  Advanced imaging systems empower librarians, museum curators, scientists and preservation specialists to glean important new information about their collection objects for evaluation and study. These systems yield not just quality images, but also provide institutions with integrated image data and metadata needed to offer new insights into manuscripts and printed works.  In addition, current multispectral and other imaging systems now provide researchers with access to what was previously unseen: features and artifacts that are not visible on the actual object with the naked eye.
Digital imaging capabilities developed from forensics, earth resource, and national security applications are now becoming effective tools for multidisciplinary studies in museums and libraries.  Their use poses challenges for scholars, scientists and institutions working to digitize and integrate large amounts of data from multiple sources for study, access and permanent storage.  To properly apply these tools, advances in digital imaging technology must capitalize on parallel developments in information technology to ensure the continued utility of standardized cultural heritage data for collaborative study across disciplines by researchers in multiple institutions.  This includes the integration of not just technologies, but also work processes and skills into an efficient system appropriate to the technology, institution and culture.  This systems approach in support of a diverse set of researchers is changing the face of museum and library studies and analyses.
Integration of spectral imaging systems into libraries, museums and other cultural heritage institutions supports ongoing studies of modern and ancient manuscripts.  These range from studies of early papyrus texts to the ancient palimpsests of Archimedes, Galen and those in St. Catherine’s Monastery in the Sinai and the Vatican Library, to the more modern Top Treasures in the Library of Congress and David Livingstone’s Diaries in Scotland.  The free online availability of image products from these projects advances scholarly and conservation studies by allowing broader global analysis, research and collaboration.  The integration of spectral imaging and processing, work processes and global access technologies into useful end-to-end technical systems provides standardized metadata and data about these and other antiquities for study by researchers across cultures and institutions.
Digital imaging of historic artifacts poses significant challenges with the large and potentially complex data files that serve as facsimiles of the physical object.  Linking advanced technologies with online tools and talented personnel offers researchers a rich reward of data for antiquities research.  With integrated systems, institutions can help ensure data from these advanced tools can be efficiently shared globally with commonly available software, infrastructure, and storage systems for collaboration across institutions and disciplines.
Multispectral imaging offers new insights into historic objects, including those at the Gilcrease Museum. Based on techniques and technologies developed during imaging work in a range of institutions – from the Walters Art Museum to the Library of St. Catherine’s Monastery in the Sinai and University College London – multispectral imaging studies have revealed obscured text, drawings and production details.  With this new tool for non-destructive analysis, museums are now able to effectively integrate multispectral imaging with current workflows to support interdisciplinary research.  At the Gilcrease Museum Helmerich Center for American Research, multispectral imaging carried out with the support of the digital curation team provided new insights into objects from the museum collection, including the 1775 Paul Revere Certification and a copy of the Declaration of Independence.  In each, spectral imaging and processing provided new insights into the provenance and history of these manuscripts (Fig 1).
Advanced multispectral imaging systems empower curators, scholars, preservation specialists and scientists to obtain important new information about their cultural heritage objects for evaluation and study. The latest generation spectral imaging equipment yields not just quality images, but provides institutions with the ability to collect, process, manage and exploit the large amounts of standardized multispectral image data and metadata that meet their goals.  Current systems are based on techniques and technologies developed during over a decade of imaging to support institutions and scholars around the globe – from the Harvard University Library to the Library of St. Catherine’s Monastery in the Sinai and University College London.  Multispectral imaging has been used to support scholarly and digitization studies of a variety of materials in institutions around the globe, including Top Treasures at the Library of Congress – the Waldseemüller 1507 Map, and drafts of the Gettysburg Address and Declaration of Independence; hide paintings and early photographs at the New Mexico Museums; Scandinavian maps in Sweden and Denmark; damaged and burnt manuscripts in Germany; David Livingstone’s Diaries; palimpsests in the Vatican Library and St. Catherine’s Monastery in the Sinai; papyrus samples in the Duke University and John Rylands Libraries and the Berkeley University Tebtunis Collection; artwork at the Getty Museum and others.  For these and other paper, parchment and papyrus objects, narrowband multispectral imaging systems have produced standardized image products that has allowed further analysis and research to advance scholarship and conservation.
With standardized spectral imaging systems that include image processing and data management, now libraries, museums and universities are capitalizing on these technologies to conduct their own multispectral imaging of important objects – revealing that which had not been seen before.  Multispectral imaging studies at these and other libraries and museums have demonstrated that spectral imaging equipment, techniques and work processes can effectively support research into a range of cultural heritage materials – from artifacts to manuscripts.
Since narrowband spectral imaging was used in the Archimedes Palimpsest project over 15 years ago, technical experts and scientists have refined and adapted multispectral imaging capabilities to meet cultural heritage needs for institutions with valuable collections.  These include not just the imaging system, but also work processes, training and image processing capabilities required for an effective, full-life-cycle multispectral imaging program.
For over a decade, multispectral imaging systems have provided not just quality images, but also the ability to process, manage and exploit standardized data and metadata to meet near- and long-term digitization and scholarly goals.  This has been used in studies of text, drawings and maps on paper and parchment; papyrus texts; books; photographs and paintings; and artifacts including pottery, decorative items and textiles.
Spectral Imaging System
The latest spectral imaging systems include commercial-off-the-shelf (COTS) hardware and software for digital spectral image capture and viewing with an integrated system.  Systems include an achromatic camera, object positioning devices and mounting accessories used in many institutions.  Multispectral imaging is conducted in a dark room with narrowband illumination provided by low heat and low maintenance, long-lifetime light emitting diodes (LEDs) that provide illumination for imaging in multiple distinct ultraviolet, visible and infrared spectral bands.  The object is illuminated in about a dozen specific wavelengths in the spectral range of 370 nanometers (nm) to 950nm – from the ultraviolet (UV) through the visible and into the near infrared (IR) with two integrated illuminators, each with multiple LEDs.  In addition, to capture fluorescence from an object, a motorized filter wheel containing optical glass filters increases the range of captured information to include fluorescence emissions. This allows the characteristic spectra of substrate, colorant, and contaminant materials (e.g. parchment, iron gall inks, and mold, respectively) to be more visible for analysis.   Operating software allows simplified system operation and unified metadata capture with integrated control of the digital camera back, filters and illumination as a single system.   Fully integrated systems also include customized image processing software to combine and highlight features from the multiple images in the high spectral and spatial resolution image sets.
Two case studies highlight the potential offered to cultural heritage research, preservation and access by advanced multispectral imaging: the Syriac Galen Palimpsest and Mummy Mask Cartonnage.
Syriac Galen Palimpsest
In 2002, the owner of the Archimedes Palimpsest purchased another palimpsest for study: a Syriac translation of the Greek physician Galen of Pergamon that had been overwritten with a liturgical hymn book. Galen’s influential text On the Mixtures and Powers of Simple Drugs was translated from the original Greek to Syriac and Arabic as it influenced the development of early medical practices across cultures. In 2009 a team used multispectral imaging in an initial study at the Walters Art Museum to try to reveal the undertext by Galen.  This study yielded positive results, allowing multispectral imaging of the entire disbound manuscript in 2010.  Working from this initial data, scholars discovered that a 6th-century translation of Galen’s On Simple Drugs was scraped off and overwritten with an 11th-century Syriac religious text. Following the data and metadata management and hosting techniques developed for the Archimedes Palimpsest, the team hosted all the data on digitalgalen.net.  These included all the captured narrowband multispectral images and a set of pseudocolor and greyscale images processed from combinations of the captured images.
Scholars used the online Galen images to study the Syriac translations of Galen’s teachings and compare them with other translations and expand their knowledge of the use of “herbals” in ancient medicine.  They also used the images in conjunction with catalogues of other Syriac texts to track down detached leaves from the original manuscript in libraries in the Vatican, the Sinai Desert, Harvard, and Paris.  Multispectral imaging revealed the undertext in these rare manuscript leaves from 2013 to 2015, and they have all now been reunited digitally and are freely available for global study at digitalgalen.net. The scholarly study of this text is covered in the New York Times article Medicine’s Hidden Roots in an Ancient Manuscript
Mummy Mask Cartonnage
Egyptian mummy masks and cartonnage were often fabricated using a papier-mâché process with layers of papyrus, including some with texts of historical interest on them. Texts written with carbon-based inks were common in the Ptolemaic and early Roman period (30 BC-641 AD). These inks were mainly composed from soot, gum arabic, and water. In the past, researchers have destroyed the masks to access the papyri, denying future researchers access to the primary historical artefacts.  An international team developed a multidisciplinary project to assess the feasibility of integrating non-destructive digital imaging technologies for making texts visible in images of papyrus in mummy mask cartonnages for open research and analysis.  Given the difficulty of this task, the team also developed intermediate goals, such as simply detecting the presence of text on mummy masks.  This was an initial attempt to offer scientifically valid approaches for documenting the initial state of objects and their production for future research and analysis.
Imaging of mummy cartonnage requires penetration of the paint and papyrus layers, as well as binders and other materials from which the artefact was constructed.   In this phased approach, the team worked to develop increasingly complex techniques, starting with optical multispectral imaging of actual mummy cartonnage and fragments, as well as surrogate test objects that do not require difficult conservation care, handling and transportation. Complementary to the optical imaging, this was followed with optical coherence tomography imaging and imaging with the higher energy levels of x-rays.
With advanced imaging technologies the various teams assessed, identified and attempted to read the texts contained within mummy cartonnage, minimizing damage to such artefacts.  This was a significant scientific challenge, but pioneered new methodologies for non-destructive imaging and analysis, sharing of standardized data for interdisciplinary studies, and focus attention the need to avoid the current destruction of these important cultural heritage objects for textual analysis.  This project built on the extensive experience of teams that have pioneered imaging with multispectral imaging systems, Synchrotron Rapid-Scanning X-ray Fluorescence (SRS-XRF), x-ray micro Computer Tomography (micro CT), and Optical Coherence Tomography (OCT), as well as data management and integration, and program management. Team members come from the fields of digital humanities, medicine, particle physics, imaging science, data management and systems integration.  SRS-XRF was conducted at the SLAC Stanford Synchrotron Radiation Lightsource (SSRL), micro CT imaging at the Berkeley Advanced Light Source (ALS), and Optical Coherence Tomography imaging at Duke University.