We began using RTI during excavations at Antiochia ad Cragum, in Turkey. This summer we will image the Jeffers Petroglyph Site in Minnesota. How can we capture the most petroglyphs this summer? Our shooting scenario: capturing several thousand carvings deposited on exposed bedrock, in full sun, with changing lighting conditions.
A case study on a project using RTI to reveal the texture of watermarks on watercolors with a backing that prevented traditional transmitted light photography.
This project combines data obtained from RTI, multi-band imaging, and photogrammetry into a single interactive file using Blender, an open-source 3D graphics and animation software. The part of the project presented here explores the value of combining RTI and photogrammetry. RTI, multiband imaging, and photogrammetry complement each other and together can yield high-resolution renderings that emphasize surface topography, reflectance and absorbance properties, and physical shape. Their combined benefits can facilitate more detailed study of an artifact remotely and can minimize handling of ancient material through "virtual handling."
This early 7th-century dry-lacquer Buddha (19.186) from China is hollow and made of fabric and dry lacquer. It is one of just three known examples that have survived from this period. Our conservation department needed to build a mount that supports the Buddha from the inside during display and transportation. To create a 3D model, our imaging team combined photogrammetry (using an iPhone and a Hasselblad camera) as well as laser scanning.
The Museum of Fine Arts (MFA) in Boston recently tested the potential and versatility of a CrimeScope CS-16 500 unit from Horiba Scientific for conservation purposes. The CrimeScope was designed for forensic applications and has started to be used in museums. Different MFA conservation labs (i.e., paper, furniture, and objects) analyzed this tunable radiation source to examine and document artwork under lab and gallery conditions. This specific instrument can emit nineteen (19) distinct bands of light, ranging from 300 nm to 830 nm, that can be helpful for distinguishing different materials. The MFA compared the efficacy of the CrimeScope with different ultraviolet and infrared emission sources currently owned and regularly employed for imaging purposes at the museum. The CrimeScope produced satisfactory results, but individual dedicated radiation sources did generate slightly crisper images. The instrument did expand the current multi-band imaging capacities by allowing conservators to examine the visible-induced visible luminescence (VIVL) of artworks. Overall, conservators found the CrimeScope easy to use and in particular found the emission intensity and large stop size very useful; however, there were persistent reservations about the cost of the instrument.
Latest developments in the coordinated attempts between conservators and photographers in The Metropolitan Museum of Art to produce workflows for the capture of broad-spectrum images. Inserted in the broader context of standardization and repeatability, the talk will more specifically illustrate a workflow for the use of targets in Adobe Photoshop Lightroom.
The Arc/k Project would like to present some of the photogrammetric challenges encountered from using crowd-sourced data for digital cultural preservation.
Qeros are Andean ceremonial drinking vessels decorated with a complex and beautiful polychrome technique: low-relief carving filled with sheets of a plant resin pigmented with a range of organic and inorganic colorants. RTI reveals the technology of manufacture and the remarkable three-dimensional surface qualities of these vessels. Imaging techniques such as RTI, Reflected IR, and photogrammetry are powerful tools for non-destructive analysis and dissemination.
Reflectance Transformation Imaging (RTI) was directly enabled by the introduction of digital vs. film photography. We take it for granted now, but every photograph taken with a digital camera on a tripod is perfectly registered to the one before it, as long as the subject or camera didn't move. This was not true when using film, and digital photography made capturing quantitative models of how an object responds to varying lighting (called the reflectance function) trivial. We briefly describe the development of RTI in the late 90's and follow up with a look at three offshoots of RTI that may prove useful in the future when applied to conservation and cultural heritage. First, it is possible to find "informative" lighting directions automatically from an RTI using a concept called image entropy. Second, we show it is possible to make "prints" from an RTI that respond to lighting in the same ways that the real objects originally photographed to make an RTI do. Third, we show off a system that performs RTI capture and display at video rates, allowing the user to simply hold an object up to the camera and see faint details often not visible to the human eye.
This presentation reports on tests in order to integrate multispectral imaging and Reflectance Transformation (RTI) through a pilot on Roman mummy portraits held in the collections of the UCL Petrie Museum of Egyptian Archaeology. The aim was to assess the extent to which the combination of RTI and spectral imaging provides potential for future research on portrait-manufacturing techniques and the biographies of these objects, as well as providing a cost-effective tool for planning conservation work and subsequent artefact monitoring. The integrated spectral RTI capture utilized a method developed by Todd Hanneken's (St. Mary's University) together with processing software, Spectral RTI Toolkit (see Github), the continuing development of which is supported by the National Endowment for the Humanities. The pilot was conducted using the UCL Centre for Digital Humanities Multi-Modal Digitisation Suite's spectral imaging system comprised of an integrated PhaseOne 60MP IQ260 digital back and iXR body with 120mm lens, 6-position motorised filter wheel, and narrowband LEDs at 12 different wavelengths covering the spectral range of 370nm to 950nm. Rotolight NEO™ LEDs and a Canon Speedlite 600EX-RT were trialed and compared for the RTI portion of the capture sequence. The results took the form of Accurate Colour, Extended Spectrum, & PCA Pseudocolour RTI files. These super-high-resolution images delivered deep texture and extended color information with robust documentation of the portraits' complex surfaces: e.g., paint strokes and sequence, chips of missing paint, craquelure, fine cracks in underlying wood, accretions such as textile. The pros and cons of the technique are briefly assessed and follow on questions raised.
Photometric stereo techniques can recover useful shape information from reflectance transformation imaging (RTI) data despite systematic errors inherent in such approaches. For example, flat objects with inscribed details can be reconstructed using an optimization technique that constrains the solution to match expected lighting falloff on flat surfaces. We used this approach to provide conservators at the Art Institute of Chicago additional materials for the analysis of printmaking processes employed by Paul Gauguin. The addition of a second camera attached to the movable illumination source can additionally improve the accuracy of shape recovery during an otherwise conventional RTI capture process. The fusion of multiple views from this second camera with the fixed RTI camera exposures can produce geometry estimates that outperform independent photometric and multi-view reconstructions. Extending RTI captures in this way covers a middle ground between conservation tasks that rely solely on relighting and those that rely on more resource-intensive, high-fidelity 3D acquisition systems.
Lithic artifacts (knapped stone tools) are difficult to image digitally due to low-relief, low-feature contrast, non-uniformity, and the specular nature of many of the materials used (e.g., obsidian, chalcedony). Line drawings are often used instead of photographs, but time and cost considerations limit their use. RTI "normals" images are an excellent, inexpensive alternative, providing detail comparable or superior to line drawings at a fraction of their cost. Methods for processing normals images for optimal detail are discussed, including complications. Preliminary results on groundstone tools are presented, including the use of normals data to visualize surface smoothness.