Multispectral Imaging (MSI) is commonly used in cultural heritage for the recovery of lost and damaged writing. Through MSI, illegible text can be recovered (see research question 4) without destroying, damaging or otherwise changing the original artefact. The non-invasive and non-destructive process of MSI consists of collecting spectral information through imaging and later revealing the lost writing through image processing. Registered, high-resolution images are collected during the imaging sequence from a single manuscript page or object. Depending on the chemical composition of the materials, such as inks and writing supports, a difference in reflectance, fluorescence or transmittance at single wavelengths can be detected for each pixel. In statistical image processing, lost writing and other useful features are revealed, dependent on the state of degradation, by digitally enhancing low signal and/or low contrast information. A set of final images with the lost writing, made as legible as possible for the scholars, is produced, shared and discussed.

The CSMC MSI system is using 19 discrete LEDs as light sources to illuminate the object ranging from the ultraviolet (365 nm) to the near infrared (1050 nm). This approach is not only much gentler on the object than using a set of hot halogen lamps, which project all the light on the object at once and then need filters to separate the wavelengths, but it also enables us to generate extremely accurate color renderings of the original. Our system has 2 ultraviolet wavelengths (365 nm and 385 nm) that are limited to a maximum of two minutes exposure time during a single imaging sequence, which further eliminates potential risks. To secure and fix the object during imaging sequence, such as pages of the codex or single fragments, we use conservator-approved materials. Individual handling solutions are agreed upon with conservators prior to every project. When all illumination sources are used, a total of 49 images per item (e.g., leaf, page, side, etc.) are accurately captured. These greyscale images are flattened, calibrated and compiled into a three-dimensional “image cube” that serves as the basis for image processing whereby computations are performed to enhance or reduce the contrast of specific pixels, making the desired text more legible (see Digital Images). See MSI data. 

Multispectral Imaging (MSI) in cultural heritage is commonly used for recovering lost, damaged or otherwise illegible writing (see research question 4) without destroying, damaging or altering the original artefact. The process of MSI consists of collecting spectral information through imaging and later revealing the lost writing through image processing. During imaging, the object is photographed under 19 distinct wavelengths of light that range from the ultraviolet (UV) to the infrared (IR) regions (365–1050 nm). These wavelengths are employed in four modalities – reflectance, fluorescence, raking, and transmissive – and result in up to 50 images per capture sequence. These registered, high-resolution, greyscale images are combined into a three-dimensional ‘image cube’ to which mathematical processes are then performed in order to enhance or reduce the contrast of specific pixels, usually with the goal of making the desired text more visible (see Digital Images). A final set of “processed” images with the lost writing made legible is produced without damaging the object.