Intriguing new digital imaging technology is unveiling previously illegible and even unseen text on ancient manuscripts recovered from caves near the Dead Sea, the ashes of Vesuvius, and the ruins of Petra. Known as multispectral imaging (MSI), the technology has outperformed conventional photographic techniques in many applications, capturing images that are shedding new light on history. This initial success at producing superior textual images of fragile, deteriorating antiquities--and electronically preserving those images for future study and appreciation--has attracted widespread interest and bodes well for the continued use of MSI technology in archaeological fieldwork.
This article outlines the development of MSI, describes its advantages over both black-and-white and infrared photography, and reviews the exciting results of its recent applications to fragments of the Dead Sea Scrolls and to the carbonised scrolls of Petra and Herculaneum.
Development of multispectral imaging
Space exploration has been a rich source of spin-off technologies for terrestrial applications as diverse as robotics, biomedicine, and material science. So it may come as no surprise that MSI, though inspired by techniques developed and first used for remote exploration of the solar system, has begun to significantly affect the study of the past and its artefacts.
Multispectral imaging was first applied in the LANDSAT series of satellites, whose imaging hardware used four or five spectral bands to capture images in either the visible or near-infrared ranges of the light spectrum (Elachi 1987: 104-10). Most of that instrumentation was designed for use on moving platforms such as spacecraft, satellites, and aircraft. In the early 1990s, a shift in NASA's planetary missions from large science payloads (thousands of kilograms) to small ones (up to 100 kilograms) required the development of smaller instruments that retained the same science capability as before.
One result was that new imaging spectrometers (sophisticated multispectral imagers) were developed that did not require motion for image acquisition. This advance allowed MSI technology to be used in other areas; for example, multispectral image detectors can now be mounted on a microscope for biological study or on a fundus camera for medical examination or, in our case, taken into the field or a museum for archaeological imaging.
In contrast to remote sensing from space at hundreds of kilometres, archaeological MSI is typically performed at close range (measured in tens and hundreds of centimetres) in makeshift laboratories in settings that vary from humid jungles and caves to museums and conservatories in faraway lands.
Advantages over conventional photography
By responding to ultraviolet and infrared light, the image detectors in multispectral cameras reveal information that is concealed from the human eye. Multispectral imaging (1) (see technical note at the end) divides that portion of the light spectrum into a number of frequency bands and records each of the images separately, typically as a set of monochrome images. Multiple images of the same scene, each viewed at different wavelengths, form a multispectral image cube. This data cube can then be processed to extract information related to spectral differences of the images within the cube (see Figure 1).
[FIGURE 1 OMITTED]
By contrast, colour pictures use a very reduced set of frequency bands operating over the spectrum of visible light to produce an image from cyan, magenta, and yellow image planes as shown in Figure 2. Multispectral imaging divides the same spectrum into multiple finer-image data sets and adds images from outside the visible light range. As a result, multispectral imaging produces a set of images with much more information than a single black-and-white or colour photograph. Examining these additional images, one can often see faint details emerge from the background at spectral locations where clutter disappears, ink becomes dark and the background light, or pigments appear. The quality of the response of a single point in an image (a picture cell, or pixel) is a function of the illuminating light frequency. Digitally recording of the image cubes as a collection of pixels provides data sets for comparison and later …