Abstract. We have recovered undertext from a palimpsest using multispectral imaging. Moreover, we have developed a method for generating simulated multispectral images from previously acquired digitised images of the manuscript, using knowledge of how a colourchecker chart appears in the multispectral images and in the standard digitised images. The ability to identify the undertext was generally better in the real multispectral images, though there were examples of improved identification in the simulated images. However, the method was unsuccessful when applied to freely available images of the Archimedes Palimpsest, suggesting that the mapping between multispectral and standard images must be calculated using the same camera systems that were used to acquire the original images. This approach may allow deeper analysis of previously acquired digitised images using this low-cost simulated multispectral image approach.
