Hyperspectral Imaging (HSI) is considered a capable tool for the analysis of faded or deliberately erased historic documents. In HSI, the spectral reflectance properties of the documents are registered with high precision across the visible light (VIS) and near infrared (NIR) regions of the electromagnetic spectrum (EMS). The resulting data set contains the photographic (spatial) information combined with the spectral information for each pixel. Varying reflectance intensities of different materials can allow for an increased contrast of paper/parchment and ink. By using supervised and unsupervised multivariate image processing techniques, the visibility of the text can be enhanced or restored making use of the data’s statistics. Most of the techniques used are derived from the field of remote sensing. To validate the effectiveness of several techniques proposed in the literature, their exemplary application to different historic manuscripts is presented in this work. To enhance the spectral and spatial quality of the scans, pre-processing steps like noise reduction and spectral sharpening are employed. Then, unsupervised transformation methods (principal component analysis – PCA, independent component analysis – ICA and minimum noise fraction transformation – MNF) are applied to different analytical tasks and their potential usefulness in legibility enhancement is confirmed. Supervised classification methods (spectral angle mapping – SAM and mahalanobis distance classification – MahDC) showed to deliver high quality classification images, if the scriptures are well enough separable. In this work, the high versatility of HSI for legibility enhancement of historic manuscripts was confirmed. The technique provides a toolset for various issues and can be adapted to the diverse objects that are expected in this field of application, like the separation of different, overlaying scriptures or the attenuation of disturbing background effects. In the process of digitization, cultural institutions like museums and libraries are increasingly focusing on the online presentation of their collections and research. HSI can provide highly informative images that allow for a direct communication of scientific findings to scientists, working with the historic texts, as well as to a large general online audience.     «

Hyperspectral Imaging (HSI) is considered a capable tool for the analysis of faded or deliberately erased historic documents. In HSI, the spectral reflectance properties of the documents are registered with high precision across the visible light (VIS) and near infrared (NIR) regions of the electromagnetic spectrum (EMS). The resulting data set contains the photographic (spatial) information combined with the spectral information for each pixel. Varying reflectance intensities of different mater...     »