ABSTRACT
Hyperspectral imaging has been a particularly useful technique for pigment characterizing on paintings and other types of art manifestations in previous years, thanks to its non-invasive character and because it is complementary to chemical analysis, providing high efficiency in results. This technique consists of capturing a number of images in the visible spectrum and the near infrared band using frequency intervals between 400 and 1700nm. False-colour composites facilitate the contextualization of images at different wavelengths as opposed to the image obtained by direct light, which is useful to highlight the information not easily seen by the human eye. One of the main problems for the generation of this type of composites lies in the capturing conditions of the images, obtained through cameras with different resolutions and scales, and influenced by the position, angle and distance to the portrayed object. This makes it difficult for the overlapping of RGB (red, green, and blue) channels on the direct light image with new channels from images captured with different techniques and camera conditions. Thus, the purpose of this research is the design and implementation of software tools that allow the automatic generation of false-colour composites that take into account the different conditions of the images being captured. The application substitutes each of the RGB channels of direct light photograph through the information obtained after applying techniques such as X-ray and hyperspectral imaging. The result is an artificial composite that facilitates researchers’ work in characterizing the pigments needed to create a more complex and improved interpretation of the artwork and its evolution in the restoration process. These interpretations will be of great support in the analysis, restoration and conservation of cultural heritage. Our approach provides a simple intuitive tool for quick preliminary analysis using non-invasive techniques, enabling the statistical analysis and subsequent segmentation. Preliminary tests on 14th Century altarpieces and Flemish paintings demonstrate improvement in the rendering of the paintings; as a result of the visible spectrum provided by these particular composites, the identification of pigments brought to light hidden information in the paintings.
