In the process of artwork restoration a restorers report is written, which among others contains materials research of taken microscopic samples. The aim of this research is to identify organic and inorganic compounds using microanalytical methods. The results are used to interpret the painting technique of original paint and re-paints and to describe secondary changes of painting materials. The images of microscopic samples (in visible spectra, ultraviolet spectra and using electron microscope) are acquired in restoration process. The aim of proposed project is to design suitable methods for automatic description and following classification of these cross-sections and grain morphology of materials by means of digital image processing. Such description is crucial for material determination and following classification. The aim of the first stage is an exhaustive analysis of input data, which comes out of expert knowledge. We consider this stage an important preparation for finding suitable methods of image data description. The second stage will deal with specific proposal and description testing. The last stage will be dedicated to the classification design and implementation issues together with integration to Nephele system [1]. The whole project will be accomplished in close cooperation with Academy of Fine Arts in Prague.