DISCRIMINATION OF CERAMIC TYPES USING DIGITAL IMAGE PROCESSING BY MEANS OF MORPHOLOGICAL FILTERS

摘要

Digital processing of images can be used to analyse ceramic pastes, contributing in this way to the discrimination of ceramic types. The method involves several stages of analysis: digitization of images of ceramic pastes, pre-processing of the images, particle segmentation, calculation, quantification and classification of the particle sizes. This procedure uses interactive algorithms of segmentation of images and grain size analysis derived from mathematical morphology theory. To prove the efficiency of the method, we processed digital images of polished thin sections of ceramic samples of Cypriot wares from the Late Bronze Age (1550-1450 bc) and their Egyptian reproductions. Such groups have been previously analysed by optical microscopy, XRD, XRF, ICP-MS and EPMA. The results of the analysis were able to discriminate between these two different wares.