Abstract
Our paper presents a low-cost, low-technology quantitative approach to the study of the Cucuteni pottery matrix from the point of view of porosity characterisation. The study focuses on 9 samples previously characterised through complex physicochemical analyses, providing thus an interpretation baseline. In a first stage we assess the reliability of the images obtained through optical microscopy against SEM micrographs. We then use non-supervised partitional clustering in order to explore the informative potential of the digital image analysis of the photographs obtained through optical microscopy. We conclude that the method successfully discriminates between different types of secondary pores, but larger sample collections need to be used for a thorough characterisation of primary pores and discrimination from secondary shrinkage cracking.