Texture Analysis of Grey-Tone Images by Mathematical Morphology: A Nondestructive Tool for the Quantitative Assessment of Stone Decay

Abstract

Both a low-cost and easily handled nondestructive methodology and its validation criterion are proposed. The methodology is based on image analysis by mathematical morphology for the assessment of decayed stone surfaces in historic limestone buildings. It is adapted to follow the evolution, at macroscopic time and space scales, of stone materials used in art pieces and monument building stones. This methodology is applied to the quantitative analysis of textures of static grey-tone CCD video camera images representative of flat stone structures that cannot be handled. These structures line the walls inside an important church, Basilica da Estrela, built in the 18th century and located in the city of Lisbon. Detailed visual/tactile observation shows that these flat vertical structures are more or less damaged depending on their position inside the church. The damage is possibly associated with different environment conditions both from the atmosphere inside the church and from contact with the walls of the church. The presence of these pathologies breaks down the order and organization of sound rock textures introducing changes in the topographical and optical characteristics of the texture of the surfaces. This new methodology is based on the granulometry and covariance analysis of grey-tone images corresponding to the structures studied. The validation criterion allows the results of the proposed methodology to be compared to the results of a previous qualitative study made by experts in the field using visual assessment and monument mapping as a valid methodology to assess the degree of decay. This procedure is based on optical and topographical characteristics and type of decay of the stone surface, and the quantity of material lost from the surface of the panels during the year. As discussed in this paper, in the context of cultural heritage, the proposed methodology was validated and its results can be considered an improvement on and complementary to expert visual analysis and also to other nondestructive techniques.