Studio della porosità attraverso la segmentazione di immagini in microscopia elettronica a scansione

Porosity is a primary feature of materials, due to his control on materials physical-mechanical properties and his influence on theirs deterioration rate. Porosity is quite a variable feature, both because the wide variability of pore’s areas and geometries, both for pore’s interconnections morfology. So this make porosity very difficult to describe. The better way for the description is a multitechnique approach. One of this techniques, Image Analysis, it has become one of the most used techniques, due to his capability of measure the total porosity and quantify pore size and geometries, and distribution of pore size. The master thesis will be focused on the study of measure’s variations of porosity and pores size distribution of two bricks samples. Measurements were obtained by segmentation of backscattered electrons SEM images (SEM-BSE), of different dimensions, but with the same dimensional ratio for the same sample. This work is starting from the work of Coletti et al. Published in 2016 on Materials Characterization, in which was built a panoramic image from 30 smaller images that showed an excessive variability of porosity due the variability of pore sizes. In this thesis work I have determined how porosity, pore sizes and theirs distributions changes in relation to image dimension, with a image magnification of 50X and 500X. I have analysed also the relationships between porosity value of smaller images with same dimentions. At last, the problems that occour during segmentation process were analysed, as well as resolution strategy. Results confirm that porosity variabilità it depends from larger pores size and from the dimensional distribution of pores. The choice of the dimensions of the image for the porosity’s analysis therefore depends on the porosity’s characteristic, and so on the factors that affect it, such as the nature of the sample and his termic and reologic history