ON AUTOMATICALLY EXTRACTING THE STATISTICAL INFORMATION OF PORES IN HETEROGENEOUS MATERIALS FROM THE SEM MORPHOLOGY
Haolin Li, Shuhao Dong, Na Qin, Jiantao Liu, Yaoxiang Yu, Zhengqing Zhang, Muqing Wu, Chen Zhu · Journal of Porous Media · 2021
Heterogeneous materials, whose macroscopic properties largely depend on the distribution, shape, size, and number of internal pores, as well as inhomogeneities and defects, have been widely used in engineering applications. SEM imaging is an effective measure to quantitatively identify the microscopic characteristics of these materials by extracting and analyzing the relevant micro-composition. In this paper, a novel strategy is developed to automatically extract the statistical microscopic information about porous media containing pores of various complex shapes and sizes from SEM images by using the Bwlabel function to detect the connected components and fitting the relevant outlines using a proposed scheme. It is noteworthy that our method can be applied to analyze the pore constitution and extract the statistical information of porous materials with a maximum porosity of 0.3 and improve the integrity and generality of information extraction. The materials' micro-information obtained by this method provides a good support for better simulating the actual performance of materials. Two typical examples are presented to demonstrate our developed strategy.