Pore structure characterization and classification using multifractal theory—An application in Santanghu basin of western China

It is difficult to extract quantitative information of pore space from thin section. A novel method to characterize the pore space using multifractal analyses is presented, with the help of mercury injection capillary pressure (MICP) and nuclear magnetic resonance (NMR) transversal relaxation time ( T 2 ) data. Thin sections were processed via multi-threshold segmentation using particle swarm optimization (PSO) algorithm to obtain binary images containing pore space and mineral matrix. After multifractal analyses of binary images, the interrelationship between multifractal parameters and pore structure is investigated in detail. A crossplot for pore structure classification is put forward with types predefined by MICP and NMR experiments. The result shows that different pore types demonstrate their unique multifractal characteristics. Multifractal dimensions are positively correlated with the median capillary pressure, whereas negatively correlated with average pore radius and the T 2 geometric value. D min and D 0 can be served as good indicators of pore structure types in the case study. This study provides an effective way of pore structure characterization and classification based on thin section, especially for regions without MICP and NMR experiments.