Image Processing Variability in Pore Structural Investigations: Conventional Thresholding vs. Machine Learning Approaches

Abstract

Summary This work compares the conventional thresholding and advanced machine learning methods in image processing of broad ion beam - scanning electron microscopy (BIB-SEM) mapping data. The test data set consists of representative BIB-SEM maps acquired from 12 Middle Miocene mudstone samples within a single stratigraphic interval in the Vienna Basin, Austria. Pore space segmentation was processed using the deterministic thresholding method in the software ImageJ and the machine learning-based pixel classification in the open-source toolkit ilastik. Total porosity, as well as pore size distributions and pore morphology resolved by ImageJ and ilastik were compared to determine the impact of different image processing approaches on the resulting segmentations. Porosity and pore size distributions obtained by ImageJ and ilastik are comparable; however, variations in pore boundary delineation and feature identification inevitably influence the results.