Segmentation of sandstone thin section images with separation of touching grains using optimum path forest operators

Abstract The segmentation of detrical sedimentary rock images is still a challenge for characterization of grain morphology in sedimentary petrography. We propose a fast and effective approach that first segments the grains from pore in sandstone thin section images and separates the touching grains automatically, and second lets the user to correct the misclassified grains with minimum interaction. The method is mostly based on the image foresting transform (IFT)—a tool for the design of image processing operators using optimum connectivity. The IFT interprets an image as a graph, whose nodes are the image pixels, the arcs are defined by an adjacency relation between pixels, and the paths are valued by a connectivity function. The IFT algorithm transforms the image graph into an optimum-path forest and distinct operators are designed by suitable choice of the IFT parameters and post-processing of the attributes of that forest. The solution involves a sequence of three IFT-based image operators for automatic segmentation and the interactive segmentation combines region- and boundary-based object delineation using two IFT operators. Tests with thin section images of two different sandstone samples have shown very satisfactory results, yielding r 2 and accuracy parameters of 0.8712 and 94.8% on average, respectively. Biases were the presence of the matrix and rock fragments.