June 24, 2019
Journal Article

Downscaling-based segmentation for unresolved images of highly heterogeneous granular porous samples

Xiaofan Yang
Svyatoslav Korneev
Ilenia Battiato
Direct numerical simulations of pore-scale flow and transport in natural sediment columns require the accurate knowledge of pore-space topology. The limited resolution of X-ray tomography is often insufficient to fully characterize pore space structure within fine-grained regions. As a result, single and multilevel threshold-based segmentation approaches are customarily employed to identify solid, pore and porous solid regions in the image by means of grey intensity thresholds. While the choice of cutoff thresholds is arbitrary, it dramatically affects the effective properties as well as the dynamical response of the reconstructed porous structure. We propose a simple and efficient algorithm of downscaling followed by segmentation to reconstruct the unresolved pore space from X-ray computed tomography (XCT) images of natural geological porous media columns. Contrary to widely employed single and multilevel threshold-based segmentation, the proposed method is based on a continuous map between pixel intensity and local pixel porosity that does not rely on the definition of arbitrary thresholds. The approach is used to generate a high-resolution binary image of the porous medium from poorly resolved grey-scale images while preserving important sub-resolution information. First, we validate the method on a synthetic unresolved image of a heterogeneous porous medium and compare its known pore space distribution with the extracted one. The comparison shows that the method better represents the initial distribution than a comparable threshold-based segmentation and is more computationally efficient than stochastic reconstruction. Second, we apply the method to extract the pore space distribution from unresolved XCT images of two natural sediment columns, and use it to parametrize a capillary bundle model. The latter is then used to estimate the hydraulic conductivity and breakthrough behavior of passive solute transport. The comparison between model predictions and data is excellent and demonstrates that the algorithm is accurate across multiple length scales.

Revised: June 24, 2019 | Published: April 1, 2018

Korneev S., X. Yang, J.M. Zachara, T.D. Scheibe, and I. Battiato. 2018. "Downscaling-based segmentation for unresolved images of highly heterogeneous granular porous samples." Water Resources Research 54, no. 4:2871-2890. PNNL-SA-125744. doi:10.1002/2018WR022886