Estimating multi-phase pore-scale characteristics from X-ray tomographic data using cluster analysis-based segmentation

Recent advances in experimental techniques have made it possible to characterize and distinguish such micro-scale entities as fluid phase ditributions and pore geometry in porous media. In particular, non-destructive synchrotron based X-ray computed microtomography allows 3D resolution of individual pores and interfaces. However, separation of the various phases (fluids and solids) in the grey-scale tomographic images has posed a major problem to quantitative analysis of the data. We present an image processing technique that facilitates identification and separation of the various phases present in grey-scale X-ray tomographic images. The approach is based on a cluster analysis technique, used in combination with various other filtering and skeletonization schemes. We apply this segmentation algorithm to analyze multiphase
pore-scale flow subjects such as hysteresis and interfacial characterization. The
results clearly illustrate the advantage of using X-ray tomography together with cluster analysis-based image processing techniques. We were able to obtain detailed information on pore scale distribution of air and water phases, as well as quantitative measures of air bubble size and air-water interfacial areas. The method for segmentation of the data presented here represents an important advance and opens the door to an endless number of processes and phenomena that can be investigated using synchrotron based X-ray microtomography techniques.