Improved Surface Area Estimation Based on Surface Curvedness.

Accurate determination of interfacial/surface areas in multiphase systems is of essence to enhance our understanding of multiphase flow and mass transfer processes in porous media. X‑ray micro-CT provides promising means to estimate surface areas from three-dimensional segmented images. However, surface area estimates for a continuous object based on its discrete representation are approximations. Surface area calculations are many-to-one mappings where various objects with slightly different shapes or sizes might exhibit the same discrete representation; hence yield similar surface areas. Several classes of estimators have been proposed in literature, including methods that assign weights to specific voxel configurations, that reconstruct approximate surfaces with polygons or integrate surface voxels intersecting with a set of uniformly distributed lines. Especially, weighted-voxel based methods that determine weights by optimizing planes that approximate curved surfaces yield poor estimates for planar surfaces when compared to simple surface-voxel-face counts. To overcome this limitation, we present a new surface area estimator that is based on surface curvedness computed from principal curvatures. A curvedness threshold is applied to discern surface voxels with either curved or planar surface neighborhoods. While for voxels with curved neighborhoods an optimized weighted-voxel method is applied, voxels with planar neighborhoods are treated with the surface-voxel-face count method. Preliminary results for various surface configurations and comparison with standard methods will be presented.