Effective conductivity of a class of multicomponent heterogeneous materials by digital image analysis

Abstract

We develop computer vision techniques to calculate the effective (electrical or thermal) conductivity of multicomponent heterogeneous materials based on information collected from digital images of their microstructure. Image processing techniques are used in order to obtain the geometric and material properties of circular or near- circular inclusions embedded in a matrix. More specifically, a border-following algorithm is used to extract digital contour data pertaining to the interphase boundaries from grayscale images. The effective conductivities are then determined using an integral boundary equation formulation using the contour data extracted from images. Tests conducted on various images of multicomponent microstructures produce accurate results throughout a wide range of image resolution, demonstrating the efficacy and robustness of the developed approach. ☐ Keywords: digital image processing, border-following algorithm, effective conductivities, heterogeneous materials, composite materials.