Estimate of the Thermoelastic Properties of Pyrolytic Carbon based on an Image Segmentation Technique

Abstract

AbstractThe submicrostructure of Pyrolytic Carbon (PyC) can be described as an aggregate of approximately coherent domains with different orientations. The domains themselves exhibit a hexagonal material symmetry. In order to apply statistical homogenization schemes [1] which predict the thermomechanical properties in terms of the microstructural characteristics, a quantitative description of PyC microstructures based on experimental data is a precondition. In the study, a new efficient image segmentation technique is applied which extracts from micrographs the one‐ and two‐point correlation functions of domain orientation which are necessary for computing bounds of the thermo‐elastic properties of PyC. The segmentation technique, which segments the micrograph into regions with approximately coherent orientations, uses the distribution of Local Binary Patterns (LBPs) [4] for determining the similarity of adjacent image regions. This region‐based algorithm yields a first coarse image segmentation. (© 2010 Wiley‐VCH Verlag GmbH & Co. KGaA, Weinheim)