Stereological reconstruction of polycrystalline materials

Abstract

Laguerre tessellations are suitable models for many polycrystalline materials. In this work, we present a reconstruction-based approach to fit a spatial Laguerre tessellation model to a plane section of a cellular material under the condition that one section of the model resembles the observed section of the material. To account for this special requirement, we introduce a novel Euclidean distance-based criterion for the model fitting. The model fitting itself is based on Simulated Annealing. If the structure under consideration is a Laguerre tessellation, we found a nearly perfect reconstruction of its spatial cell characteristics in the model. Even for a real sample of a sintered alumina the observed section is captured quite well by the model.