Computational approaches for structural analysis of wood specimens

Abstract

The structure tensor (ST), also named a second-moment matrix, is a popular tool in image processing. Usually, its purpose is to evaluate orientation and to conduct local structural analysis. We present an efficient algorithm for computing eigenvalues and linking eigenvectors of the ST derived from a material structure. The performance and efficiency of our approach are demonstrated through several numerical simulations. The proposed approach is evaluated qualitatively and quantitatively using different two-dimensional/three-dimensional wood image types. This article reviews the properties of the first- and second-order STs, their properties, and their application to illustrate their usefulness in analyzing the wood data. Our results demonstrate that the suggested approach achieves a high-quality orientation trajectory from high-resolution micro-computed tomography (μCT)-imaging. These orientations lead to establishing a description of fiber orientation states in thermomechanical models for fiber-reinforced composite materials. We conclude with an overview of open research and problem directions.