Thickness determination of laminate core layers from microscopy images

Abstract

In sandwich laminates made of metal–polymer–metal (MPM) layers under forming conditions, the thickness changes along the thin-layered specimens are of particular interest. To measure this distribution, it is common to take microscopic images and determine manually the thicknesses at particular discrete points. We compare and provide several methods determining the thickness distribution from microscopic images. First, an interpolation scheme is chosen to obtain a continuous function of the interfaces between the layers instead of the discrete pixel data. Afterwards, the methods—ruled surface approach, orthogonal projection, embedded circle scheme, and a dull but simple method—are applied and compared with each other. For very smooth data, most of the schemes show equivalent results. However, the “brute force”—called the dull approach—turns out to be the most robust scheme, particularly, if noisy data is studied. The Gaussian error propagation concept is applied to study the uncertainties resulting from noisy pixel data. The schemes are adapted to finite element simulation results as well so that a direct comparison of experimental and numerical data would be possible.