Quantitative coating thickness determination using a coefficient-independent hyperspectral scattering model

Liesbeth M Dingemans,Roger M Groves,Aurèle J L Adam,Ping Liu,Vassilis M Papadakis

doi:10.1186/s41476-017-0068-2

Abstract

BackgroundHyperspectral imaging is a technique that enables the mapping of spectral signatures across a surface. It is most commonly used for surface chemical mapping in fields as diverse as satellite remote sensing, biomedical imaging and heritage science. Existing models, such as the Kubelka-Munk theory and the Lambert-Beer law also relate layer thickness with absorption, and in the case of the Kubelka-Munk theory scattering, however they are not able to fully describe the complex behavior of the light-layer interaction.MethodsThis paper describes a new approach for hyperspectral imaging, the mapping of coating surface thickness using a coefficient-independent scattering model. The approach taken in this paper is to model the absorption and scattering behavior using a developed coefficient-independent model, calibrated using reference sample thickness measurements performed with optical coherence tomography.ResultsThe results show that this new model, by considering the spectral variation that can be recorded by the hyperspectral imaging camera, is able to measure coatings of 250 μm thickness with an accuracy of 11 μm in a fast and repeatable way.ConclusionsThe new coefficient-independent scattering model presented can successfully measure the thickness of coatings from hyperspectral imaging data.

Highlights

Hyperspectral imaging is a technique that enables the mapping of spectral signatures across a surface
Point measurement techniques include destructive sampling and inspection under a microscope, ultrasonic techniques based on pulse-echo ultrasonics [1], spectroscopic ellipsometry and low-coherence interferometry [2]
In this work we propose a new method of thickness measurement for semi-transparent coatings, based on hyperspectral imaging

Summary

Introduction

Hyperspectral imaging is a technique that enables the mapping of spectral signatures across a surface It is most commonly used for surface chemical mapping in fields as diverse as satellite remote sensing, biomedical imaging and heritage science. Existing models, such as the Kubelka-Munk theory and the Lambert-Beer law relate layer thickness with absorption, and in the case of the Kubelka-Munk theory scattering, they are not able to fully describe the complex behavior of the light-layer interaction. Multiple layers are applied either with different material properties, or to facilitate the drying process when applying a thick coating. The point measurement techniques provide limited spatial information and the field techniques are typically slow and introduce challenges in scanning and automation

Objectives

Methods

Results

Discussion

Conclusion