Classification of Visual Smoothness Standards Using Multi-Scale Areal Texture Parameters and Low-Magnification Coherence Scanning Interferometry

The ability to objectively specify surface finish to ensure consistent visual appearance addresses a vital need in surface coating engineering. This work demonstrates how a computational framework, called surface quality and inspection descriptors (SQuID™), can be leveraged to effectively rank different grades of surface finish appearances. ISO 25178-2 areal surface metrics extracted from bandpass-filtered measurements of a set of ten visual smoothness standards taken on a coherent scanning interferometer are used to quantify different grades of powder-coated surface finish. The ability to automatically classify the standard tiles using multi-scale areal texture parameters is compared to parameters obtained from a hand-held gloss meter. The results indicate that the ten different surface finishes can be automatically classified with accuracies as low as 65% and as high as 99%, depending on the filtering and parameters used to quantify the surfaces. The highest classification accuracy is achieved using only five multi-scale topography descriptions of the surface.