Fabric Surface Roughness Evaluation using Wavelet-Fractal Method

Abstract

An objective and reliable evaluation method of fabric pilling using a three-dimensional scanning system with higher accuracy is presented. The overall fabric surface roughness together with the pilling characteristics were evaluated to comprehensively understand the fabric pilling phenomena and exactly grade the degree of pilling. The fractal dimension calculated by the wavelet-fractal method and the surface average mean curvature were used as descriptors of fabric surface roughness. Localization and characterization of pills was achieved by wavelet reconstruction. The number, area, and population density of pills were extracted as the parameters of pilling characteristics. In order to select features and then reduce dimensions, a Karhunen–Loève (K–L) transform was employed. Bayes, minimum distance, k-nearest neighbors, and neural network classifiers were used to classify the fabric pilling into objective grades. The experimental results demonstrated that the fabric pilling evaluation system developed in this study represented both the fabric surface properties and the pilling properties and also showed high accuracy in grading the degree of pilling.