Stereovision for three-dimensional measurements of fabric pilling

Abstract

This paper introduces a three-dimensional (3D) imaging system designed for objective evaluation of fabric pilling. Thesystem was aimed at reconstructing high-fidelity 3D surfaces of fabric by using only two side-by-side images of a pilling fabric captured by a pair of regular digital cameras without special lighting. The robust calibration and stereo-matching algorithms were implemented to make the system insusceptible to fabric structures, colors, fiber contents and other factors. The depth data provide the most relevant information for pilling segmentation and measurements, because pilling is the protrusion of entangled fibers. 3D measurement data can be used for not only evaluating pilling appearance, but also for understanding pilling mechanisms in different abrasive treatments.