Development of Fabric Image Models and Invariance Property of Variance-Area Curve Within Fabric

Abstract

This thesis is concerned with the development of two prediction models for fabric quality, 2-D and 3-D images for visualization of fabric qualities, and the application of variance-area curves in fabric quality rating. Two types of variance-area curves, CV(A) curve and CB(A) curve, are introduced. They together provide a new way to judge the fabric quality quantitatively in its dependence on the measured area. An invariance property of the variance-area curves within a fabric was also investigated by using fabric prediction models and measured yarn data.

Based on two different ways in which weft yarns are mapped onto a fabric, two models are developed by simulating the fabric using measured yarn data and, subsequently, 2-D macro and 3-D micro fabric imaging methods were developed. Application of variance-area curves to woven fabrics and its invariance property were also studied. These are done by using MATLAB® programs.

The actual yarn data were captured by an advanced yarn measurement system incorporated by a CTT (Constant Tension Transport) unit, a 12 bit CCD line scan camera, and a program written in C, called "camera1", running on a Linux Operating System platform.