Estimation of filament orientation in three dimensional CT image of textile fabric based on principal component analysis

Abstract

In this paper, a method to estimate a filament orientation is proposed in a three dimensional computer tomography (3D CT) image of a textile fabric for a structure analysis. In this study, yarn positional information is obtained as the structure analysis of the textile fabric by using the estimated filament orientation at each voxel of the 3D CT image. In previous study, the filament orientation was estimated by correlating a filament model, which is cylindrical solid model, with the 3D CT image. Since the maximum value of the correlation is calculated by an iterative calculation method, an extraordinary amount of time is required to estimate the filament orientation over the 3D CT image. In this paper, a different approach to estimate the filament orientation is proposed in order to shorten the processing time: a principal component analysis-based method is proposed. The axis of the first principal component is regarded as the filament orientation. The validity of the proposed method is confirmed by experimentally applying the proposed method to actual textile fabric. The effectiveness of the proposed method is then indicated by comparing the processing times of the filament orientation estimation by the proposed and previous methods.