Hair density distribution profile to evaluate yarn hairiness and its application to fabric simulations

摘要

This paper describes a new method for yarn hairiness evaluation that calculates the hair densities at varying distances from the yarn core using an image processing technique. The method is based on integrating the number of pixels for distances incremented by a pixel size from the yarn core edge. The experiments with various yarns showed that the hair density distribution profile exhibits two different exponential behaviours, one below and the other above, approximately 0.75 mm from the core. A total hairiness index is also calculated using the total number of hair pixels. A good correlation is observed between the proposed THw and Uster's H index especially for grey cotton and cotton/polyester samples. A novel technique that generates realistic yarn simulations using the total hairiness and the hairiness distribution data along with the diametric irregularity is introduced. The simulations created are compared with actual yarn images in a qualitative manner. A simple single jersey knitted fabric simulation algorithm is also described utilizing the yarn simulation including hair distribution data, which gives a more realistic simulated fabric appearance.