Image analysis to evaluate removal of particles from fabric surface

Abstract

We propose an objective method to quantify the solid-particle removal rate from fabric. The method extracts the ratio ( K/S) λ of light absorption K to scattering S at wavelength λ from fused digital images captured under a stereo photometric system in which illumination is from four directions. Three different white polyester fabrics were contaminated with iron oxide particles. Digital images of the fabrics were obtained before contamination, then before and after cleaning. The ( K/S) λ ratios extracted from images were used in a fabric-detergency formula to determine the solid-particle removal rate. Digital image acquisition conditions were optimized to minimize the effects of fabric structural factors. Our method was faster, more accurate, and cheaper than existing methods. Moreover, it is nondestructive and does not require a tracer. The average accuracy of the proposed method was improved by 44.77% compared with the existing surface reflectance method and by 42.51% compared with the binary image method. Moreover, the accuracy was further increased by calculating ( K/S) λ for a signal that corresponds to the color of the contaminant particles. This method can be used to quantify the effectiveness of self-cleaning textiles and garment-care machines.