Image analysis of fabric pills and fuzz

Image analysis was used to characterize pills and fuzz on fabric surfaces in order to extend measurements beyond the limit of human visual ability. Pill features were described in terms of number, size, contrast, dispersion, shape and orientation. Fuzz was described in terms of quantity, arrangement of dense-transition-sparse regions, and height. To verify validity of the software developed for this project, pill features derived from machine vision were correlated to conventional pill ratings using human vision. Fuzz features derived from machine vision were measured for various fabric structures familiar to humans to validate machine-based measurements since no conventional human-based fuzz analysis method existed for comparison. Factors influencing pills and fuzz growth under different abrasion conditions were examined to provide systematic information for those phenomena previously taken for granted. Relationships between pills and fuzz also were investigated with machine vision.