Objective evaluation method of fabric pilling based on image analysis

摘要

Resistance to pilling is an important test among all fabric performance tests. So far, the subjective evaluation based on visual by expert is not ideal in the consistency and accuracy of rating results. A new objective evaluation method of fabric pilling using image analysis is put forward in this paper. After acquiring fabric pilling images treated with histogram equalization, median filtering, image segmentation based on Pulse Coupling Neutral Networks (PCNN) and area filtering, etc. Pill images are segmented from background. Five features of pills are extracted from two value image, which are the number of pills, the area of pills, the density of pills, the area standard deviation of pills and the largest area of pills. At last, template matching classifying method is adopted to realize the objective evaluation of fabric pilling grading. Different fabric samples are tested in the experiment. The method is proved validly and the accuracy is about 90 percent.