Objective measurement of pilling resistance in knitted fabrics with image processing techniques

Abstract

In the textile industry, measurement techniques based on image processing principles prepare to take over subjective evalution. In this study, MATLAB software was used to evaluate pilling of knitted fabrics as objective. Knitted fabric images were taken in the cycles of 1000, 2000, 3000, 5000 and 7000. Fabric's surface digitization, pills detection and segmentation were carried out from these images. Texture analysis was performed with Gray Level Co-occurrence Matrix (GLCM). After this phase, pill quantizations were made using images in matrix format obtained from image processing studies. Standart deviation, entropy and mean of matrix elements, pill count, pill area and contrast values increased with the increase of rubbing cycles applied on the fabric. Furthermore, the increase of rubbing cycles caused decrease in energy and homogenite.