Application of image processing technology on testing blending ratio and blending irregularity of blended yarns

Abstract

The blending ratio and blending irregularity of the blended yarn determine its property and quality. In this article, the image processing method was introduced to accurately and conveniently measure the blending ratio and blending irregularity of blended yarn by using images of the blended yarn cross sections. The Euclidean distance peeling method was applied to segment the adhered fibers to make them individual. Then, the fibers were classified according to characteristic parameters of fiber cross section such as color, shape, and area. The experiments were implemented to verify the accuracy and applicability of the image processing method to test the irregularity of blended bundles, slivers, and yarns by different fibers including polyester, viscose, and wool. The results show that the fiber identification error rate of the image processing method for blended bundles, compared with the designed, is less than 2%; the blending irregularity difference rate between image processing method and the manual counting method is less than 5% for blended sliver by sliver blending; the difference rate of the blending ratio of polyester/viscose blended yarn between designed and measured by image processing method is less than 5%, while manual counting method is less than 10%. The findings mean that the image processing method proposed in this article can greatly save labor and time on identifying and counting fibers, and it is a more accurate, convenient, and reliable method to test blending irregularity.