Abstract: In order to detect the length and root number of yarn hairiness more accurately, based on the combination with video microscope and image processing technology, a novel hairiness detection method was proposed. The yarn images that captured with a MOTIC SME-140 video microscope were processed with grayscale conversion, image segmentation, morphology opening and image thinning sequentially in order to generate both the images of yarn core and thinned hairiness. Then considering the yarn core edge as the measurement baseline to judge the hairiness cut-point, hairiness root numbers of different lengths were finally obtained. The detection result of hairiness root number in each piece shows a significant stability, and is close to the result of visual counting. Therefore, the proposed approach of hairiness detection in this study is more accurate and reliable than the traditional photoelectric method.

Key words: yarn, hairiness, image processing, automatic detection