Abstract:

Aiming at the low efficiency and subjectivity of the visual observation of yarn hairiness of blackboard, a new hairiness detection method based on image processing was proposed. Image was acquired by the scanner. Then, the image was processed by median filtering, binaryzation, morphological operation and local threshold to obtain the blackboard hairiness images, and the number of hairy fiber pixels were counted. In addition, M Index was proposed to evaluate yarn hairiness on blackboard. The experiments were carried out by using different types of yarn material, yarn density and spinning system. The hairiness M Index of 18 kinds of yarns were measured. The regression model was established with the H value tested by a Uster tester, and the correlation coefficient between them is 0.975.The verification results of six kinds of yarn samples show that the hairiness detection method proposed in this paper can be used to extract the whole blackboard yarn hairiness relatively completely, and the M Index can assess the quality of blackboard yarn hairiness, with high algorithm accuracy and good reliability.

Key words: yarn blackboard, hairiness, image processing, local threshold