Abstract:

In order to evaluate the yarn apparent evenness more precisely and continuously, a method based on sequence images of yarn was developed and applied to measure the yarn evenness. At first, an image acquisition system was adopted to acquire the sequence images of yarn. Then, the images were segmented to binary images based on FCM ( Fuzzy C-mean ) clustering. Finally, a threshold value was set to remove the isolate points and burr points in the binary image and acquire yarn core image. The number of pixels of each row was accumulated in the yarn core image to calculate the yarn diameter value. To verify the accuracy of the proposed method, seven kinds of compact-spinning pure cotton yarns with different counts were measured and the results compared with those by USTER® Tester 5-S800. It shows that the results of yarn evenness measured by sequence images are highly correlated with those by the USTER tester, proving the proposed method is accurate and feasible.

Key words: yarn diameter, yarn evenness, sequence image, fuzzy clustering, relevant analysis