Abstract:

In order to solve the empirical and time-consuming problems associated with wool blending in wool mills, FCM algorithm is proposed to count the proportions of different color fibers in wool top so as to enhance the wool blending efficiency. On the basis of the images captured by a scanner, the images are converted from RGB mode to uniform color space of L*a*b* mode. Then FCM algorithm is used to cluster and analyze all the pixels in the image. Finally, the percentages of different color pixels are obtained and blending ratio are also obtained through statistical analysis of classification results. The effects of scanning resolution and the size scanning area on test results are also discussed. The results show that the percentages of different color pixels is relatively stable and close to the actual value when the scanning resolution and scanning area are not less 3000dpi and 40mm×40 mm respectively.