Automatic Analysis Of Matching Color Wool Ratio Based On Digital Image Processing

In wool spinning enterprise, a few kinds of wool fibers with different colors are blended with a certain percentage, which will make the color and tone of yarn reach some require. The process are called matching wools. This process is completed manually until now. It is time-consuming, strenuous and being affected easily by people's subject. The knowledge of color and dying must be mastered by matching wool people and the matching wool people must have higher technological capability of matching wool and rich work experience. How to increase the matching wool efficiency, decrease labor time and lower the demand of experience of matching wool people is the problem needed to be solved.In order to realize the automatic analysis of matching wool ratio and eliminate the time-consuming and experiential problems existed in matching process and so on, the digital image processing is used tentatively and FCM algorithm is proposed to process the captured images digitally and the proportions of fibers with different colors are counted.First, a few fibers of different colors are blended with some percentage on matching wool machine to get analytical sample and the images of blended wool are captured by iQsmart3 scanner. Then the colors of image are classed and the percentages of pixels with different colors in captured image are counted. In order to get the excellent effect of color separation, the RGB color space of images is converted to Lab color space. The pixels of images are classed in uniform color space using FCM algorithm.Scanning resolution and scanning area have a great influence on the experience analysis results. The images with different scanning resolutions, the images with different scanning areas but with same scanning resolution, are tested and analyzed in this paper. Experience results show: the higher scanning resolution, the better quality of image is and the more exact analysis results are but the larger calculate amount. In the case of the images having the same resolution, the bigger the area of image, the more stable analysis results that is to say the experience results are close to actual blended percentage of each color fiber but the workload of specimen preparation and calculate amount of image analysis are larger simultaneously. The scanning resolution of 3000dpi and scanning area of 40×40mm2 are chosen in this paper by comparing more schemes.The analysis results can be influenced by the choice of scanning background color of image, especially when the distribution of fibers is sparser. In order to obtain the exact matching wool ratio, the background color chosen should have a big difference with the colors of wool fibers.The matching color wool ratio of wool tops of three colors, four colors and five colors, the yarns of two colors, three colors, four colors, are analyzed automatically in this paper. The research results show: the method proposed in this paper has the features of fastness and accuracy, which can replace artificial matching wool method and has a good applied prospect.