Research On The Fiber Identification Of Cashmere And Wool Based On Image Preprocessing

Textile fiber composition and content of label express is required in today’s trade. The development planning of Aqsiq also request to optimize the fiber identification technology and establish scientific and complete fiber supervision standards.Cashmere and wool fibers have similar structure form and the physical and chemical properties, and also because of the high quality and expensive price of the cashmere fiber, many undesirable merchants adulterate the cashmere products with wool fibers, so accurately and efficiently identifying the cashmere and wool fiber is very necessary. On the basis of computer digital image processing technology, the images of the cashmere and wool fibers are processed, and the characteristic parameters of the two fibers are measured and recognized in this paper. The main work is as follows:On the acquisition of cashmere and wool fiber images, the automatic fiber slice instrument is used to obtain good quality samples, and the differential interference phase contrast microscope is applied to acquire distinct fiber images in cashmere and wool fiber detection.In the part of the image processing of the cashmere and wool,the image enhancement methods, the image de-noising methods, the image segmentation methods, and the image modification methods are used in this paper. The MATLAB software is applied to do the image processing experiments and identify the suitable fiber image processing scheme for this paper. Eventually the edge information of the fiber and its scale are extracted in the single pixel binary map.In the part of extracting the characteristic indexes of the cashmere and wool fibers, The intuitive and relative characteristic parameters of the two fibers are compared and analyzed. The fiber diameter, height and diameter ratio scales are selected and measured.Finally based on the extracted data of the characteristic parameters of the two fibers, the BP neural network classification method is used to realize the automatic identification of cashmere and wool fibers.