Research Of Classification Techniques Of Cashmere And Wool Fiber Based On Image Processing

With the large variety of export of Wool blended fabric, the mixed identification of cashmere and wool fabric has been increasingly important. In order to quickly and accurately identify cashmere and wool fiber, more and more researchers have set their sights on the calculation method of image processing. This thesis focuses on the study classification and recognition technology of cashmere and wool fiber based on digital image. The main work of this paper is extraction, measurement and analysis to a large number of feature parameter. Cashmere fiber images dealed with in this paper are obtained by the optical microscope. After image segmentation, grayscale, binary, image growth, image denoising, edge detection, and expansion, refinement that a series of operations, the original image forms a binary image having the single pixel wide and continuous edge. And then the characteristic parameters are extracted. The improved central axis algorithm is designed in this paper to extract and measure the fiber diameter and the scale. The pixel measurement algorithm is designed in this paper to extract and measure the scale characterization area and the scale characterization circumference.The diagonal approximation algorithm is designed in this paper to extract and measure the scale factor outside the square. The center of expansion algorithm is designed in this paper to extract and measure the scale factor within the square. According to these measured parameters, the ratio of diameter and high, the scale density, the relative scale area, the relative scale circumference and the scale circumference factor, the five relative indexes is calculated.Finally, the automatic measurement of the above parameters is reality by programming and the appropriate database is established to store data for facilitating subsequent analysis and processing.