The Characterization Study Of Hooked Fiber Based On Image Processing

In recent years, the qualities of finished and semi-finished textiles are more and more important with the improvement of living standards, and they determine the grade of the products. The number and direction of hooked fiber are so crucial for spinning that they determined the fiber unbend parallelism and yarn evenness. So it is important for spinning the yarn with high quality to define, character about hooked fiber, and provide related precise indicators, evaluate the hooked fiber during each working procedure from carding to spinning, minimize the number of the hooked fiber as far as possible.Previous studies of hooked fiber only stay in the adjustment of related process, or the judgment by copying the hooked fiber morphology, but it did not give a specific definition of hooked fiber and related indicators. At first, this paper is mainly to define the hooked fiber, which gives a reference to judge the hooked fiber in the follow-up work. It is convenience to extract the characteristic values of the fibers through characterizing the hooked fiber. The main contents of this paper can be outlined as follows:(1) Introduce the background the hooked fiber and analyzes the product process and their causes and harm of it, as well as how to reduce it;(2) According to the existing research, redefined and characterized of hooked fiber through the comparative analysis:the forms of fiber in cotton net and semi-product are folded in half and looped, crimp, tilt and hooked fiber (leading-hooked fiber, tail-hooked fiber, hooked fiber in two directions). According to the definition, we can identify the hook fiber. According to the characteristics of the hook fiber, it proposed fiber array index such as contour length, the effective length, parallel straight degree, characterization of parameters through the image acquisition on hooked fiber;(3) Discuss the preparation process of hook fiber, collect and sort images by using CCD camera, use MATLAB7.0as the software application platform tools to analysis the hook fiber image, introduce the process of pretreatment and algorithms of the gradation conversion, filtering, Gaussian-fitting, morphological processing, prune branches removed principles, and so on, and the images processing results are presented. The results show that it is more convenient to extract eigenvalues from binary image after image preprocessing. It laid the foundation for subsequent work.(4) Thirty images are selected to be recognized according to the definition and characterization of hook fiber, extract characteristic parameters of hook fiber and get the fiber contour length, effective length, straight degree, degree of hook, and other indicators. Based on the analysis and comparison, we conclude that fiber hooked-length, curvature K and hooked angle alpha comprehensively reflect the degree of fiber hook. When the length of hooked-fiber is the same, the bigger curvature is, the smaller hook angle is, which is to say the curvature is more obviously. It is more clear about hooked fiber features by using comprehensive auxiliary judgment.(5)The collected images has been judged by experts, and the results of expert judgment are compared with the results obtained by image processing analysis, which found that only by experience to judge did not give quantitative indicators and its floatability is unstable, however, each eigenvalue index processed by the image show the character of hook fiber more clearly, which has been a detailed reference for the follow-up process that. Hooked fiber pattern based on the image processing can make much better show on all kinds of characteristic value of the hook fiber. It provides a better basis for the study on formation cause of the hook fiber and their solution, which can develop new technology from the experiment data to prevent the formation of hook fiber in the greatest degree, so as to improve the quality of semi-finished and finished products.This paper is devote to characterize the hooked fiber based on image processing technology, which can provide theoretical basis for objective evaluation about hooked fiber, and technical support for the semi-product quality detection.