Detection Of Structure And Parameters Of Single Weft Knitted Fabrics Based Image Processing

In the industrial design of knitted products, analysis and recognition of fabric need to test the fabric density, stitch length, unfilled coefficient and so on,which are the main factors in affectine the physical and mechanical properties of knitted fabrics. In addition, sample processing or imitation of the design usually takes the organizational structure of knitted fabrics. To complete previously tests of these parameters and the structure of the fabric mainly rely on manual with a magnifying glass, the cloth mirror and other tools. This method demand rich experience and good vision of Technical staff. Simultaneously Recognition cycle of analysis and identification is a relatively long period and the analysis process is also dull and boring, thus which lead to inevitable human errors. Therefore, to solve error and time-consuming issues in the manual measurement of the structural parameters of knitted fabric, it has become necessary to research on how to get and analyze fabric classification, yarn diameter, fabric weave structure with computer automatically. Uses of image acquisition device transfer the fabric image to the computer. And we use effective recognition algorithm to extract information from the fabric and automatically detect and identify structural parameters. Further analysis of fabric images accurately to get fabric classification, Organizational rules and technical parameters.This paper presents some algorithms to get and analyze fabric Construction parameters and weave structure using Digital Image Processing and Pattern Recognition technology, and develop the flow and technique route of automatic fabric analysis and recognition. First, we analyzed the method of fabric image getting and pre-processing to represent more information of fabric construction parameters and weave structure. Then we depict particularly the algorithm of getting fabric construction parameters-density, yarn diameter, stitch length, unfilled coefficient, etc. using gray sum to get wave spectrum of organizations cycle, then the wavelet transform to eliminate jagged peaks, difference method used to pick up the location and number of peak points, Then calculate the wale and course of knitted fabric. By selecting a gray value as the threshold with optimum threshold algorithm, we can convert the fabric image to black and white. While the balck represent yarn. Then we can calculate the diameter of yarn. We get tightness factor by the same method. In the end, we contrast the result got by algorithm with the result calculated manually, and this technique can measure the knitted fabric's parameters quickly and correctly.To Identification of the weave structure for the fabric, this paper firstly used two-dimensional wavelet transform to descript and extract low-frequency, high-frequency signal feature. Secondly we used gray co-occurrence matrices to extract image gray scale space-related characteristics, which include angular second moment, entropy, contrast, and correlation to characterize fabric. And then use Learning Vector Quantization Neural Network to classify the fabric. We reach the purpose of automatical identification and analysis of the characteristics and advantages of this algorithm.In this thesis, research and development based on MATLAB, using MATLAB's powerful functions, achieve fast and efficient computing and image processing.And it proposes a method which can detect structure and parameters of simple single weft knitted fabric.