Research On Recognition And Analysis Of Woven Fabric Structure And Parameters Based On Image Processing

Nowadays, in the field of Textile industry, the structure of woven fabric analysis and recognition mainly depend on manual work or special equipment. Though this way is authoritative, it is not easy to manipulate, and hard to master. Moreover, it's time consuming and tedious. So it has become necessary to research on how to get and analyze fabric construction parameters,weave structure with computer automatically.In this work, we provide 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 technical 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------warp and weft densities, the tex of warp and weft: decompose the pre-processed images with discrete wavelet transform, the decomposed images include longitude and latitude fabric information, adopting the method of data average to get rid of the impact caused by fabric deformation and scan image deviation, and turn the images into Binary image, then we can get the warp and weft densities from the Binary image, the tex of warp and weft can be calculated with yarn average diameter. In the end, we contrast the result got by algorithm with the result calculated manually.In this work, We provided the flow of fabric weave structure analysis and recognition: First, find the yarn's junction by analyzing the brightness changes of images in the horizontal and vertical directions, Thus making it possible to determine the location of the intersection points; Then, acquire the size of the weave repeat unit with autocorrelation function, and get the weave repeat unit; Then, classify the intersection points: do the first classification according to the different brightness of different positions, afterward, Identify the intersection points in the same float through edge extraction of the weave repeat unit, thus can divided the intersection points into two categories: one is warp over weft, another is weft over warp; Finally, Drawing pattern draft according to the distribution of the intersection points.On the issue of recognition of complex fabrics including different structures, process texture segmentation to fabric image at first; After that we can get regions including one structure, also can get the proportion and position of different regions; Thus can using the methods of structure recognition to analyze the single region. We provided different algorithms of texture segmentation for fabric gray image and color image: Use texture segmentation based on texture spectrum to process gray image; for color image, Integrate color and spatial information for segmentation on the basis of the color clustering; These two algorithms both gained good results.The technical route and algorithms we provided above, have gained feasibility validation and achieve applicable results, have certain theoretical value and use for reference in the domain. Especially pattern draft recognition method has some innovation in this field. The texture segmentation methods of complex fabrics is also useful and practical.