Analysis And Research On Fiber Distribution Of Hybrid Web Based On Image Processing

The mixed fiber material is made up of many kinds of fiber materials,which are usually made by the mixed fiber net through strengthening processes.And the distribution of the fiber on the mixed fiber network directly affects the structure and properties of the mixed fiber material.The traditional method of fiber mesh detection can detect the homogeneity and fiber orientation of single fiber network,but the fiber distribution of mixed fiber network can not be well detected.Therefore,it is significant for the application of mixed fiber materials to judge the distribution of fibers in the mixed fiber network quickly and accurately.This paper presents a method of extracting fiber distribution features from mixed fiber web images based on image processing in order to solve the above problems.Firstly,the whole design scheme and composition of the system are discussed in detail,and the image acquisition is carried out by using the CCD lens with IMX386 CMOS pixel of 12 million,so as to add the light compensator in the process of acquisition.Secondly,Sobel filter and weighted average method are used on Denoise and grayscale the collected image,and multistage stochastic resonance technique is used to enhance the target fiber information in the image.Then,the image cutting method is improved based on the global threshold method and the local adaptive threshold method to make it suitable for the image cutting in this paper.The target fiber contour in the mixed fiber network is extracted successfully,and the small size image is recovered and recognized.Finally,eight groups of experiments of four kinds of glass/polyester blended fiber and basalt/polyester mixed fiber were designed,according to the experimental steps above,and the images in each position of each group were processed separately.The relationship between the number of the target area and the collection location and fiber ratio in the mixed fiber network is analyzed.The experiment and analysis show that the distribution of the glass fiber in glass/polyester fiber network and the basalt fiber in basalt fiber/polyester fiber network can be distinguished by the established detection system.Compared with the traditional method,the method is more objective,detailed and reproducible.