Research On Inspection System Of Fluff Fabric Surface Quality Based On Machine Vision

Soft and dense fluff on the surface of fabric is generated by fluff-raising technology for improvement of quality such as appearance,style of the fabric,Wear resistance,warm-keeping,supple,wrinkle-resistance and so on.With the improvement of fluff-raising technology and product quality,the traditional manual detection methods can not meet the demand of high-speed,high efficiency for fluff fabric production.For this purpose,this paper proposes surface quality detection method of fluff fabric based machine vision which can increase detection speed with high accuracy by replacing manual operation.This paper is organized as follows:Firstly,the method of collecting fluff contour images is analyzed.Combining the tangential imaging of the pile fabric and the back-illumination method,contour image of pile fabric surface are obtained without interference by the texture and color characteristics of the fabric surface.By fusing the features of the fluff region and the background region,the characteristic region in the fluff surface contour image can be extracted efficiently.Secondly,the pre-processing of fluff images is proposed.The noise generated during the fluff image acquisition process is eliminated by dispelling high frequency signal from initial image.Subsequently,residual signal is fused with low frequency one to match the goal of reducing noise interference.At the same time,linear transformation technology is utilized to increase the contrast and light intensity of the fluff image.Thirdly,the method of edge feature extraction for fluff image is demonstrated.According to the distribution characteristics of the gray histogram of the fluff image,the maximum inter-class variance method is used for image segmentation.The defects such as voids in the segmented fluff region are identified.The linear and circular structural elements are constructed to perform morphological calculation on fluff region,and the complete calculation is obtained.Based on the Canny algorithm,the edge feature of the fluff is effectively extracted by the improved Zernike algorithm.Fourthly,a parameter model for evaluating the surface quality of fluff fabric is established.The image of the fluff region is analyzed by transforming the image coordinates to the Cartesian coordinates.In order to evaluate the fluff height,a least squares fitting algorithm is performed on the fluff edge contour to obtain a heightevaluation reference line,and a fluff height parameter model is established.The statistical histogram of the fluff edge in the sampling interval was calculated based on the average height of the fluff for determining fluff coverage parameter.The surface quality of the fleece fabric was comprehensively evaluated by the fluff height parameter and the coverage parameter.Finally,an experimental platform is built to validate the inspection technology based on machine vision for fluff surface quality.The design parameters of platform are determined and its structural components are designed for complete final test platform with computation software of the vision system.Experiments were carried out with three different colors of fluff fabrics.The test results indicate the detection method proposed by author can objectively reflect the surface state of the fluff fabric,which is consistent with the manual detection.In addition,53 figures,9 tables,80 references are included in this paper.