Research On The Method Of Identifying And Classifying Grey Cloth Defects

Fabric defect detection plays an important role in the detection of the quality of the cloth.The traditional method of fabric defect detection with low detection speed and detection accuracy,detection of low level requirements workers experience higher shortcomings,can not meet the requirements of mass production of modern textile cloth,and testing equipment imported price is too high,not in the domestic large-scale promotion,therefore,fabric defect online detection system with independent intellectual property rights develop is significant.This paper studies the algorithm of fabric defect detection.Firstly,on the basis of the analysis of the traditional local binary pattern operators,the mean square error statistics are introduced,and the refined local binary pattern SLBP.The operator not only considers the relationship between the gray values of the local neighborhood pixels,but also the relationship between the variance of the whole image and local neighborhood variance is taken into account,Which is superior to the traditional LBP operator.Secondly,the strategy of complementary advantages is adopted to remedy the defects of different characteristics,combining SLBP detection with DFT phase transformation detection.During the detecting period,using the nonlinear gain and refined mathematical morphology methods to enhance the defect characteristics,and then the defects are detected and positioned by using standard deviation-based defect segmentation method.To effectively classify defects,this paper through principal component analysis(PCA)method and BP neural network method to classify the defects.In the defect classification process,firstly,the gray level co-occurrence matrix(GLCM)of the image to be measured is extracted,then,the principal component analysis(PCA)method is used to select the feature matrix,and finally through the BP neural network trained by defect classification of the principal component data obtained.The object of this paper is the 5 kinds of defects often appear in grey fabric,verified by experiment,the method of average classification accuracy rate reaches 93.06%.