Fabric defect detection system based on digital image processing
Zhaozhun Zhong1, Pengjie Qi2, Miao Guan1, Yuedong Xia2, Yuanhui Fu3
1School of Iron and Steel Soochow University, Suzhou City, Jiangsu Province, PR China, 215021
2School of Mechanical and Electrical Engineering Soochow University, Suzhou City, Jiangsu Province, PR China, 215021
3Humanetics Innovative Solutions, Inc., Plymouth, USA, 48170
A fabric defect detection system based on digital image processing for textile fabric is proposed in this paper. The approach for the classification and identification of three commonly encountered classes of fabric defects (holes sto, missing end and mispick) is studied. The developments of both the hardware and software structures are presented. Firstly, median filter preprocessing and image segmentation based on Otsu threshold are applied to localize the fabric defects. Secondly, the features based on grey-level histogram and geometry are extracted. Thirdly, the classification and identification are accomplished by the method of artificial neural network based on the extracted features. Finally, a variety of textile images with different defects are tested to evaluate the performance of the proposed defect detection system. The experiment results indicate that the proposed system works efficiently with high accuracy, which can meet the requirements of the textile industry.