Development And Application Of Intelligent Fabric Defect Detection System Based On Digital Image Processing Technology

With the development of large-scale textile industry, quality control of textileproducts is becoming increasingly important. In the background that our labor pricesrise, and costs of manual detection rise, but prices of computer hardware and softwarefall, in order to response to the state’s call to promote traditional industries intelligence,investment of research and development based on image processing fabric defectdetection system is urgent and meaningful.With observation of the fabric production process and research on other defectdetection algorithms, we propose a new type of automated fabric defect detectionalgorithm, then design and develop the intelligent fabric defect detection system in thisarticle. The system uses CCD cameras to collect image data, and completes the defectdetection of high-speed moving fabric using digital image processing technology toanalysis and calculate the image data. When discovering defect, the machine should beshut down promptly to reduce losses, and it should notice workers do remedial work.First, using the image processing part of the system, the input images are segmented.According to the features, the preliminary judgment of defects is achieved. For theundetermined case, using the achieved information, original images are binarized bythe otsu adaptive method and processed by mathematical morphology. Finally, thetwo-valued results are analyzed to determine the defects.During practical applications in the textile factory, the system which is subjectedto tests of mass data has a high reliability and accuracy. The system which replaces thetraditional manual detection and saves manpower is real-time and has good accuracy.It controls the fabric production and quality effectively, uniforms the product qualitystandards, saves resources and costs, and it is conducive to the development of thetextile industry.