Research On Defect Detection And Classification System Of Glass Fiber Electronic Fabric Based On Machine Vision

Glass fiber electronic cloth production process,can be divided into drawing,winding yarn,warping,weaving,post-processing,inspection cloth and other processes.The cloth detection process is an important link in the production of electronic cloth,and it is a key link for acceptance and control of product quality.At present,domestic electronic cloth companies still adopt the manual cloth detection method in the cloth detection process.Due to the problems of slow detection speed and low detection rate due to manual detection,the long-term detection has an impact on employees’ vision health.Therefore,there is an urgent need for an electronic cloth detection method capable of stable operation over a long period of time with high accuracy and speed.In this paper,through the field investigation and investigation of related companies,combining detection objects and project requirements,a hardware platform with multiple CMOS line array cameras,industrial computers,light sources,and mechanical structures is built on the cloth inspection machine.Based on the analysis of the collected images,a defect detection algorithm using DOG as a filter is studied and designed.The detected defects are first manually classified,then gray-level co-occurrence matrix feature extraction is performed,and multi-class SVM classifiers are trained to implement the defect classification function.The C# language is used to develop the electronic fabric detection and classification system software in the Visual Studio 2010 programming environment.SQL Server 2008 is used as the system database to save information.The software includes user management,real-time images,point navigation,report export,and other auxiliary functions,which are of good practicality.Through the installation and commissioning of the system in the company’s cloth detection workshop,the appearance test of the electronic cloth is performed.The results showed that the various performance indicators of the system achieved the expected goals.Finally,according to the system usage,this paper summarizes and analyzes the achievements and deficiencies in the study,and looks forward to the next step of the scientific research work of this topic.