Yarn Recognition And System Implementation Based On Machine Vision

With the release of 13th Five-Year Development Plan for Textile Industry,textile industry of China is gradually towards the era of industry 4.0.At the same time,with deep integration of internet textile industry,big data and other technologies,a variety of digital intelligent textile products emerge in an endless stream,it also lays a foundation for transformation of textile enterprises to intelligence.Yarn tube is a necessary winding tool for textile enterprises in production process.At present,There are three problems in yarn tube processing in textile enterprises: 1.kinds of yarn tubes and classification complicated.2.low sorting efficiency and high cost.3.mixed yarn and non-yarn,classification difficult.How to efficiently sort out these mixed yarn tubes is urgent problem.The detection method combined with machine vision technology not only performs excellently in efficiency of pipeline detection,but also has an irreplaceable position on detection accuracy,which becoming the mainstream product line detection technology at present.Therefore,this paper proposes a yarn identification and detection system based on machine vision.First of all,this paper introduces the overall design and technical index of yarn tube identification detection system.According to actual situation,this paper analyzes selection of industrial camera,industrial lens and light source.On the other hand,in order to improve accuracy of yarn detection,this paper mainly introduces the placement scheme between light source and detection target.Based on above system design,this paper studies the detection algorithm of yarn residue and sorting algorithm of yarn color.In this paper,a statistical algorithm of maximum connected domain based on image gradient method is proposed,which combines advantages of gradient algorithm and connected domain algorithm can accurately identify yarn tube with single loop yarn.In yarn color sorting algorithm,color histogram matching algorithm is used to realize the accurate classification of yarn color.Corresponding image processing software system is designed,which based on above image detection algorithm.Image processing software can be divided image acquisition function module,image algorithm processing function module,parameter setting function module,image storage function module,status test function module and registration upgrade function module.Data interaction with lower computer is realized through serial port communication,so as to complete precise sorting of yarn.