Design And Application Research Of Nonwoven Needle Defect Detection System Based On Machine Vision

With the arrival of "Made in China 2025",the development of intelligent manufacturing has become the main direction of innovation and upgrading of China’s manufacturing industry,which will comprehensively improve the intelligent level of production process.In recent years,machine vision technology has gradually become a hot topic in various fields of research.With the continuous development of machine vision technology,visual detection technology has been widely used in many fields with the advantages of non-contact,high precision and speed,especially in the field of product defect detection.As a common necessity and consumable in the knitting industry,non-woven fabric needle has higher requirements for its detection because of its small size,light weight,complex production process and high requirements for product quality.At present,most knitting manufacturers still rely on manual eye inspection for the detection of non-woven needles.This detection method not only has high labor intensity and low accuracy,but also can easily cause visual fatigue and reduce work efficiency for a long time.Therefore,it has important application value to study the intelligent system of non-woven needles detection process.In view of a series of disadvantages brought by manual eye inspection in the detection of non-woven needles,a non-woven needle defect detection system based on machine vision was studied to replace the traditional manual detection method,further improve the detection efficiency and accuracy,and realize the automation of the non-woven needle defect detection process.The main research contents of this paper include:(1)The main structure design of defect detection system.First of all,the detection system should have the function and detection requirements are analyzed,based on which the overall scheme of detection system and detection system workflow are designed.Secondly,the image acquisition system related hardware design and selection,including the choice of camera,lens,light source and lighting mode design;Thirdly,the mechanical structure design of the defect detection system was completed according to the overall scheme of the system.The structure design of the system was carried out by using Solidworks three-dimensional modeling software,which mainly included the design of needle feeding mechanism,needle cleaning mechanism,needle transmission mechanism,eliminating mechanism and needle connecting mechanism,as well as the selection of parts.Finally,all parts of the test system are integrated and assembled.(2)Design of PLC control system for defect detection system.PLC control system mainly includes hardware design and software design of two parts,first,the design of the control system overall scheme;Secondly,the hardware part of the control system design,including PLC selection,human-computer interaction interface module design and sensor selection;Finally,the control system software part of the design,according to the planned control system flow chart and electrical schematic diagram,the use of GX Works2 software to complete the control system PLC program writing and debugging,in order to achieve the detection system between each component coordination work and smooth operation.(3)Research on defect detection algorithm of non-woven needle.The research on non-woven needle defect detection algorithm is carried out,including three parts: image preprocessing,image edge feature extraction and defect detection.Firstly,in the part of image preprocessing,binary processing,filtering denoising and morphological processing are carried out on the acquired original image.Secondly,the edge feature of the preprocessed image is extracted to highlight the edge feature information.Finally,the design and experimental research of non-woven needle defect detection algorithm are carried out.The experimental results show that the designed detection algorithm can accurately identify non-woven needle.(4)System test and analysis.On the spot installation,commissioning and trial operation of the non-woven needle defect detection system,and taking the unit detection time,detection accuracy,false detection rate of good products and missing rate of defective products as evaluation indexes,the detection system was evaluated and analyzed.