Design And Research Of A Defect Detection System For Glass Panel

Defect detection of glass panel is one of the important links in quality control of electronic products.How to detect the defects of glass panels accurately and efficiently,is an urgent problem to be solved in industrial production.In recent years,deep learning has made great progress in many fields and has become a key technology for breakthroughs in the field of machine vision.In order to detect defects in glass panels quickly and accurately,this paper studies glass panel defect detection algorithms based on deep learning,focusing on solving two types of tasks in glass panel defect detection,namely,identifying whether there are defects in glass panels,defecting and classifying the types of defects in the glass panel.At the same time,the detection algorithm is evaluated.In order to identify whether there are defects in glass panel,this paper studies a glass panel defect detection algorithm based on Generative Adversarial Networks(GAN).Firstly,this paper constructed a model for image repair based on GAN.If there are defect areas in the samples,the network can repair these defect areas.Then,the difference between the original image and the repaired image can be used to determine the defect area in the glass panel.The results show that the defect detection algorithm used in this paper performs well.In order to detect the defect location and category in the glass panel,this paper studies a glass panel defect detection algorithm based on Faster R-CNN.This paper constructs a data set based on the defect pictures of glass panels collected on site,and expands the existing defect pictures.In this paper,K-means++ clustering method is used to improve the anchor generation method.Then different feature extraction networks and anchor generation methods are used for comparative analysis,and the appropriate feature extraction network and anchor generation method are selected.By appropriately reducing the number of anchors selected after Non-Maximum Suppression,the speed of defect detection is improved.The results show that the defect detection algorithm has high detection accuracy and speed.Finally,This paper studies a glass panel real-time defect detection platform,which mainly includes image acquisition module,motion control module and defect detection module,etc.,covering the entire process of glass panel from image acquisition to defect detection.