Full-Text Search:
Home|Journal Papers|About CNKI|User Service|FAQ|Contact Us|中文
《Computer and Modernization》 2018-07
Add to Favorite Get Latest Update

TFT-LCD Circuit Defects Detection Based on Faster R-CNN

HE Jun-jie;XIAO Ke;LIU Chang;CHEN Song-yan;School of Physical Science and Technology,Xiamen University;  
The detection of tiny and complex defects in the border circuit of Thin Film Transistor-Liquid Crystal Display( TFTLCD) has been a difficult point in Automatic Optical Inspection( AOI). This paper detects TFT-LCD border circuit defects by using improved Faster Region-based Convolutional Neural Network( Faster R-CNN). The algorithm extracts features from shared convolutional layers firstly,and then generates candidate regions accurately through the multilayered Region Proposal Network(RPN),which can recognize and locate the targets combining with classification information. We analyze the performances of the method with different network structures we designed,and compare with different algorithms. The experiments trained in a border circuit dataset show that the method achieves excellent performance,and the detection system can recognize and locate six kinds of TFT-LCD border circuit defects in one image simultaneously within 0. 12 s and achieve an accuracy of 94. 6%.
【Fund】: 福建省高校产学合作项目(2016H6026)
【CateGory Index】: TN873.93;TP183
Download(CAJ format) Download(PDF format)
CAJViewer7.0 supports all the CNKI file formats; AdobeReader only supports the PDF format.
©2006 Tsinghua Tongfang Knowledge Network Technology Co., Ltd.(Beijing)(TTKN) All rights reserved