| The film transistor liquid crystal display(TFT-LCD)industry has successfully replaced the CRT display,with rapid development.Defecting detection is an important part in the production line of TFT-LCD.There will be some inevitable defects on the surface of TFT-LCD.And the large proportion of defects are speck and line defects.So it is significant to study the defection technology of speck and line defects.This thesis mainly focuses on the following points:(1)First of all,the structure,production process and displaying principle are deeply studied.After analyzing the reasons of and taking classification of speck and line defects,it makes a list of speck and line defects,which lays a good foundation for the latter work.(2)The thesis studies two methods to suppress the texture background in the defect defection: two-dimensional Fourier transform method and low-pass filter for texture background suppression in the frequency domain.The singular value decomposition is used to suppress texture background by setting the singular value of the first K representing the texture background to zero.Finally,The two algorithms are verified by experiments.(3)The thesis proposes a weighted template subtraction algorithm based on bilateral filtering.The proposed algorithm can eliminate the influence of uneven illumination.In the actual production,there may be some residual texture background points due to the difference of camera focus.Bilateral filtering can solve this problem.The maximum entropy is used to segment the defect region after the suppression of texture background.Finally,the feature of the defect region is extracted,including location,area and squareness.(4)The experiments of the algorithm with three different defects are setup.The thesis compares and analyzes the defection effect,defection accuracy and processing time.The results show that the proposed algorithm can extract the defects better,which can lay the foundation for the deeper research on defects of TFT-LCD. |
PDF Full Text Request