Design And Research Of LCD Surface Defect Detection Platform

As an important part of the current display,LCD has been widely used in all walks of life.Its quality affects the function of the display.The research on defect detection technology,one of the important means of LCD quality evaluation,has important theoretical and practical significance.According to the problem of LCD surface defect detection,this paper designs a corresponding detection platform,including its workflow and structure,which provides an effective solution for LCD surface defect detection.In order to realize automatic detection on production line,the manipulator is guided by computer vision to complete the grabbing of LCD.Firstly,the centroid coordinates of the image are obtained by a series of image processing operations.Then the mapping relationship between the image coordinate system and the world coordinate system is determined by calibrating the camera.Then the actual position information of the LCD is calculated,which provides the necessary conditions for the automatic grasping of the manipulator.Based on the analysis of the basic theory of deep learning and neural network,the convolution neural network is applied to the surface defect detection of LCD,which avoids a series of tedious defect feature extraction problems in the current mainstream image processing methods.By making the collected LCD images into data sets that meet the input requirements of convolution neural network,and according to the characteristics of LCD defects,a convolution neural network model is designed for training.Finally,the comprehensive detection accuracy rate reaches 91%.The validity of deep learning method in surface defect detection of LCD is verified.The test bench designed in this paper realizes the automatic detection of surface defects of LCD,and the test results are efficient,reliable and robust,which can replace the traditional manual detection methods.