Research On Mobile Phone Screen Defect Detection Based On Machine Vision

As the rapid development of mobile communication technology,people's demand for mobile phones is rising.As the basic component of mobile phones,the 3D glass cover of mobile phone screens has a market size of nearly 19.2 billion yuan.The manufacture of glass covers has a huge market stock and growth potential,and is the focus of competition among many countries and enterprises.Among them,detection is the final process of glass cover production,which is the key to product quality control.The defects of mobile phone screens have various types,different shapes,and low contrast,which makes the detection of defects difficult and the accuracy of detection is low.In this paper,based on the problems of mobile phone screen defects,a mobile phone screen defect detection method based on machine vision is established.The main research work of the thesis is as follows:(1)The principle of the mobile phone screen defect detection system is introduced,the hardware selection of the image acquisition system is designed,the image acquisition platform is built,and the image acquisition scheme is designed.(2)In view of the error detection caused by the inconsistency of the template sample and the image to be measured during the defect detection process,this paper uses the Pat Max algorithm to locate the mobile phone screen image,and establishes the training model containing the pose information as the reference coordinate system.Then the affine transformation algorithm is used to correct the image after positioning,eliminating the influence of the inconsistency of the proportion and tilt angle of the graphics,and at the same time cropping the background part,reducing the target area,and achieving pixel-by-pixel alignment of the template sample and the image to be measured.(3)Aiming at the problems of irregular gray value distribution and low contrast of mobile phone screen images,an image enhancement method based on the combination of shear transformation and gray morphology is proposed.The method first uses shear transformation to decompose the image into two parts: low frequency and high frequency,then uses gray-scale closed operation and N×M median filtering to process the low frequency and high frequency images,and finally generates an enhanced image by inverse shear transformation.Experiments show that the proposed image enhancement method effectively eliminates image noise,small gaps in the target area are stitched,ensuring the integrity of the defect,and the contrast of the defect is significantly enhanced.(4)Aiming at the shortcomings of mobile phone screen defect types and the defects of traditional differential image method and Blob analysis method,a mobile screen defect detection method combining differential image method and Blob analysis method is proposed.First,select the defect-free image as the template image and denoise and blur it.Then use the improved difference image method to obtain the residual image of the mobile phone screen.Finally,a series of algorithms of Blob analysis are used to obtain the parameter information such as the coordinates,perimeter and area of defects in the residual image.On the basis of the above hardware selection design and image processing algorithm,a mobile phone screen defect detection system is built,and a mobile phone screen image instance is selected to perform a performance test on the defect detection system.Experimental results show that the mobile phone screen defect detection system designed in this paper has high accuracy and detection speed.At the same time,the defect detection system in this paper has a certain versatility,which can provide theoretical and technical support for the application and promotion of two-dimensional defect detection on the product surface.