Research On Key Techniques Of Quality Visualization Inspection Of Light Guide Plate

As an important part of the LCD backlight module,the light guide plate will directly affect the display effect of the screen and affect the user experience.However,in industrial production,due to manufacturing processes and other reasons,the light guide plate products may have defects such as scratches,bright spots,black spots,and dirt.At present,there are still many shortcomings in the domestic manual optometry of the light guide plate.In this paper,the research on the single-side light guide plate is carried out.According to the inspection requirements of the enterprise,a set of single-side light guide plate quality visual inspection system is developed,which includes the hardware part and the software part.Through offline testing and online testing,the system has high detection rate and stability.Specific research and development contents are as follows:(1)Hardware device designThe hardware device design mainly includes three aspects:mechanical architecture,machine vision equipment and electrical equipment.The mechanical structure is responsible for supporting the operation of the entire system.The light guide plate enters the inspection equipment from the production area and is transported through the belt.Before the test,it is limited by the clamp cylinder so that it does not exceed the detection range.After entering the detection area,the photoelectric sensor will be triggered to generate the signal,the PLC controls the light source to illuminate,and the line scan camera captures the picture.By analyzing the product characteristics and defect types of the light guide plate,the camera,the lens,the light source,the capture card,and the processor are selected,and how the light lighted is decided to ensure that the defects can be clearly imaged.The electrical part processes the associated sensor signals to ensure that the system operates stably.(2)Preprocessing algorithm designThe image of the light guide plate collected has some problems such as image noise,excessive useless area and tilting direction.In order to facilitate subsequent algorithm detection,image denoising,ROI extraction and tilt correction processing are first performed.By comparing the effects of mean filtering,median filtering and Gaussian filtering on the image,the mean filtering is used to reduce the noise of the light guide plate image.In order to shorten the detection time and obtain a better segmentation effect,OTSU is used to extract the main portion of the light guide plate.Finally,when rotating the image,bilinear interpolation is used to ensure that the details of the image are not distorted.(3)Detection algorithm designSince the image of the light guide plate is collected by using a high-resolution line scan camera,the size of a single picture is 421 MB.The density of the dots at different positions is defferent.In order to improve the detection efficiency and accuracy,the algorithm will guide the light panel image divided into five regions according to the dot density,the first four regions are sparse regions,and the last region is a dense region.For the sparse zone,because the dots have a great influence on the detection of defects,this paper designs a dots separation algorithm.Firstly,the Gaussian operator and the Laplacian operator are used to sharpen the image,and the image is convolved by the point filtering template.The image is divided into two parts:the dot and the background,and then defects will be extracted and determined according to the features in the respective parts.For the dense area,the Gaussian gradient filter function and the digital morphology operation are used to detect the defect according to the defect type existing in the dense area.Finally,through off-line testing and on-line testing,the test proves that the proposed algorithm has a high detection rate for defects such as bright spots,scratches,crushes,and dirt,and the missed detection rate is as low as 0.06%.(4)Software system designThe detection system is jointly developed with the Dalsa camera software library and the Halcon image processing library.Based on the Windows 10 operating system,the MFC framework is used to build the system on the VS2015 development platform,and a light guide panel quality visual inspection system is developed.