Research On Imaging Mechanism And Detection Technology Of Polarizer Appearance Defects

A polymeric polarizer is a crucial part of thin-film transistor(TFT)liquid crystal display(LCD)panels.Its aesthetic defects have a serious impact on the quality of TFT-LCD panels.Currently,the aesthetic defects are mainly detected by manual work;however the manual inspection is really labor-intensive,hard to furtherly improve the speed,and lack of reliability and consistency.Hence,an automated detection technology based on machine vision for such defects with high accuracy and speed is highly important for the relevant industries.Some special and transparent defects can hardly imaging and be characterized under uniform illumination.In order to inspect these special defects,the project team adopts the linear stripe structured-light illumination to realize the defects imaging enhancement,and studied the simulation system of transparent defect imaging based on Trace Pro.On the basis of the above research,the following studies are carried out in this paper:1.Study on the mechanism of the structured-light imaging enhancement.In this paper,the variation of defect imaging with image distance is studied through the simulation and experimental work.When the image distance is smaller,the defect is imaged as a brighter spot in the bright stripe;with the increase of the image distance,the defect brightness decreases gradually,and the brightness is the same as the background(completely unable to image);then gradually the defect becomes a dark spot in the bright stripe with better contrast——considering the imaging contrast and technical feasibility,which is the best working image distance.The experimental results are consistent with the simulation results,which verify the correctness of the defect model and imaging enhancement mechanism: for a point transparent defect,the optical path is equivalent to a micro lens imaging system.In the case of uniform illumination,without aperture it has weak defect imaging;in the case of structured-light,dark and bright stripes are equivalent to the introduction of a one-dimensional aperture,which greatly enhances the imaging of the defect(micro lens).2.In order to solve the problem that some subtle defects are still difficult to be imaged under the condition of structural-light illumination,the better enhancement of contrast is realized by the method of over exposure.The influence of the stripe width of the over exposure method on the contrast of the image is studied experimentally.The results show that the optimal width of the black and white stripe width is in the range of 1.2-1.8mm,and the ratio is better close to 1:1.3.The image processing algorithm based on spatial texture filtering is studied.Two kinds of algorithms,rangefilt and stdfilt,are used to separate the defect(foreground)from the stripe image(background).Both of them have high precision and fast running speed.4.The method of indentation depth measurement based on defect imaging is proposed.The method includes the way to make indentation defect samples,and the measurement of indentation depth by digital microscope.According to the correlation between microscope observation data and defects depth,a nondestructive measurement method of the defect detection based on internal indentation depth is proposed.By using the active optical step-by-step scanning method,about 200 samples were tested experimentally.The inspection accuracies of the structured-light imaging with rangefilt and stdfilt reach up to 100%.For a defect image 1901×1745 pixel size,it takes about 1.6s and 2.6s for rangefilt and stdfilt algorithms respectively.The experimental results show that the proposed method has a good application value and is expected to be used for on-line defect detection of polarizers and its similar film products.