CN-121998934-A - Internal bubble detection method for polaroid production based on image processing

CN121998934ACN 121998934 ACN121998934 ACN 121998934ACN-121998934-A

Abstract

The invention discloses an internal bubble detection method for polaroid production based on image processing, and particularly relates to the technical field of polaroid detection; the invention extracts multidimensional fusion characteristics such as circularity, gray scale contrast and the like by collecting transmitted low-angle light and multi-polarization angle polarization images, combines a defect sample set to judge category belongings, effectively distinguishes internal bubbles from surface dust, scratches and the like, reduces erroneous judgment and missed judgment, and simultaneously, accurately positions bubble positions through processing such as local threshold segmentation, closing operation and the like, thereby laying a reliable foundation for subsequent evaluation and tracing.

Inventors

HAN DESHENG
LI DULIANG
LI QIAOLIANG
WANG KUN
LI XIA

Assignees

云南锦鼎光电科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260122

Claims (9)

1. The method for detecting the internal air bubbles in the production of the polaroid based on image processing is characterized by comprising the following steps of: Collecting a plurality of characteristic images of the same polaroid area as a characteristic image set; The characteristic image set comprises a transmission image, a low-angle light image and a polarized image, wherein the polarized image comprises images acquired under different polarization angles; analyzing the feature image set, extracting multi-dimensional fusion feature vectors of each suspicious region after positioning the suspicious region, identifying internal bubble defects according to the fusion feature vectors, and positioning the position; Based on judging defect types of different suspicious regions, accumulating suspicious region areas belonging to the same defect type, and integrating defect weights set by different defect types to perform weighted fusion, and outputting a defect evaluation coefficient of the polaroid; The method comprises the steps of determining physical position coordinates of a starting point and an ending point of each process section in the length direction of the polaroid, defining the starting point and the ending point of different process sections in a coordinate system, dividing the process influence areas for the positions of internal bubble defects as the process influence areas, determining hidden danger processes based on dividing results, extracting process parameters of the hidden danger processes in the current polaroid corresponding to the production time period, performing association analysis, judging hidden danger degrees of the hidden danger processes, and sending the hidden danger degrees to management staff.
2. The method for detecting internal bubbles in the production of an image processing-based polarizer according to claim 1, wherein: logic to locate the suspicious region in the step; In the transmission image, a local threshold of the image is acquired ; According to the threshold value Generating a binary mask image ; Expressed as: ; representing gray values at coordinates (x, y); For a pair of And performing corrosion-before-expansion closed operation, and then performing connected domain marking, and extracting the circumscribed rectangle of each connected domain as a suspicious region.
3. The method for detecting internal air bubbles in the production of an image processing-based polarizer according to claim 2, wherein: identifying internal bubble defects and positioning the positions; extracting multi-dimensional fusion feature vectors of each suspicious region, wherein the multi-dimensional fusion feature vectors comprise transmission features, low-angle light features and polarization features; for the transmission characteristic, the low-angle light characteristic and the polarization characteristic of each suspicious region, comprehensively processing the fusion characteristic vector of each pre-constructed defect sample set, and determining the category attribution between each suspicious region and each defect sample set; the corresponding type defect with the lowest category attribute is used as the judging defect of the suspicious region; and extracting suspicious regions with the defects judged to belong to the internal bubble defects, and acquiring positions based on pixel information of the transmission image.
4. The method for detecting internal bubbles in the production of an image processing-based polarizer according to claim 3, wherein: calculation logic for the transmission characteristics, the low angle light characteristics, and the polarization characteristics; The transmission features include circularity and gray scale contrast; the gray scale contrast is obtained by calculating the absolute difference between the pixel gray scale average value in the suspicious region and the pixel gray scale average value in the background region; The low-angle light features are contrast ratio of bright and dark fields; Calculating an absolute difference value between a gray average value of a suspicious region in a low-angle light image and a gray average value of a background region, and carrying out ratio calculation by taking the absolute difference value as a numerator and the gray contrast of a transmission image as a denominator to obtain a bright-dark field contrast ratio; the polarization characteristic is curve smoothness; For different polarization angles And calculating the average gray level of the suspicious region under each angle, carrying out normalization processing, fitting the curve, and calculating the root mean square of the second derivative of the fitted curve as the smoothness of the curve.
5. The method for detecting internal bubbles in the production of an image processing-based polarizer according to claim 1, wherein: determining hidden danger technological process in the steps; the process influence area comprises a coating process influence area, a stretching process influence area and a curing process influence area; Counting the number of bubble defect pixels in different process influence areas, converting the actual area based on the image resolution, selecting the process influence area with the highest bubble defect area, and identifying the process type as a hidden danger process; the hidden danger processes comprise a coating hidden danger, a stretching hidden danger and a curing hidden danger, and correspond to a coating process influence area, a stretching process influence area and a curing process influence area respectively.
6. The method for detecting internal air bubbles in the production of an image processing-based polarizer according to claim 5, wherein: the parameters of the association analysis in the step specifically comprise; The technological parameters related to the coating hidden trouble comprise slurry viscosity, scraper pressure and coating speed; the technological parameters related to the stretching hidden trouble comprise stretching partition temperature, stretching multiplying power and stretching speed; the process parameters associated with the curing hidden trouble include curing oven temperature and curing oven wind speed.
7. The method for detecting internal air bubbles in the production of an image processing-based polarizer according to claim 6, wherein: if the hidden danger process is a coating hidden danger in the step, judging the hidden danger degree of the hidden danger process; extracting the slurry viscosity, the scraper pressure and the coating speed of the polaroid at each time point in the production time period of the coating process, and respectively calculating the average value and then outputting the slurry average viscosity, the scraper average pressure and the coating average speed; The average viscosity of the slurry, the average pressure of the scraper and the average coating speed are comprehensively treated by combining the coating process standard, and then the hidden danger coefficient of the coating process is output; if the coating process hidden danger coefficient is higher than the set coating threshold hidden danger coefficient, calculating a difference value between the coating process hidden danger coefficient and the coating threshold hidden danger coefficient, and calculating the ratio by taking the difference value as a numerator and the coating threshold hidden danger coefficient as a denominator to serve as hidden danger degree.
8. The method for detecting internal air bubbles in the production of an image processing-based polarizer according to claim 6, wherein: Judging the hidden danger degree of the hidden danger process if the hidden danger process is a stretching hidden danger in the step; extracting stretching partition temperature of the polaroid in a stretching time period, wherein the stretching partition comprises an inlet, a middle section and an outlet; respectively carrying out average value calculation on the inlet partition temperature, the middle partition temperature and the outlet partition temperature at each time point in the stretching time period, and outputting inlet average temperature, middle average temperature and outlet average temperature; Combining the interval temperature difference with the stretching multiplying power and the stretching speed in the stretching time period, and outputting the hidden danger coefficient of the stretching process after comprehensively processing by combining the stretching process standard; If the potential risk coefficient of the stretching process is higher than the potential risk coefficient of the stretching threshold, calculating a difference value between the potential risk coefficient of the stretching process and the potential risk coefficient of the stretching threshold, taking the difference value as a numerator, and calculating a ratio by taking the potential risk coefficient of the stretching threshold as a denominator as the potential risk degree.
9. The method for detecting internal air bubbles in the production of an image processing-based polarizer according to claim 6, wherein: judging the hidden danger degree of the hidden danger process if the hidden danger process is a curing hidden danger in the step; Extracting the temperature and the wind speed of the polaroid at each time point in the curing time period of the curing furnace, respectively calculating the average value, and then outputting the average temperature and the average wind speed in the furnace; the average temperature in the furnace and the average wind speed in the furnace are comprehensively treated by combining the furnace curing standard, and then the hidden danger coefficient of the curing process is output; if the hidden danger coefficient of the curing process is higher than the hidden danger coefficient of the curing threshold, calculating the difference between the hidden danger coefficient of the curing process and the hidden danger coefficient of the curing threshold, taking the difference as a numerator, and calculating the ratio by taking the hidden danger coefficient of the curing threshold as a denominator as the hidden danger degree.

Description

Internal bubble detection method for polaroid production based on image processing Technical Field The invention relates to the technical field of polaroid detection, in particular to an internal bubble detection method for polaroid production based on image processing. Background The polaroid is used as a core component of a liquid crystal display panel, the optical performance of the polaroid directly determines the display effect and quality stability of a terminal display product, internal bubble defects are extremely easy to generate due to temperature control deviation and other factors in the production process, the defects can seriously influence the optical performance of the polaroid, the problems of bright spots, dark spots and the like of the terminal display product are caused, and the product qualification rate is reduced. Currently, an online detection method based on image processing gradually replaces manual visual inspection, but the following technical bottlenecks still exist: on one hand, the detection precision of a single type of image is limited, so that the defects of internal bubbles, surface dust, scratches, crystal points and the like are difficult to distinguish effectively, and misjudgment and missed judgment are easy to occur; On the other hand, the detection process can only realize the identification and the positioning of the defects, the correlation between the defects and production process parameters cannot be established, the generation sources of the bubble defects are difficult to trace, hidden danger processes in the production process cannot be found and adjusted in time, the defect occurrence rate cannot be reduced fundamentally, and the detection intelligent degree is low. For this reason, an internal bubble detection method for polarizer production based on image processing is proposed. Disclosure of Invention In order to overcome the above-described drawbacks of the prior art, embodiments of the present invention provide an internal bubble detection method for polarizer production based on image processing. In order to achieve the above purpose, the present invention provides the following technical solutions: an internal bubble detection method for polarizer production based on image processing, comprising: Collecting a plurality of characteristic images of the same polaroid area as a characteristic image set; The characteristic image set comprises a transmission image, a low-angle light image and a polarized image, wherein the polarized image comprises images acquired under different polarization angles; analyzing the feature image set, extracting multi-dimensional fusion feature vectors of each suspicious region after positioning the suspicious region, identifying internal bubble defects according to the fusion feature vectors, and positioning the position; Based on judging defect types of different suspicious regions, accumulating suspicious region areas belonging to the same defect type, and integrating defect weights set by different defect types to perform weighted fusion, and outputting a defect evaluation coefficient of the polaroid; The method comprises the steps of determining physical position coordinates of a starting point and an ending point of each process section in the length direction of the polaroid, defining the starting point and the ending point of different process sections in a coordinate system, dividing the process influence areas for the positions of internal bubble defects as the process influence areas, determining hidden danger processes based on dividing results, extracting process parameters of the hidden danger processes in the current polaroid corresponding to the production time period, performing association analysis, judging hidden danger degrees of the hidden danger processes, and sending the hidden danger degrees to management staff. In particular, the method comprises the steps of,Logic to locate the suspicious region in the step; In the transmission image, a local threshold of the image is acquired ; According to the threshold valueGenerating a binary mask image; Expressed as:; representing gray values at coordinates (x, y); For a pair of And performing corrosion-before-expansion closed operation, and then performing connected domain marking, and extracting the circumscribed rectangle of each connected domain as a suspicious region. In particular, the method comprises the steps of,Identifying internal bubble defects and positioning the positions; extracting multi-dimensional fusion feature vectors of each suspicious region, wherein the multi-dimensional fusion feature vectors comprise transmission features, low-angle light features and polarization features; for the transmission characteristic, the low-angle light characteristic and the polarization characteristic of each suspicious region, comprehensively processing the fusion characteristic vector of each pre-constructed defect sample set, and determining the category a