OLED display quality inspection method based on machine vision

Abstract

The present invention relates to the field of image processing technology, and more specifically, the present invention relates to a method for detecting the quality of an OLED display screen based on machine vision, comprising: obtaining a grayscale image and a color image of an OLED display screen surface image; calculating the difference between the grayscale value of each pixel and the overall grayscale mean, and calculating the degree of compensation of the pixel point. The greater the difference, the more obvious the brightness non-uniformity, and the greater the degree of compensation of the pixel point. Then, different weights are assigned to each pixel point according to the brightness non-uniformity. The present invention calculates the final grayscale value of the pixel point through the degree of compensation of the pixel point and the correction coefficient, and the degree of compensation is determined according to the difference of the neighborhood gradient value and the overall grayscale difference in the grayscale image, and the correction coefficient is determined according to the color difference of the pixel point in the color image. Mura defect areas with insignificant brightness difference but obvious color distortion can be identified, thereby improving the accuracy of quality detection of the OLED display screen.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and more specifically, to a method for detecting the quality of an OLED display screen based on machine vision. Background Art

[0002] OLED (Organic Light-Emitting Diode) display is a new type of display technology, which consists of multiple layers of organic materials, including anode, cathode, hole transport layer, electron transport layer and light-emitting layer. Its working principle is to inject electrons and holes into the organic material layer under the action of electric field, and compound them in the light-emitting layer to generate light. OLED transparent display can be used in medical imaging equipment. Since the organic materials in OLED display may be unevenly distributed during the manufacturing process, Mura defects may appear on OLED display, so it is necessary to conduct quality inspection on OLED display.

[0003] In the related art, for example, a Chinese patent application document with publication number CN110503923A discloses a method and device for compensating for mura defects in a display screen, wherein a characteristic image of mura is extracted by filtering the captured image to obtain a mura defect area map; brightness bias compensation: sampling the characteristic image of mura to obtain a brightness bias compensation map; feedback calibration: reading a brightness bias compensation value according to the brightness bias compensation map, and controlling the drive to reduce the LED brightness of the corresponding mura area according to the brightness bias compensation value.

[0004] At present, when using Demura technology to identify Mura defects in OLED displays, it is only based on brightness differences, which easily ignores the color difference between the Mura defect area and the normal area. Therefore, some Mura defects with obvious color distortion but insignificant brightness difference may be missed, resulting in a lower accuracy rate for quality inspection of OLED displays. Summary of the invention

[0005] The present invention provides an OLED display quality inspection method based on machine vision, aiming to solve the problem that in the related art, identification is performed only based on brightness difference, which easily ignores the color difference between the Mura defect area and the normal area, and may miss some Mura defects with obvious color distortion but insignificant brightness difference.

[0006] The present invention provides a method for detecting the quality of an OLED display screen based on machine vision, comprising: obtaining a grayscale image and a color image of the surface image of the OLED display screen; calculating the compensation degree of the pixel points in the grayscale image, and determining the compensation degree of the pixel points; The degree of compensation for: ; Pixel The gradient value With pixels The gradient value difference, Pixel The gradient direction With pixels The gradient direction difference, Pixel Pixels in the eight-neighborhood, is the normalization function, Pixel The weight reflects the difference between the initial grayscale value of the pixel and the overall grayscale mean of the grayscale image; according to the color difference between each pixel in the color image and the pixels in its neighborhood, the local fluctuation degree corresponding to the pixel is calculated, and based on the size of the local fluctuation degree, the pixel is divided into two categories, the first category is the defect edge area, and the second category is the inside of the defect area and the normal area, and the difference between the yellow-blue hue of the pixel and the minimum yellow-blue hue is used to divide the inside of the defect area and the normal area, and the corresponding correction coefficient is set based on the pixels in different areas; the initial grayscale value of the pixel is corrected using the correction coefficient and compensation degree corresponding to the pixel to obtain the final grayscale value, and the quality of the OLED display screen is detected based on the average value of the sum of the difference between the initial grayscale value and the final grayscale value of each pixel. According to the overall grayscale value difference of the pixel and the neighborhood gradient value difference, the compensation degree of each pixel is obtained, the correction coefficient of the pixel is calculated according to the color difference between the pixels, and the initial grayscale value of the pixel is compensated based on the compensation degree and correction coefficient of each pixel, so that the final grayscale value of each pixel can be accurately calculated.

[0007] Furthermore, the quality of the OLED display screen is tested, including: normalizing the average value; if the normalized value of the average value is greater than a first threshold, the quality of the OLED display screen is unqualified and the screen needs to be returned to the factory for repair; if the normalized value of the average value is less than or equal to the first threshold, the quality of the OLED display screen is qualified. Based on the sum of the differences between the initial grayscale value and the final grayscale value of each pixel, the quality of the OLED display screen can be accurately judged.

[0008] Furthermore, obtaining the final grayscale value includes: multiplying the sum of the correction coefficient and the compensation degree corresponding to the pixel point and the initial grayscale value as the final grayscale value of the pixel point.

[0009] Furthermore, in response to the pixel Located in the normal area, the pixel The final gray value for: , Pixel The initial gray value, Pixel The degree of compensation; in response to the pixel point Located at the edge of the defect, the pixel The final gray value for: , Pixel The initial gray value, Pixel The degree of compensation, Pixels in the defect edge area Correction coefficient of the pixel point Located inside the defect area, the pixel The final gray value : , Pixel The initial gray value, Pixel The degree of compensation, is the pixel inside the defect area The correction factor of is a natural function An exponential function with base .

[0010] Furthermore, corresponding correction coefficients are set based on the pixels in different regions, including: setting the correction coefficients of the pixels in the defect edge region to ; Set the correction coefficient of the pixel points inside the defect area to ; Set the correction coefficient of the pixel points in the normal area to ;in, ,and , , is a natural constant.

[0011] Furthermore, obtaining a color image of the surface image of the OLED display screen includes: performing color correction processing on the surface image using a color correction matrix, and converting the processed surface image from the RGB space to the Lab space to obtain a color image. The use of a color correction matrix and a processing method for converting an image from the RGB space to the Lab space can effectively improve the color accuracy and consistency of the image and reduce color errors caused by device differences and lighting changes.

[0012] Furthermore, the calculation formula for the local fluctuation degree corresponding to the pixel point is: ; In the formula, Indicates The local fluctuation degree corresponding to the pixel point, , , Respectively represent the The brightness, red-green and yellow-blue of the pixel in Lab space. , , Respectively represent the The brightness, red-green and yellow-blue of the pixel in Lab space. Indicates The number of pixels in the eight-neighborhood of a pixel.

[0013] Furthermore, based on the magnitude of the local fluctuation degree, the pixels are divided into two categories, including: obtaining the third quartile of the local fluctuation degree corresponding to all the pixels; dividing the pixels into two categories according to the third quartile; if the local fluctuation degree corresponding to the pixel is greater than the third quartile, it is divided into the first category; if the local fluctuation degree corresponding to the pixel is less than or equal to the third quartile, it is divided into the second category. According to the characteristic that the local fluctuation degree is larger, it often corresponds to the edge area of ​​the mura defect, the pixels belonging to the edge area of ​​the mura defect are divided.

[0014] Furthermore, the internal defective area and the normal area are divided, including: if the absolute value of the difference between the yellow-blue hue of the pixel and the minimum yellow-blue hue is less than the division threshold, the pixel belongs to the normal area; if the absolute value of the difference between the yellow-blue hue of the pixel and the minimum yellow-blue hue is greater than or equal to the division threshold, the pixel belongs to the internal defective area. Further, the overall grayscale mean of the grayscale image is obtained, including: dividing the grayscale image into areas of equal size; calculating the entropy value of the grayscale values ​​of all pixels in each area, and taking the average value of the grayscale values ​​of all pixels in the sub-area corresponding to the minimum entropy value as the overall grayscale mean of the grayscale image. Determine the division area of ​​the pixel points according to the color difference of the pixel points in the color image, so as to facilitate the subsequent calculation of the correction coefficients of the pixel points in each area.

[0015] Beneficial effect: The final grayscale value of the pixel is calculated by the compensation degree and correction coefficient of the pixel. The compensation degree is determined according to the difference of the neighborhood gradient value (dramatic brightness change) and the overall grayscale difference (uneven brightness change) in the grayscale image. The correction coefficient is determined according to the color difference of the pixel in the color image. Mura defect areas with insignificant brightness difference but obvious color distortion can be identified, thereby improving the accuracy of quality inspection of OLED displays. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The following detailed description is read with reference to the accompanying drawings, which illustrate several embodiments of the present invention in an exemplary and non-limiting manner, and in which like or corresponding reference numerals represent like or corresponding parts, wherein:

[0017] Figure 1 FIG. 4 is a flow chart schematically illustrating the calculation of the final grayscale value of each pixel according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0019] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0020] In one embodiment, the present invention evaluates whether the quality of the OLED display screen meets the standard based on the total compensation degree of all pixels in the OLED display screen image, and returns the OLED display screen that does not meet the standard to the factory for repair. Specifically, the total compensation degree of a pixel can be obtained based on the compensation degree of the pixel and the correction coefficient. The degree of compensation is determined based on the difference in neighborhood gradient values ​​(dramatic brightness changes) and the overall grayscale difference (uneven brightness changes) in the grayscale image. The correction coefficient is determined based on the color difference of the pixel in the color image. In this way, Mura defect areas with insignificant brightness differences but obvious color distortion can be identified, thereby improving the accuracy of quality detection of OLED displays. The specific implementation steps are as follows.

[0021] It should be noted that Mura defect refers to the phenomenon of uneven brightness or color on the display panel, which has the characteristics of randomness, blurred edges, irregularity, etc. In the existing method, Demura technology is used to identify and eliminate the Mura defect area in the OLED display.

[0022] like Figure 1 As shown, S101: obtaining a grayscale image and a color image of the surface image of the OLED display screen.

[0023] In one embodiment, the process of obtaining the surface image of the OLED display screen includes: placing the produced OLED display screen in the center of a horizontal desktop, then installing a high-resolution industrial camera at an appropriate distance above the center of the horizontal desktop, and then photographing the OLED display screen with the high-resolution industrial camera to obtain the surface image of the OLED display screen. Among them, the light should be kept stable for the acquisition environment to avoid the influence of external light sources on the image quality. Obtaining the grayscale image of the surface image of the OLED display screen includes: bilateral filtering the three channels of R, G, and B of the surface image respectively to eliminate the interference of random noise points generated by the heating of the electronic components of the camera, and retaining the edges while removing the noise. The surface image after noise removal is converted into a grayscale image, and the gradient value of each pixel in the grayscale image is obtained using the Sobel operator for subsequent use. Obtaining the color image of the surface image of the OLED display screen includes: using a color correction matrix to perform color correction processing on the surface image, and converting the processed surface image from the RGB space to the Lab space to obtain a color image. It should be noted that in the above process, bilateral filtering, Sobel operator, conversion grayscale image and color correction matrix (CCM) are all prior art and will not be repeated here.

[0024] It should be noted that the Lab color space is a uniform color space based on human eye perception proposed by the International Commission on Illumination (CIE). It consists of three coordinate axes: Indicates brightness, which reflects the lightness or darkness of the color, and its value range is from 0 (black) to 100 (white). Indicates the red-green color, reflecting the position of the color on the red-green axis. Positive values ​​represent red, and negative values ​​represent green. Represents yellow-blue chromaticity, reflecting the position of the color on the yellow-blue axis. Positive values ​​represent yellow, and negative values ​​represent yellow-blue chromaticity. According to the distribution on the red-green axis in the Lab color space, all pixels on the positive half axis of the red-green axis are removed, that is, the pixels in the red area of ​​the color image are removed. This is to reduce the interference of red pixels on subsequent mura defect detection. The reason is that yellow-blue OLED materials usually have lower luminous efficiency, which means that under the same current, the luminous intensity of yellow-blue pixels is not as high as that of red pixels, and is not uniform and stable. Red OLED materials usually have higher luminous efficiency and are more uniform and stable. Yellow-blue pixels are more prone to aging, brightness attenuation and unevenness after long-term use. These factors make mura defects more likely to occur in the yellow-blue chromaticity area.

[0025] S102: Calculate the compensation degree of each pixel.

[0026] In one embodiment, the degree of compensation of each pixel is the degree to which the grayscale value needs to be compensated. The degree of compensation for pixels belonging to the mura defect area is large, and the degree of compensation for pixels in the normal area is small. Specifically, the degree of compensation of the pixel is calculated by calculating the difference between the grayscale value of each pixel and the overall grayscale mean. The greater the difference, the more obvious the brightness non-uniformity, and the greater the degree of compensation of the pixel. Then, different weights are assigned to each pixel according to the brightness non-uniformity, and then a weighted neighborhood is designed in combination with the weight of each pixel, and the gradient value difference and gradient direction difference of each pixel and the pixel in its weighted neighborhood are calculated to characterize the intensity of the brightness change. The greater the difference in gradient value, the more intense the brightness change, and the pixel may have defects, and the greater the degree of compensation of the pixel.

[0027] In one embodiment, a calculation formula for calculating the compensation degree of each pixel is provided, and the calculation formula is: ; Pixel The degree of compensation, Pixel The gradient value With pixels The gradient value The difference Pixel The gradient direction With pixels The gradient direction The difference Pixel The weight of Pixel The number of pixels in the eight-neighborhood of is a normalization function; the weight is positively correlated with the difference between the initial grayscale value of the pixel and the overall grayscale mean of the grayscale image. The larger the weight of The greater the compensation, the greater the

[0028] In one embodiment, the grayscale image is divided into regions of equal size, and then the weight of each pixel is calculated using the following formula: .in, Pixel The weight of Pixel The gray value of Pixel The regional grayscale mean, Represents pixel Gray value With pixels The regional grayscale mean The difference between Represents pixel The regional grayscale mean The overall grayscale mean The difference between Represents normalization processing, which can be done by using maximum and minimum value normalization, etc. Among them, to obtain the overall grayscale mean of the grayscale image , including: calculating the entropy value of the grayscale values ​​of all pixels in each area, and taking the average value of the grayscale values ​​of all pixels in the area corresponding to the minimum entropy value as the overall grayscale mean of the grayscale image.

[0029] S103: Obtaining a correction coefficient for each pixel.

[0030] In one embodiment, only by brightness difference to identify defects, it is easy to ignore the color difference between the Mura defect area and the normal area, and miss some Mura defects with obvious color distortion. Therefore, according to the color difference of the pixels in the color image, the correction coefficient of each pixel is calculated, and then the Mura defect area with insignificant brightness difference but obvious color distortion is identified.

[0031] In one embodiment, to obtain the correction coefficient of each pixel point, the local fluctuation degree corresponding to each pixel point must first be calculated, and based on the magnitude of the local fluctuation degree, the pixel points are divided into two categories, the first category is the defect edge area, and the second category is the inside of the defect area and the normal area. Specifically, the third quartile of the local fluctuation degree corresponding to all the pixels is obtained; the pixels are divided into two categories according to the third quartile; if the local fluctuation degree corresponding to the pixel point is greater than the third quartile, it is divided into the first category; if the local fluctuation degree corresponding to the pixel point is less than or equal to the third quartile, it is divided into the second category. Among them, the third quartile is to divide the data into four parts from large to small, and the third quartile is the 75th position in the sorted data set.

[0032] In one embodiment, the calculation formula for the local fluctuation degree corresponding to each pixel point is: ; It's a pixel The degree of local fluctuation corresponding to all pixels in its eight neighborhoods. , , Respectively represent the pixel points in the color image Brightness, red-green, and yellow-blue in Lab space, , , Respectively represent the The brightness, red-green and yellow-blue of the pixel in Lab space. Represents pixel The number of pixels in the eight-neighborhood.

[0033] In one embodiment, two embodiments are provided in the present invention for dividing the pixels inside the defective area and the normal area, setting corresponding correction coefficients based on the pixels in different areas, and finally calculating the final grayscale value of the pixels. The two embodiments are as follows.

[0034] Embodiment 1

[0035] In one embodiment, the difference between the yellow-blue luminance of the pixel and the minimum yellow-blue luminance is used to divide the pixels inside the mura defect area and the normal area again. Specifically, a division threshold is set, the absolute value of the difference between the yellow-blue luminance of the pixel and the minimum yellow-blue luminance is calculated, and the absolute value of the difference is normalized. If the normalized value is less than the division threshold, the pixel belongs to the normal area; after obtaining the absolute value of the difference between the yellow-blue luminance of the pixel and the minimum yellow-blue luminance, the absolute value of the difference is normalized. If the normalized value is greater than or equal to the division threshold, the pixel belongs to the mura defect area, where the empirical value of the division threshold is 0.5. The reason is that the pixels inside the mura defect area have the lightest yellow-blue luminance than the pixels in the normal area, and the lighter the yellow-blue luminance (the yellow-blue luminance is a negative value), the closer the yellow-blue luminance is to 0, and the greater the difference with the minimum yellow-blue luminance should be, which is more in line with the characteristics of the mura defect area. Finally, the corresponding correction coefficient is set based on the pixels in different areas.

[0036] In one embodiment, when setting corresponding correction coefficients based on pixels in different regions, the following reasons should be considered: when a pixel belongs to a normal region, there is a small fluctuation in grayscale and color angle, that is, a slight compensation is required to prevent image distortion caused by excessive compensation. Therefore, the correction coefficient of the pixel in the normal region is set to When the pixel point is inside the mura defect, color angle compensation is required. Since the gray value inside the mura defect area is larger than that in the normal area, it is necessary to compensate and reduce its gray value. The correction coefficient of the pixel point inside the defect area is set to , and reduce its gray value to a certain extent. When the pixel point belongs to the edge area of ​​mura defect, color angle compensation is required. Since the gray value of the edge area of ​​mura defect is smaller than that of the normal area, it is necessary to compensate and increase its gray value. Therefore, the correction coefficient of the pixel point in the edge area of ​​the defect is set to .in, ,and , , is a natural constant. In addition, when the pixel point belongs to the red area, its correction coefficient is 0.

[0037] S104: Correcting the initial grayscale value of the pixel point using the correction coefficient and compensation degree corresponding to the pixel point to obtain a final grayscale value.

[0038] In one embodiment, the product of the sum of the correction coefficient and the compensation degree corresponding to the pixel point and the initial grayscale value is used as the final grayscale value of the pixel point. , pixel The initial gray value is , pixel The degree of compensation is , pixel The correction factor is , then the pixel The final grayscale value is equal to , control correction factor The final grayscale value is controlled between (1,2) times the initial grayscale value before compensation. For the pixel points inside the mura defect area , pixel The initial gray value is , pixel The degree of compensation is , pixel The correction factor is , then the pixel The final grayscale value is equal to , control correction factor The final grayscale value is controlled between (0,1) times the initial grayscale value before compensation. , pixel The initial gray value is , pixel The degree of compensation is , pixel The correction factor is , then the pixel The final grayscale value is equal to , correction factor The empirical value is 0.1 or 0.2, etc., and only a slight compensation is made to the initial grayscale value.

[0039] It should be noted that in Example 1, only the difference between the yellow-blue hue of the pixel and the minimum yellow-blue hue is used to divide the pixels inside the defective area and the normal area. In order to improve the accuracy of the division and the accuracy of the final calculation of the correction coefficient, the change in pixel brightness is also introduced, and the characteristics of pixel brightness and pixel yellow-blue hue are combined to divide the pixels inside the defective area and the normal area, and the final grayscale value of each pixel is calculated, which improves the accuracy of the classification and calculation results. The reason is that the brightness of the pixels inside the mura defect is the largest, while the brightness of the pixels in the normal area is less than the brightness of the pixels inside the mura defect, and greater than the brightness of the pixels in the edge area of ​​the mura defect.

[0040] Embodiment 2

[0041] In one embodiment, the correction coefficients corresponding to the pixels in different areas may be calculated by the following method, and the original grayscale value of the pixel may be corrected again using the correction coefficient and compensation degree of each pixel to obtain the final grayscale value.

[0042] In one embodiment, if the pixel point belongs to the pixel point in the edge area of ​​the mura defect in the first category, the calculation formula of the correction coefficient of the pixel point is: . The defect edge area Correction coefficient corresponding to the pixel. Represents pixel points Brightness , yellow-blue degree The minimum value of brightness and the minimum value of yellow-blue The difference between For the The local fluctuation degree corresponding to the pixel point, is an exponential function with the natural constant e as the base. Therefore, for the pixel points inside the mura defect area , pixel The final gray value The calculation formula is: , Pixel The initial gray value, Pixel The degree of compensation, Pixels in the edge area of ​​mura defect The correction coefficient is used to control the gray value after compensation to be between (1,2) times of the gray value before compensation, thus improving the gray value to a certain extent.

[0043] In one embodiment, if the pixel point belongs to the second type of pixel point inside the mura defect area or the normal area, then it is necessary to determine the calculation formula according to the conditions satisfied by the pixel point. , When The more likely a pixel is to be inside the mura defect area, the calculation formula for the correction coefficient of the pixel is: In the formula, is the pixel inside the defect area The corresponding correction factor is, Represents pixel points Brightness , yellow-blue degree The minimum value of brightness and the minimum value of yellow-blue The difference between Pixel The corresponding local fluctuation degree, is a normalized function. The reason is that the pixel brightness inside the mura defect area is the highest, the corresponding yellow-blue color is the lightest, the local fluctuation is small, and it is closer to the normal area. That is to say, when the The brightness of the pixel , yellow-blue degree The minimum value of brightness and the minimum value of yellow-blue The larger the normalized value of the difference between The smaller the local fluctuation degree corresponding to the pixel point, the The more likely the pixel is to be inside the mura defect area, the Pixel correction factor The bigger it is.

[0044] For pixels inside the mura defect area , pixel The final gray value The calculation formula is: , Pixel The initial gray value, Pixel The degree of compensation, Pixels inside the mura defect area The correction factor of For natural function An exponential function with as the base. Thus, the gray value after compensation is controlled between (0,1) times the gray value before compensation, and its gray value is reduced to a certain extent.

[0045] In one embodiment, when a pixel satisfies , When the conditions are met, explain The more likely a pixel point is to belong to the normal area, the calculation formula for the correction coefficient corresponding to the pixel point is: ; In the formula, Normal area pixels The corresponding correction factor is, Represents pixel points Brightness , yellow-blue degree The minimum value of brightness and the minimum value of yellow-blue The difference between Pixel The corresponding local fluctuation degree, is the normalization function. That is, when The brightness of the pixel , yellow-blue degree The minimum value of brightness and the minimum value of yellow-blue The smaller the normalized value of the difference between and the smaller the local fluctuation, the pixel is considered The more likely the pixel is to be in the normal area, the smaller the correction coefficient is. It should be noted that since there is a certain local gray value fluctuation in the normal area, only a slight compensation gray value is required. Therefore, for the pixel in the normal area , pixel The final gray value The calculation formula is: , Pixel The initial gray value, Pixel At this point, the final grayscale value of each pixel can be obtained.

[0046] S105: Detecting the quality of the OLED display screen based on the average value of the sum of the differences between the initial grayscale value and the final grayscale value of each pixel.

[0047] Specifically, the average value is normalized; if the normalized value of the average value is greater than a first threshold, the quality of the OLED display screen is unqualified and is returned to the factory for repair; if the normalized value of the average value is less than or equal to the first threshold, the quality of the OLED display screen is qualified.

[0048] In another embodiment, the quality test result of the OLED display screen may also be calculated using the following formula: ; In the formula, This is the quality test result of OLED display. Represents pixel The difference between the initial grayscale value and the final grayscale value, N represents the number of pixels, is an exponential function with the natural constant e as the base. When the value is greater than the preset threshold, it can be considered that the quality inspection result of the OLED display screen is good, otherwise the OLED display screen needs to be returned to the factory for repair. Among them, the empirical value of the preset threshold is 0.8. In other embodiments, the empirical value of the preset threshold can be 0.85 or 0.7, etc., which can be adjusted according to the specific implementation situation. At this point, the quality inspection of the OLED display screen is completed.

[0049] Through the above steps, the total compensation degree of a pixel can be obtained based on the compensation degree of the pixel and the correction coefficient, and finally the final grayscale value of each pixel is obtained. The compensation degree is determined according to the difference of the neighborhood gradient value in the grayscale image (dramatic brightness change) and the overall grayscale difference (uneven brightness change), and the correction coefficient is determined according to the color difference of the pixel in the color image. In this way, Mura defect areas with insignificant brightness difference but obvious color distortion can be identified, thereby improving the accuracy of quality inspection of OLED displays.

[0050] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0051] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0052] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent application. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for detecting the quality of an OLED display screen based on machine vision, characterized in that: include: Obtaining grayscale images and color images of the surface image of the OLED display screen; Calculate the compensation degree of pixels in the grayscale image. The degree of compensation for: ; Pixel The gradient value With pixels The gradient value difference, Pixel The gradient direction With pixels The gradient direction difference, Pixel Pixels in the eight-neighborhood, is the normalization function, Pixel The weight reflects the difference between the initial gray value of the pixel and the overall gray mean of the gray image; According to the color difference between each pixel in the color image and the pixels in its neighborhood, the local fluctuation degree corresponding to the pixel is calculated, and based on the magnitude of the local fluctuation degree, the pixel points are divided into two categories, the first category is the defect edge area, and the second category is the inside of the defect area and the normal area. The difference between the yellow-blue hue of the pixel point and the minimum value of the yellow-blue hue is used to divide the inside of the defect area and the normal area, and the correction coefficient of the pixel point in the defect edge area is set to ; The correction coefficient of the pixel points inside the defect area is set to ; The correction coefficient of the pixel points in the normal area is set to ;in, ,and , , is a natural constant; The product of the sum of the correction coefficient and compensation degree corresponding to the pixel and the initial grayscale value is taken as the final grayscale value of the pixel, and the quality of the OLED display is tested based on the average value of the sum of the differences between the initial grayscale value and the final grayscale value of each pixel.

2. The OLED display screen quality detection method based on machine vision according to claim 1, characterized in that: Test the quality of OLED display screens, including: Normalize the average value of the sum of the differences between the initial grayscale value and the final grayscale value of each pixel; If the normalized value of the average value is greater than the first threshold value, the quality of the OLED display screen is unqualified and the display screen is returned to the factory for repair; If the normalized value of the average value is less than or equal to the first threshold, the quality of the OLED display screen is qualified.

3. The OLED display screen quality detection method based on machine vision according to claim 1, characterized in that: Get the final grayscale value, including: Responding to the pixel Located in the normal area, the pixel The final gray value for: , Pixel The initial gray value, Pixel the degree of compensation; Responding to the pixel Located at the edge of the defect, the pixel The final gray value for: , Pixel The initial gray value, Pixel The degree of compensation, Pixels in the defect edge area Correction factor of Responding to the pixel Located inside the defect area, the pixel The final gray value : , Pixel The initial gray value, Pixel The degree of compensation, is the pixel inside the defect area The correction factor of is a natural function An exponential function with base .

4. The OLED display screen quality detection method based on machine vision according to claim 1, characterized in that: Get color images of the OLED display surface image, including: The surface image is color corrected using a color correction matrix, and the processed surface image is converted from RGB space to Lab space to obtain a color image.

5. The OLED display screen quality detection method based on machine vision according to claim 4 is characterized in that: The calculation formula for the local fluctuation degree corresponding to the pixel point is: ； In the formula, Represents pixel The corresponding local fluctuation degree, , , Respectively represent the pixel points in the color image Brightness, red-green, and yellow-blue in Lab space, , , Respectively represent the The brightness, red-green and yellow-blue of the pixel in Lab space. Indicates The number of pixels in the eight-neighborhood of a pixel.

6. The OLED display screen quality detection method based on machine vision according to claim 1, characterized in that: Based on the magnitude of local fluctuation, pixels are divided into two categories, including: Obtain the third quartile of the local fluctuation degree corresponding to all the pixel points; Dividing the pixels into two categories according to the third quartile; If the local fluctuation degree corresponding to the pixel point is greater than the third quartile, it is classified into the first category; If the local fluctuation degree corresponding to the pixel point is less than or equal to the third quartile, it is classified into the second category.

7. The OLED display screen quality detection method based on machine vision according to claim 1, characterized in that: The internal and normal areas of the defect area are divided into: If the absolute value of the difference between the yellow-blue hue of the pixel and the minimum yellow-blue hue is less than the division threshold, the pixel belongs to the normal area; If the absolute value of the difference between the yellow-blue hue of the pixel and the minimum yellow-blue hue is greater than or equal to the division threshold, the pixel belongs to the defect area.

8. The method for detecting the quality of an OLED display screen based on machine vision according to any one of claims 1 to 7, characterized in that: Get the overall grayscale mean of the grayscale image, including: Dividing the grayscale image into regions of equal size; The entropy value of the grayscale values ​​of all pixels in each area is calculated, and the average grayscale value of all pixels in the sub-area corresponding to the minimum entropy value is taken as the overall grayscale mean of the grayscale image.

Images