Method and system for automatically correcting brightness photographing data of curved surface area of OLED panel

Abstract

The invention provides a method and a system for automatically correcting brightness photographing data of a curved surface area of an OLED (Organic Light Emitting Diode) panel. The method mainly comprises the steps that an industrial camera shoots brightness data of different gray scales of the curved-surface OLED display panel, shot pictures are synchronously processed to obtain brightness processing data, the extracted brightness data are automatically subjected to curved-surface area and non-curved-surface area recognition and division, a recognition result generates a mask through a miscellaneous point removal method, and the mask is used as a mask for the curved-surface OLED display panel. And determining the curved surface region of which the brightness data needs to be corrected through the mask, and finally determining the correction target brightness in the connected region based on the Mura form distribution, and correcting the data of the curved surface region.

Description

Technical Field

[0001] The present invention relates to an image segmentation algorithm, belongs to the field of image processing, and particularly relates to a method and system for automatically correcting brightness photographic data of a curved surface area of ​​an OLED panel. Background Art

[0002] As a major technological highlight in the field of consumer electronics in recent years, flexible OLED panels have broken the physical form limitations of traditional rigid screens. Flexible screens allow the screen to bend, fold or curl to a certain extent. Flexible screen mobile phones on the market have developed from the initial single-sided bending to double-sided bending, and then to a fully foldable design, providing users with a new interactive experience. With the change of the flexible screen form, the Mura problem of the OLED panel itself still exists, and the brightness data photography process in the Demura process of the traditional rigid screen is not well applicable. The brightness data extraction technology at non-curved positions is relatively mature at present, with high accuracy, and can provide reliable data support for key performance indicators such as the overall brightness uniformity of the screen. However, due to the complex and changeable physical form of the curved position, the traditional plane measurement method may not accurately reflect its true brightness information, including factors such as light refraction and reflection caused by curvature changes, which may affect the effective acquisition of brightness data. How to achieve high-precision brightness data extraction at the curved position requires a suitable processing algorithm to truly restore and quantify the brightness performance of the OLED panel on the curved part. Summary of the invention

[0003] The present invention provides a method for automatically correcting the brightness photographing data of the curved surface area of ​​an OLED panel:

[0004] Step 1: The industrial camera captures different grayscale brightness data of the curved OLED display panel;

[0005] Step 2: Brightness data extraction;

[0006] Step 3: Input the data to be corrected;

[0007] Step 4: Correct area identification;

[0008] Step 5: Determine whether the area recognition is normal. If it is normal, proceed to the next step; if it is not normal, the parameters are adaptively adjusted until the area recognition is successful, and then proceed to the next step;

[0009] Step 6: Interference removal;

[0010] Step 7: Determine whether the interference removal is successful. If successful, proceed to the next step; if unsuccessful, perform parameter adaptive adjustment until the interference removal is successful.

[0011] Step 8, data correction: determine whether the upper and lower surfaces need to be processed, if yes, process the upper and lower surfaces, if not, determine whether the left and right surfaces need to be processed, if yes, process the left and right surfaces, if not, proceed directly to the next step;

[0012] Step 9: Output the corrected data.

[0013] Furthermore, step one includes:

[0014] Step 1.1: Data preprocessing:

[0015] Convert the captured OLED curved panel brightness data into grayscale space;

[0016] Step 1.2, define the window size:

[0017] For each pixel in the image, a neighborhood window of size w×w is selected, where w is determined based on the characteristics of the actual image, the noise level, and the desired segmentation effect;

[0018] Step 1.3, calculate the local threshold:

[0019] For each pixel p in the window, calculate the statistical characteristics (mean) of the grayscale values ​​of all pixels in the window, denoted as T p ; Use the average grayscale in the neighborhood as the threshold basis, and then add a constant C;

[0020] Among them, for pixel S p , its adaptive threshold expression is:

[0021] T p =mean(I neighborhoo )+C

[0022] Step 1.4, binary decision:

[0023] For pixel S p , according to the calculated local threshold T p Binarization:

[0024] If S p >T p , then set pixel p to white (usually 255 for an 8-bit grayscale image);

[0025] If S p <T p , then set pixel p to black (usually 0 for an 8-bit grayscale image);

[0026] Step 1.5, result output:

[0027] Repeat the above steps until all pixels in the image are traversed, and the resulting binary image is segmented by adaptive thresholding, denoted as I.

[0028] Furthermore, step 2 includes dilation corrosion, which is performed as follows:

[0029] Image erosion: For a binary image I, a structural element B, and for each pixel (x, y) in the image, the result of the erosion operation, E, can be expressed as:

[0030]

[0031] Image dilation: For a binary image I, a structural element B, and for each pixel (x, y) in the image, the result ID of the dilation operation can be expressed as:

[0032]

[0033] Here, the structural element B is a small two-dimensional matrix, usually with a defined shape and size, whose element values ​​are only 0 or 1, and is used to define the shape relationship within the neighborhood. In the erosion operation, only when the structural element completely falls in the foreground of the image (usually white or high brightness area), the corresponding output pixel is retained as the foreground; in the expansion operation, as long as the structural element intersects any part of the image, the output pixel will be marked as the foreground.

[0034] Furthermore, step four includes:

[0035] The edge is obtained by calculating the difference between the image after dilation and erosion, and its morphological gradient I G for:

[0036] I G =I D -I E

[0037] Among them, I D is the result of image expansion, I E It is the result of image corrosion. This method can effectively highlight the mutation area in the image, that is, the potential edge, thereby forming a mask to divide the curved area and the non-curved area.

[0038] Further, step five includes identifying and separating each interconnected set of pixels in the image, including:

[0039] Mask the captured brightness data to obtain the surface brightness data that needs to be corrected, and divide it into connected areas. The specific steps are as follows:

[0040] (1) Connected component labeling

[0041] Use various algorithms (such as depth-first search DFS or breadth-first search BFS) to traverse the image, starting from each unvisited foreground pixel, compare it with the labeled pixels, and if they are adjacent, classify them into the same connected region and give them the same label;

[0042] (2) Regional growth:

[0043] Region growing is performed according to the threshold, and pixels with similar attributes to the seed point are gradually merged to form a connected region. The steps are as follows:

[0044] S1, initialization: select one or more seed points and assign initial labels to them;

[0045] S2. Growth criterion: Set a threshold function T based on gray value (or other feature value), which can usually be expressed as:

[0046] |I(p)-I(q)|≤T

[0047] Where I(p) is the gray value of the seed point p, I(q) is the gray value of the pixel q to be detected, and T is a preset threshold;

[0048] S3, growth process: starting from the seed point, check whether each pixel in its neighborhood meets the growth criteria, that is, whether the difference between its gray value and the seed point is less than or equal to the threshold T;

[0049] If the conditions are met, add pixel q to the current region, update its label to the label of the seed point, add the neighborhood of pixel q that has just been added to the region to be processed, and continue the same inspection and growth process;

[0050] S4, termination condition: when the queue is empty, or reaches the preset stop condition (such as the specified maximum region size, the upper limit of the number of iterations, etc.), stop growing;

[0051] S5. Repeat the above process: If there are multiple seed points, perform the above region growing process on each seed point until all seed points and their connected pixels are processed.

[0052] Furthermore, step six includes morphological operations:

[0053] Morphological operations such as erosion and dilation are sometimes used to remove interfering noise, fill gaps, or separate objects that are too close to each other in order to better identify and distinguish connected areas.

[0054] Finally, through the above steps, the surface areas to be corrected can be connected, and the surface areas to be corrected can be identified more accurately, which plays a key role in the subsequent brightness data correction task.

[0055] Furthermore, step eight includes:

[0056] The brightness of the adjacent edge of the connected area is used as the correction target brightness, and data correction is performed in the column / row direction to align it to the adjacent non-curved surface position.

[0057] Divide the connected area to be corrected into m×n small areas. The size of the area can be adjusted according to different application scenarios.

[0058] Calculate the mean Lv of each small area mean(m，n) ;

[0059] Determine the target value of each connected area: Lv tar (m, n);

[0060] Calculate the corrected gain for each connected region:

[0061]

[0062] Traverse and calculate the Gain (x, y) of each point, and obtain it through Gain interpolation of adjacent connected areas;

[0063] The correction calculation for a certain pixel is:

[0064] Lv new (x,y)=Gain(x,y)*Lv(x,y). BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 The figure shows a schematic diagram of the brightness and darkness of the data captured by an OLED curved panel;

[0066] Figure 2 The figure shows a data flow chart of automatically correcting the brightness of a curved area of ​​an OLED panel for photographing according to the present invention;

[0067] Figure 3 Shown are schematic diagrams of the curved surface (dark) and non-curved surface (normal) areas;

[0068] Figure 4 Shown is a schematic diagram of the binary division result of the present invention (partial);

[0069] Figure 5 Shown is a schematic diagram (partial) of removing noise from spots of the present invention;

[0070] Figure 6 Shown is a schematic diagram of the masking result of the present invention;

[0071] Figure 7 Shown is a schematic diagram (partial) of the connected area division of the present invention;

[0072] Figure 8Shown is a schematic diagram of the surface data correction effect of the present invention. DETAILED DESCRIPTION

[0073] 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 only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] See also Figure 1-7 , a method for automatically correcting the brightness photographic data of the curved surface area of ​​an OLED panel, such as Figure 2 As shown, including:

[0075] Step 1: The industrial camera captures different grayscale brightness data of the curved OLED display panel;

[0076] Step 2: Brightness data extraction;

[0077] Step 3: Input the data to be corrected;

[0078] Step 4: Correct area identification;

[0079] Step 5: Determine whether the area recognition is normal. If it is normal, proceed to the next step; if it is not normal, the parameters are adaptively adjusted until the area recognition is successful, and then proceed to the next step;

[0080] Step 6: Interference removal;

[0081] Step 7: Determine whether the interference removal is successful. If successful, proceed to the next step; if unsuccessful, perform parameter adaptive adjustment until the interference removal is successful.

[0082] Step 8, data correction: determine whether the upper and lower surfaces need to be processed. If yes, process the upper and lower surfaces. If not, determine whether the left and right surfaces need to be processed. If yes, process the left and right surfaces. If not, proceed directly to the next step.

[0083] Step 9: Output the corrected data.

[0084] Furthermore, step one includes:

[0085] Step 1.1: Data preprocessing:

[0086] Convert the captured OLED curved panel brightness data into grayscale space;

[0087] Step 1.2, define the window size:

[0088] For each pixel in the image, a neighborhood window of size w×w is selected, where w is determined based on the characteristics of the actual image, the noise level, and the desired segmentation effect;

[0089] Step 1.3, calculate the local threshold:

[0090] For each pixel p in the window, calculate the statistical characteristics (mean) of the grayscale values ​​of all pixels in the window, denoted as T p ; Use the average grayscale in the neighborhood as the threshold basis, and then add a constant C;

[0091] Among them, for pixel S p , its adaptive threshold expression is:

[0092] T p =mean(I neighborhood )+C

[0093] Step 1.4, binary decision:

[0094] For pixel S p , according to the calculated local threshold T p Binarization:

[0095] If S p >T p , then set pixel p to white (usually 255 for an 8-bit grayscale image);

[0096] If S p <T p , then set pixel p to black (usually 0 for an 8-bit grayscale image);

[0097] Step 1.5, result output:

[0098] Repeat the above steps until all pixels in the image are traversed, and the resulting image is a binary image after adaptive threshold segmentation, denoted as I. Figure 3-4 shown.

[0099] The operation of dilation corrosion is as follows:

[0100] Image erosion: For a binary image I, a structural element B, for each pixel (x, y) in the image, the result of the erosion operation I E It can be expressed as:

[0101]

[0102] Image dilation: For a binary image I, a structural element B, for each pixel (x, y) in the image, the result of the dilation operation I D It can be expressed as:

[0103]

[0104] Here, the structure element B is a small two-dimensional matrix, usually with a defined shape and size, whose element values ​​are only 0 or 1, and is used to define the shape relationship within the neighborhood. In the erosion operation, only when the structure element completely falls in the foreground of the image (usually white or high brightness area), the corresponding output pixel is retained as the foreground; in the dilation operation, as long as the structure element intersects any part of the image, the output pixel will be marked as the foreground. Figure 5 shown.

[0105] The edge is obtained by calculating the difference between the image after dilation and erosion, and its morphological gradient I G for:

[0106] I G =ID-IE

[0107] Among them, I D is the result of image expansion, I E is the result of image corrosion. This method can effectively highlight the mutation area in the image, that is, the potential edge, thereby forming a mask to divide the surface area and the non-surface area. Figure 6 shown.

[0108] Dividing connected regions refers to identifying and separating the interconnected pixel sets in the image. It can support eight-connected or four-connected operation conditions. Eight-connected means that in a binary image, if a pixel is adjacent to another pixel in the up, down, left, right, and diagonal directions, the two pixels are considered to be eight-connected. Four-connected means that only the adjacent pixels in the four directions of up, down, left, and right are considered. Mask the captured brightness data to obtain the surface brightness data that needs to be corrected, and divide it into connected regions. The specific steps are as follows:

[0109] (1) Connected component labeling

[0110] Use various algorithms (such as depth-first search DFS or breadth-first search BFS) to traverse the image, starting from each unvisited foreground pixel, compare it with the labeled pixels, and if they are adjacent, classify them into the same connected region and give them the same label;

[0111] (2) Regional growth:

[0112] Region growing is performed according to the threshold, and pixels with similar attributes to the seed point are gradually merged to form a connected region. The steps are as follows:

[0113] S1, initialization: select one or more seed points and assign initial labels to them;

[0114] S2. Growth criterion: Set a threshold function T based on gray value (or other feature value), which can usually be expressed as:

[0115] |I(p)-I(q)|≤T

[0116] Where I(p) is the gray value of the seed point p, I(q) is the gray value of the pixel q to be detected, and T is a preset threshold;

[0117] S3, growth process: starting from the seed point, check whether each pixel in its neighborhood meets the growth criteria, that is, whether the difference between its gray value and the seed point is less than or equal to the threshold T;

[0118] If the conditions are met, add pixel q to the current region and update its label to the seed point label. Add the neighborhood of pixel q that has just been added to the region to be processed and continue the same inspection and growth process.

[0119] S4, termination condition: when the queue is empty, or reaches the preset stop condition (such as the specified maximum region size, the upper limit of the number of iterations, etc.), stop growing;

[0120] S5. Repeat the above process: If there are multiple seed points, perform the above region growing process on each seed point until all seed points and their connected pixels are processed.

[0121] The morphological operation steps are as follows:

[0122] Morphological operations such as erosion and dilation are sometimes used to remove interfering noise, fill gaps, or separate objects that are too close to each other in order to better identify and distinguish connected areas.

[0123] Finally, through the above steps, the surface areas to be corrected can be connected, and the surface areas to be corrected can be identified more accurately, which plays a key role in the subsequent brightness data correction task. Figure 7 shown.

[0124] The brightness of the adjacent edge of the connected area is used as the correction target brightness, and data correction is performed in the column / row direction to align it to the adjacent non-curved surface position.

[0125] Divide the connected area to be corrected into m×n small areas. The size of the area can be adjusted according to different application scenarios.

[0126] Calculate the mean Lv of each small area mean (m,n);

[0127] Determine the target value of each connected area: Lv tar (m, n);

[0128] Calculate the corrected gain for each connected region:

[0129]

[0130] Traverse and calculate the Gain (x, y) of each point, and obtain it through Gain interpolation of adjacent connected areas;

[0131] The correction calculation for a certain pixel is:

[0132] Lv new (x,y)=Gain(x,y)*Lv(x,y). like Figure 8 shown.

[0133] A system for automatically correcting brightness photographing data of a curved surface area of ​​an OLED panel comprises: a threshold segmentation module, an expansion and corrosion module, an edge detection module, a connected area division module, and a curved brightness data correction module.

[0134] Threshold segmentation module, used for adaptive threshold segmentation to obtain binary images and preliminarily divide the surface area and non-surface area;

[0135] The expansion and corrosion module performs expansion and corrosion operations on the image to remove noise while maintaining the morphological structure of the divided area;

[0136] The edge detection module obtains the edge by calculating the difference between the image after dilation and erosion;

[0137] The connected area division module performs mask processing on the captured brightness data to obtain the surface brightness data that needs to be corrected and divides it into connected areas;

[0138] The curved brightness data correction module takes the brightness of the adjacent edge of the connected area as the correction target brightness, performs data correction in the column / row direction, and aligns it to the adjacent non-curved surface position.

[0139] Its advantage is that it applies the adaptive threshold segmentation method combined with morphological operations to establish an accurate and efficient segmentation model, and realizes the adaptive segmentation of curved and non-curved areas. At the same time, the correction area is divided based on the regional connectivity characteristics of the OLED panel itself, and the brightness at the regional segmentation point is used as the correction target brightness, and corrections are made one by one through regional traversal. Compared with manually determining the surface segmentation parameters based on actual panel debugging and the brightness correction method based on fitting, the efficiency and accuracy of the brightness correction process are greatly improved, thereby optimizing the automation of production line operations, improving accuracy and output speed, and having important practical value.

[0140] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is limited by the appended claims and their equivalents.

Claims

1. A method for automatically correcting the brightness photographic data of a curved surface area of ​​an OLED panel, characterized in that: Step 1: The industrial camera captures different grayscale brightness data of the curved OLED display panel; Step 2: Brightness data extraction; Step 3: Input the data to be corrected; Step 4: Correct area identification; Step 5: Determine whether the area recognition is normal. If it is normal, proceed to the next step; if it is not normal, the parameters are adaptively adjusted until the area recognition is successful, and then proceed to the next step; Step 6: Interference removal; Step 7: Determine whether the interference removal is successful. If successful, proceed to the next step; if unsuccessful, perform parameter adaptive adjustment until the interference removal is successful. Step 8, data correction: determine whether the upper and lower surfaces need to be processed, if yes, process the upper and lower surfaces, if not, determine whether the left and right surfaces need to be processed, if yes, process the left and right surfaces, if not, proceed directly to the next step; Step 9: Output the corrected data.

2. The method for automatically correcting the brightness photographic data of the curved surface area of ​​an OLED panel according to claim 1, characterized in that: The step one comprises: Step 1.1: Data preprocessing: Convert the captured OLED curved panel brightness data into grayscale space; Step 1.2, define the window size: For each pixel in the image, a neighborhood window of size w×w is selected, where w is determined based on the characteristics of the actual image, the noise level, and the desired segmentation effect; Step 1.3, calculate the local threshold: For each pixel p in the window, the statistical characteristics of the grayscale values ​​of all pixels in the window are calculated as the mean, denoted as T p ; Use the average grayscale in the neighborhood as the threshold basis, and then add a constant C; Among them, for pixel S p , its adaptive threshold expression is: T p =mean(I neighborhood )+C Step 1.4, binary decision: For pixel S p , according to the calculated local threshold T p Binarization: If S p >T p , then set pixel p to white; If S p <T p , then set pixel p to black; Step 1.5, result output: Repeat the above steps until all pixels in the image are traversed, and the resulting binary image is segmented by adaptive thresholding, denoted as I.

3. The method for automatically correcting the brightness photographic data of the curved surface area of ​​an OLED panel according to claim 1, characterized in that: The step 2 includes expansion corrosion, and the operation is as follows: Image erosion: For a binary image I, a structural element B, for each pixel (x, y) in the image, the result of the erosion operation I E It is expressed as: Image dilation: For a binary image I, a structural element B, for each pixel (x, y) in the image, the result ID of the dilation operation is expressed as: The structure element B is a small two-dimensional matrix that defines the shape and size. Its element values ​​are only 0 or 1, and it is used to define the shape relationship within the neighborhood. In the erosion operation, only when the structure element completely falls in the foreground of the image, that is, falls into the white or high-brightness area, the corresponding output pixel is retained as the foreground; in the expansion operation, as long as the structure element intersects with any part of the image, the output pixel will be marked as the foreground.

4. The method for automatically correcting the brightness photographic data of the curved surface area of ​​an OLED panel according to claim 3, characterized in that: The fourth step comprises: The edge is obtained by calculating the difference between the image after dilation and erosion, and its morphological gradient I G for: I G =ID-IE Among them, I D is the result of image expansion, I E It is the result of image corrosion, which obtains the mutation area in the image, that is, the potential edge, thereby forming a mask to divide the curved surface area and the non-curved surface area.

5. The method for automatically correcting the brightness photographic data of the curved surface area of ​​an OLED panel according to claim 3, characterized in that: The step five includes identifying and separating each interconnected set of pixels in the image, including: Mask the captured brightness data to obtain the surface brightness data that needs to be corrected, and divide it into connected areas. The specific steps are as follows: (1) Connected component labeling Use the depth-first search DFS method or the breadth-first search BFS method to traverse the image, starting from each unvisited foreground pixel, compare it with the marked pixels, and if they are adjacent, classify them into the same connected region and assign them the same label; (2) Regional growth: Region growing is performed according to the threshold, and pixels with similar attributes to the seed point are gradually merged to form a connected region. The steps are as follows: S1, initialization: select one or more seed points and assign initial labels to them; S2, Growth criterion: Set a threshold function T based on gray value, expressed as: |I(p)-I(q)|≤T Where I(p) is the gray value of the seed point p, I(q) is the gray value of the pixel q to be detected, and T is a preset threshold; S3, growth process: starting from the seed point, check whether each pixel in its neighborhood meets the growth criteria, that is, whether the difference between its gray value and the seed point is less than or equal to the threshold T; If the conditions are met, add pixel q to the current region, update its label to the label of the seed point, add the neighborhood of pixel q that has just been added to the region to be processed, and continue the same inspection and growth process; S4, termination condition: when the queue is empty or reaches the preset stop condition, it stops growing; S5. Repeat the above process: If there are multiple seed points, perform the above region growing process on each seed point until all seed points and their connected pixels are processed.

6. The method for automatically correcting the brightness photographic data of the curved surface area of ​​an OLED panel according to claim 5, characterized in that: The step six includes morphological operations: It includes corrosion and expansion, eliminating interference noise, filling gaps or separating objects that are too close, identifying and distinguishing connected areas, connecting the surface areas to be corrected, and identifying the surface areas that need to be corrected.

7. The method for automatically correcting the brightness photographic data of the curved surface area of ​​an OLED panel according to claim 3, characterized in that: The step eight comprises: The brightness of the adjacent edge of the connected area is used as the correction target brightness, and data correction is performed in the column / row direction to align it to the adjacent non-curved surface position. Divide the connected area to be corrected into m×n small areas, and adjust the area size according to different application scenarios; Calculate the mean Lv of each small area mean (m,n); Determine the target value of each connected area: Lv tar (m, n); Calculate the corrected gain for each connected region: Traverse and calculate the Gain (x, y) of each point, and obtain it through Gain interpolation of adjacent connected areas; The correction calculation for a certain pixel is: Lv new (x,y)=Gain(x,y)*Lv(x,y)。 8. A system for automatically correcting brightness photographic data of an OLED panel curved surface area according to any one of claims 1 to 7, comprising: Threshold segmentation module, expansion and corrosion module, edge detection module, connected area division module, and curved brightness data correction module. Threshold segmentation module, used for adaptive threshold segmentation to obtain binary images and preliminarily divide the surface area and non-surface area; The expansion and corrosion module performs expansion and corrosion operations on the image to remove noise while maintaining the morphological structure of the divided area; The edge detection module obtains the edge by calculating the difference between the image after dilation and erosion; The connected area division module performs mask processing on the captured brightness data to obtain the surface brightness data that needs to be corrected and divides it into connected areas; The curved brightness data correction module takes the brightness of the adjacent edge of the connected area as the correction target brightness, performs data correction in the column / row direction, and aligns it to the adjacent non-curved surface position.

Images