A method and system for improving the accuracy of data captured by a Demura camera

Abstract

The present invention provides a method and system for improving the accuracy of demura camera data. Based on the brightness data acquired by an industrial camera, the system analyzes the characteristics of the brightness data and optimizes the brightness data processing to achieve a better demura compensation effect. Unlike compensation strategies that use multiple iterations, this method achieves efficient compensation effect optimization. Furthermore, the present invention can verify the mura compensation effect of a display panel. By comparing the projection information and distribution characteristics before and after compensation, the degree of mura compensation can be analyzed more comprehensively and accurately, thereby ensuring the brightness data processing parameters for each pixel to achieve the optimal compensation effect, significantly improving the brightness uniformity of the display panel.

Description

Technical Field

[0001] The present invention belongs to the field of display technology, and in particular relates to a method and system for improving the accuracy of data captured by a Demura camera. Background Art

[0002] It is an inevitable trend for flat panel display (FPD) technology to become the mainstream display technology in the future. With its advantages of lightness, high dynamic range, and low power consumption, it is widely used in smartphones, computers, TVs, car displays, wearable devices, AR / VR devices and other fields.

[0003] However, due to manufacturing process limitations, material selection constraints, design considerations, and other factors during the production process, flat panel displays inevitably exhibit non-uniformity. This non-uniformity ultimately manifests as variations in current distribution and inconsistent brightness, resulting in mura. Common types of mura include streaks, grids, clusters, sandy patches, and color unevenness. Advanced image processing techniques, known as demura compensation, are being used to improve brightness uniformity in displays. Mura is typically captured by capturing images of display samples using industrial cameras. This data is typically used directly for demura compensation and combined with engineers' actual tuning parameters to determine the final demura effect. Sometimes, if mura cannot be fully eliminated in a single pass, the demura process may need to be repeated one or more times based on the initial demura compensation results to improve mura performance. While this helps improve product yield to a certain extent, it also reduces work efficiency.

[0004] To address the above issues, the present invention proposes a method and system for improving the final demura compensation effect through data analysis and processing based on brightness data captured by industrial cameras. Summary of the Invention

[0005] The present invention provides a method for improving the accuracy of data captured by a Demura camera, including data acquisition and data analysis.

[0006] Data collection includes the following steps:

[0007] The first step is to determine the R / G / B color and grayscale of the photo according to the original Mura performance of the display panel;

[0008] In the second step, the industrial camera takes pictures of different grayscale images of the display panel to obtain a brightness data set for each pixel of the display panel;

[0009] Data analysis includes the following steps:

[0010] Step 3: Luminance Distribution Statistics and Analysis: Perform brightness statistical analysis on the display panel mura data and generate a brightness visualization to understand the characteristic distribution of pixels under different display screens. Perform spatial visualization analysis on the data to understand the spatial distribution characteristics of brightness under different display screens.

[0011] Step 4: Determine the brightness data adjustment parameters: Based on the analysis, determine the brightness data transformation and adjustment processing parameters, and adjust the brightness data to more accurately reflect the actual brightness distribution of the screen.

[0012] Furthermore, in step 1, an industrial camera is used to capture the brightness and chromaticity data of the three R / G / B images at different grayscale levels M, and the grayscale levels M are 32, 64, 128, 192 and 255 respectively.

[0013] Furthermore, in step 2, after the detection platform built based on the industrial camera completes the adaptation of the new screen according to the resolution information and pixel arrangement information of the display panel, it takes a picture according to the pre-determined display screen, and the brightness value of the pixel in the hth row and wth column on the display screen is , the brightness data of each shooting picture at different gray levels are recorded as .

[0014] Furthermore, in step 3, the maximum value, minimum value, and variance of the brightness data are statistically analyzed; the entire area of ​​the sub-pixel brightness data captured and output by the industrial camera is used as the ROI, and five grayscales are captured for each of the three colors R / G / B, for a total of fifteen groups of data The same processing steps are performed as follows:

[0015] S3.1 Normalization:

[0016] According to the maximum and minimum values ​​counted, the brightness data is normalized. The normalized value of the pixel in the hth row and the wth column is:

[0017] ;

[0018] in, For an R / G / B color image, the maximum value of the brightness data captured at grayscale level M is obtained. The minimum value of the brightness data captured at grayscale level M when the R / G / B color picture is normalized is recorded as ;

[0019] S3.2ROI generation:

[0020] ;

[0021] in The operator is The calculation result is rounded. is the value in the ROI image corresponding to the pixel in the hth row and the wth column, and the image after data visualization is recorded as .

[0022] Furthermore, in step 4, the transformation of the brightness data is to convert the spatial domain data into the frequency domain, including:

[0023] S4.1 Two-dimensional discrete Fourier transform, transform the two-dimensional discrete brightness data from the spatial domain to the frequency domain, Considered as a two-dimensional discrete signal , where h and w represent spatial coordinates, and its two-dimensional DFT is defined as follows:

[0024] ,

[0025] in, Represents the value of the frequency domain, u and v are frequency coordinates, H and W are the height and width of the data, is the rotation factor, j is the imaginary unit;

[0026] S4.2 Frequency Centering: Move the low-frequency components from the four corners of the spectrum to the center. The definition of centering is as follows:

[0027] ,

[0028] in, It is an alternating transformation factor, which rearranges the spectrum and places the low-frequency component in the center, thereby achieving frequency centralization. The result after centralization;

[0029] S4.3 High and low frequency information processing is achieved by designing filters. A circular filter is used and the part inside the circle is defined as the low frequency part, which is recorded as , the part outside the circle is defined as the high frequency part, recorded as The adjustment of high and low frequency information is achieved through the frequency adjustment factor, which is defined as and , high and low frequencies are processed as follows:

[0030] ;

[0031] ;

[0032] The processed frequency domain data is recorded as ;

[0033] S4.4 Inverse Fourier transform, after the data processing in the frequency domain is completed, the frequency domain data is converted back to the original spatial domain using the two-dimensional discrete inverse Fourier transform, which is defined as:

[0034] .

[0035] The present invention also provides a system for improving the accuracy of data captured by a Demura camera, which adopts the above method and includes:

[0036] Industrial camera brightness data acquisition unit: responsible for acquiring brightness data when the display panel displays different color grayscale images.

[0037] Brightness data analysis and processing unit: Analyzes the collected brightness data and processes the data based on discrete Fourier transform.

[0038] Demura compensation calculation unit: performs brightness non-uniformity compensation based on the processed brightness data.

[0039] The method provided by this invention uses an industrial camera to capture display panel brightness data. Through in-depth analysis and statistical analysis of this data, it enables refined data processing, significantly improving the accuracy and effectiveness of brightness compensation. This process encompasses multiple key steps and strategies, aiming to maximize the mining and utilization of information contained in the dataset. By correcting the captured data, the accuracy of the data is improved, ultimately optimizing the final image quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 Shown is a flow chart of brightness data processing and compensation value calculation of the present invention;

[0041] Figure 2 Shown is a ROI diagram of brightness data of a display panel G32 of the present invention;

[0042] Figure 3 The figure shows the spatial visualization of brightness data of a display panel G32 of the present invention;

[0043] Figure 4 The figure shows the 2DDFT spectrum of brightness data of a display panel G32 of the present invention (centralized result);

[0044] Figure 5 The figure shows a schematic diagram of the high and low frequency area division of the present invention;

[0045] Figure 6 A system for improving the accuracy of data captured by a Demura camera according to the present invention is shown;

[0046] Figure 7 Shown is a projection comparison of photographic data of a display panel of the present invention;

[0047] Figure 8 The figure shows the distribution comparison of photographing data of a display panel of the present invention. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0049] See Figure 1-8 The present invention provides a method for improving the accuracy of Demura camera shooting data. The method includes two parts: brightness data shooting and shooting data analysis and processing, as follows:

[0050] (S1-S2) Brightness data photography:

[0051] S1 photo picture and grayscale determination;

[0052] Based on the Mura performance of the display panel, determine the brightness and color data of N grayscales (for example, 32, 64, 128, 192, and 255 grayscales) captured by an industrial camera for the R / G / B images.

[0053] S2 Mura data photography;

[0054] After the detection platform built based on industrial cameras completes the adaptation of the new screen according to the resolution information and pixel arrangement information of the display panel, it takes pictures according to the pre-determined display screen. The brightness of the pixel in the hth row and wth column on the display screen ( ) value, the brightness data of each shooting picture at different grayscales can be recorded as .

[0055] (S3) Photo data analysis and processing:

[0056] (1) Statistical analysis of photographic data;

[0057] To analyze the brightness characteristics of a display, it's necessary to perform statistical analysis on the brightness data, including maximum and minimum values, and variance. These statistics help us understand the data's distribution, range, and degree of dispersion, thereby revealing key characteristics of mura. These statistics serve as the basis for subsequent data adjustments to improve image quality and enhance the performance of subsequent processing tasks.

[0058] (2) Brightness visualization processing;

[0059] In computer vision tasks, images of regions of interest (ROIs) are often visualized, known as ROI diagrams. These regions typically contain mura information that we want to identify, analyze, or process. The entire sub-pixel brightness data captured and output by an industrial camera serves as the ROI. The same processing steps are performed on fifteen sets of data, each capturing five grayscale levels for each of the three colors R / G / B. This is illustrated using data from a display panel G32 as an example:

[0060] S3.1 Normalization;

[0061] According to the maximum and minimum values ​​counted, the brightness data is normalized. The normalized value of the pixel in the hth row and the wth column is: ;

[0062] in, The maximum value of the brightness data captured by G32. is the minimum value of the G32 brightness data. The normalized G32 brightness data is recorded as ,

[0063] S3.2ROI generation;

[0064] ;

[0065] in The operator is The calculation result is rounded. is the value in the ROI image corresponding to the pixel in the hth row and the wth column. The image after data visualization is recorded as .

[0066] from Figure 2 The ROI diagram shows the approximate Mura category and distribution of the OLED screen, but the brightness and darkness differences between pixels cannot be intuitively reflected.

[0067] Spatial visualization processing Figure 3 As shown in the figure, compared to the flat ROI map display, data space visualization can not only show the planar distribution of spatial data, but also its vertical distribution and mura fluctuations. This multi-dimensional data display can provide richer information and details, and a more comprehensive understanding of the morphological characteristics of mura.

[0068] (S4) frequency domain processing;

[0069] To improve the accuracy of signal analysis, simplify the signal processing process, enhance its flexibility, and further uncover hidden signal characteristics, it is often necessary to convert spatial domain data to the frequency domain for analysis. While the spatial domain provides an intuitive understanding of the signal's morphological characteristics, it often lacks a clear and direct understanding of the signal's more complex frequency structure and dynamic characteristics. Converting spatial domain data to the frequency domain allows mathematical tools such as Fourier transforms to decompose the signal into a combination of different frequency components, enabling more precise analysis of the intensity, phase, and dynamics of each frequency component. This conversion not only provides a deeper understanding of the signal's physical nature but also simplifies signal processing and improves efficiency.

[0070] S4.1 Two-dimensional discrete Fourier transform;

[0071] 2D Discrete Fourier Transform (2DDFT) transforms the two-dimensional discrete brightness data from the spatial domain to the frequency domain. For any brightness data, such as It can be viewed as a two-dimensional discrete signal , where h and w represent spatial coordinates, and its two-dimensional DFT is defined as follows:

[0072] ;

[0073] in, Represents the value of the frequency domain, u and v are frequency coordinates, H and W are the height and width of the data, is the rotation factor, and j is the imaginary unit.

[0074] S4.2 Frequency centering;

[0075] To better observe, analyze, and process frequency domain information, making data processing more efficient and intuitive, the transformed results are subjected to frequency centering. Frequency centering involves moving low-frequency components from the four corners of the spectrum to the center. After centering, low-frequency components containing the data's primary structural information are concentrated in the center of the spectrum, while high-frequency components containing edges and noise are distributed at the edges. Frequency centering not only facilitates analysis of the image's primary information but also makes it easier to define low-frequency and high-frequency regions based on actual needs, making it easier to process data in conjunction with mura characteristics. Centering is defined as follows:

[0076] ;

[0077] in, It is an alternating transformation factor that rearranges the spectrum and places the low-frequency components in the center, thereby achieving frequency centering. The result after centralization is shown in the following figure. Figure 4 shown.

[0078] S4.3 High and low frequency information processing;

[0079] In the spectrum graph, the low-frequency area usually appears as a bright central area. This is because the low-frequency components dominate the image and have a larger amplitude. The high-frequency area appears as dark spots or bright spots at the edge of the spectrum graph. Bright spots indicate that the high-frequency components have a higher amplitude, while dark spots indicate a smaller amplitude. By observing the spectrum graph, you can intuitively see the distribution of high and low frequency components in the image. In order to achieve specific processing of data, it is necessary to distinguish between high and low frequency areas, which is usually achieved by designing filters. The filter shape is designed to be circular in the present invention, and the high and low frequency division radius can be adjusted according to actual needs. Figure 5 As shown in the figure, it is a schematic diagram of the division under a certain setting, where the part inside the circle is considered to be the low-frequency part, recorded as , the outer part of the circle is divided into high frequency parts, recorded as .

[0080] The adjustment of high and low frequency information is achieved through frequency adjustment factors, which are defined as and , high and low frequencies are processed as follows:

[0081] ;

[0082] ;

[0083] The goal of this embodiment is to improve the high-frequency part of the brightness data. is 0.8, is 1.0, and the processed frequency domain data is recorded as .

[0084] S4.4 Inverse Fourier transform;

[0085] After processing the data in the frequency domain, it is necessary to convert the frequency domain data back to the original spatial domain using the two-dimensional inverse discrete Fourier transform, which is defined as:

[0086] .

[0087] like Figure 6 The present invention provides a system for improving the accuracy of data captured by a Demura camera, which uses the above method and includes:

[0088] Industrial camera brightness data acquisition unit: responsible for acquiring brightness data when the display panel displays different color grayscale images.

[0089] Brightness data analysis and processing unit: Analyzes the collected brightness data and processes the data based on discrete Fourier transform.

[0090] Demura compensation calculation unit: performs brightness non-uniformity compensation based on the processed brightness data.

[0091] Another embodiment of the present invention is as follows: the captured brightness data is processed as in the above embodiment, and the processed data is used for demura compensation. Figure 7 The figure shows a data projection comparison of the brightness data of a display panel before and after optimization. The blue line represents the projection of the mura brightness data of the screen itself, the orange line represents the projection of the demura compensation effect data before optimization, and the green line represents the projection of the demura compensation effect data after optimization using the method of the present invention. From the data projection comparison performance, the brightness data processed by the method of the present invention can reduce fluctuations caused by shooting conditions or equipment errors, more accurately reflect the actual brightness distribution of the screen, and thus improve the uniformity of the final compensation effect.

[0092] The effectiveness of the method of the present invention is demonstrated from the perspective of local projection performance. Next, we will explain the overall performance by statistically analyzing the normal distribution results of the screen's own Mura brightness data (blue line), the Demura compensation effect data before optimization (orange line), and the Demura compensation effect data after optimization using the method of the present invention (green line). Figure 8 As shown, the brightness data processed by the method of the present invention combined with the demura compensation technology can achieve a distribution of compensated data points more closely around the target value. From the perspective of data concentration, the consistency of the compensation results is better.

[0093] In summary, the present invention provides a method and system for improving the accuracy of demura camera data. Based on the brightness data acquired by an industrial camera, the system analyzes the characteristics of the brightness data and optimizes the brightness data processing to achieve a better demura compensation effect. Unlike compensation strategies that use multiple iterations, this method achieves efficient optimization of the compensation effect. Furthermore, the present invention can verify the mura compensation effect of a display panel. By comparing the projection information and distribution characteristics before and after compensation, the degree of mura compensation can be analyzed more comprehensively and accurately, thereby ensuring the brightness data processing parameters for each pixel to achieve the optimal compensation effect, significantly improving the brightness uniformity of the display panel.

[0094] While 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 these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is limited by the appended claims and their equivalents.

Claims

1. A method for improving the accuracy of data captured by a Demura camera, characterized by: Including data collection and data analysis, Data collection includes the following steps: Step 1: Determine the color R / G / B and grayscale M for photographing based on the original Mura performance of the display panel; Step 2: The industrial camera takes pictures of the display panel with different colors and grayscales to obtain the brightness data of each pixel of the display panel. ; where h and w represent the pixel coordinates of the h-th row and w-th column spatial point respectively, Data analysis includes the following steps: Step 3: Luminance distribution statistics and analysis: Perform brightness statistical analysis on the display panel mura data and generate a planar ROI map to understand the characteristic distribution of pixels under different display screens. Perform spatial visualization analysis on the mura data to understand the spatial distribution characteristics of brightness under different display screens. Step 4: Determine the brightness data adjustment parameters: Based on the analysis, determine the brightness data transformation and adjustment processing parameters, and adjust the brightness data to more accurately reflect the actual brightness distribution of the screen; The fourth step includes: S4.1 Two-dimensional discrete Fourier transform, transform the two-dimensional discrete brightness data from the spatial domain to the frequency domain, Considered as a two-dimensional discrete signal , its two-dimensional DFT is defined as follows: ， in, Represents the value of the frequency domain, u and v are frequency coordinates, H and W are the height and width of the data, is the rotation factor, j is the imaginary unit; S4.2 Frequency Centering: Move the low-frequency components from the four corners of the spectrum to the center. The definition of centering is as follows: ， in, It is an alternating transformation factor, which rearranges the spectrum and places the low-frequency component in the center, thereby achieving frequency centralization. The result after centralization; S4.3 High and low frequency information processing is achieved by designing filters. A circular filter is used and the part inside the circle is defined as the low frequency part, which is recorded as , the part outside the circle is defined as the high frequency part, recorded as , the high and low frequency information is adjusted by the frequency adjustment factor, and the frequency adjustment factor is defined as and , high and low frequencies are processed as follows: ； ； The processed frequency domain data is recorded as ; S4.4 Inverse Fourier transform, after the data processing in the frequency domain is completed, the frequency domain data is converted back to the original spatial domain using the two-dimensional discrete inverse Fourier transform, which is defined as: 。 2. The method for improving the accuracy of data captured by a Demura camera according to claim 1, wherein: The grayscale levels M are 32, 64, 128, 192 and 255 respectively.

3. The method for improving the accuracy of data captured by a Demura camera according to claim 1, wherein: In step 2, the detection platform built based on the industrial camera completes the adaptation of the new screen according to the resolution information and pixel arrangement information of the display panel, and then takes pictures according to the pre-determined display screen. The brightness data of the pixels in the hth row and wth column on the display screen at different grayscales of the three colors R / G / B are recorded as .

4. The method for improving the accuracy of data captured by a Demura camera according to claim 3, wherein: In step 3, the maximum value, minimum value and variance of the brightness data are statistically analyzed; all sub-pixel level brightness data captured and output by the industrial camera The same processing steps are performed as follows: S3.1 Normalization: According to the maximum and minimum values ​​counted, the brightness data is normalized. The normalized value of the pixel in the hth row and the wth column is: ； in, The maximum value of the brightness data of the R / G / B color image captured at grayscale level M. The minimum value of the brightness data captured at grayscale level M for the R / G / B color image is recorded as ; S3.2ROI generation: ； in The operator is The calculation result is rounded. is the value in the ROI image corresponding to the pixel in the hth row and the wth column, and the image after data visualization is recorded as .

5. A system for improving the accuracy of data captured by a demura camera, comprising: include: Industrial camera brightness data acquisition unit: responsible for acquiring brightness data when the display panel displays different color grayscale images. Brightness data analysis and processing unit: Analyzes the collected brightness data and processes the data based on discrete Fourier transform. Demura compensation calculation unit: performs brightness non-uniformity compensation based on the processed brightness data.

Images