A display panel compensation data compression and recovery method

Abstract

The application discloses a display panel compensation data compression and recovery method. The method obtains compensation data of multiple channels of a display panel at different gray scales, and adopts a differentiated compression strategy according to the characteristics of the compensation data in different brightness regions. For low gray scale range compensation data, the correlation between different gray scale compensation data is utilized for prediction estimation, and the prediction error information is stored to realize data recovery. For middle and high gray scale range compensation data, the feature parameters of the compensation data are extracted, and the compression storage is realized by combining the quantized residual error information, so that the compensation data reconstruction is realized. The method can further improve the compression efficiency by combining the global feature parameters and the residual error information. Compared with the prior art, the application can effectively reduce the storage space requirement of the compensation data, improve the storage efficiency, and is suitable for the compression storage and recovery of display panel defect compensation data while ensuring the compensation data reconstruction accuracy.

Description

Technical Field

[0001] This invention belongs to the field of display panel brightness uniformity compensation and data compression technology, specifically relating to a grayscale compression and recovery method for display panel compensation data. Background Technology

[0002] During the manufacturing and use of OLED, Micro-LED, and other display panels, due to factors such as differences in the characteristics of light-emitting devices, the performance dispersion of thin-film transistors, manufacturing process fluctuations, and inconsistencies in driving circuits, display panels are prone to uneven brightness or color distribution when displaying the same grayscale image. This phenomenon is commonly referred to as a Mura defect. To mitigate or eliminate the impact of Mura defects on display performance, Demura compensation technology is typically employed. This involves optically inspecting the display panel to acquire display images at multiple grayscale levels, calculating corresponding compensation data based on the brightness deviation of each pixel or sub-pixel, storing the compensation data in a lower-level machine, and then compensating the input image during the display process to improve the brightness uniformity and color consistency of the display panel.

[0003] As display panel resolution increases, the amount of multi-channel, multi-grayscale, and pixel-by-pixel compensation data increases significantly. Directly storing the complete compensation matrix for each grayscale level not only consumes a large amount of space but also increases the time cost of data reading and processing. Existing compression methods typically employ a uniform compression strategy to process different grayscale data, but the data characteristics of compensation data at different grayscale levels are not the same. Low-grayscale compensation data exhibits strong correlation with adjacent grayscale levels, while mid-to-high-grayscale compensation data usually shows a locally concentrated or globally concentrated distribution in the spatial dimension. Therefore, if different compression strategies are not selected based on grayscale characteristics, unreasonable trade-offs between compression ratio and reconstruction accuracy can easily occur. Summary of the Invention

[0004] Objective of the Invention: The objective of this invention is to provide a method for compressing and restoring compensation data for display panels, addressing the problems of large data storage requirements, difficulty in balancing different grayscale data characteristics with a uniform compression method, and the challenge of achieving a balance between reconstruction accuracy and compression rate in existing display panel compensation data. This method employs appropriate compression and restoration methods for low-grayscale compensation data and medium-to-high-grayscale compensation data based on their distribution characteristics. This reduces the space occupied by compensation data storage while ensuring the effectiveness of the reconstructed compensation data and the display compensation effect, and improves the method's adaptability to different grayscale compensation data distributions.

[0005] Technical Solution: To achieve the above-mentioned objective, this invention provides a method for compressing and restoring compensation data for a display panel, comprising: acquiring original compensation data for multiple color channels of a display panel at multiple gray levels; dividing the original compensation data into low gray level compensation data and medium-high gray level compensation data according to the gray level range; performing adjacent gray level interpolation compression and storing a preset direction average error on the low gray level compensation data to obtain low gray level compressed data; performing mode representative value compression and quantization residual compression on the medium-high gray level compensation data to obtain medium-high gray level compressed data; and restoring reconstructed compensation data for the corresponding gray levels based on the low gray level compressed data and the medium-high gray level compressed data; wherein, the reconstructed compensation data is used to improve the display output effect of the display panel.

[0006] As a further improvement to the embodiments of the present invention, the original compensation data includes multiple compensation data channels corresponding to multiple color sub-pixels, and each compensation data channel includes multiple pixel-wise compensation matrices under multiple gray levels; the compensation data can be a Demura gray level offset, or it can be a compensation matrix equivalent to the Demura gray level offset.

[0007] As a further improvement to the embodiment of the present invention, the step of "performing adjacent grayscale interpolation compression and storing the preset direction average error on the low grayscale compensation data to obtain low grayscale compressed data" specifically includes: selecting the target low grayscale compensation matrix to be compressed and restored, and the adjacent reference grayscale compensation matrices located on both sides of the target low grayscale; performing interpolation calculation based on the adjacent reference grayscale compensation matrices to obtain the prediction compensation matrix of the target low grayscale; calculating the error matrix between the original compensation matrix of the target low grayscale and the prediction compensation matrix; averaging the error matrix along a preset direction to obtain the preset direction average error; and generating the low grayscale compressed data based on the adjacent reference grayscale compensation matrices and the preset direction average error.

[0008] As a further improvement to the embodiments of the present invention, the step of "recovering the corresponding gray level reconstruction compensation data based on the low gray level compressed data" specifically includes: recalculating the prediction compensation matrix of the target low gray level based on the adjacent reference gray level compensation matrix; expanding the preset direction average error to the same matrix size as the prediction compensation matrix; adding the expanded preset direction average error to the prediction compensation matrix to obtain the reconstruction compensation matrix of the target low gray level.

[0009] As a further improvement to the embodiments of the present invention, the step of "performing mode representative value compression and quantization residual compression on the medium-high grayscale compensation data to obtain medium-high grayscale compressed data" specifically includes: performing a block-based mode compression algorithm on the medium-high grayscale compensation data; wherein, the block-based mode compression algorithm includes: dividing the compensation matrix under medium-high grayscale into multiple pixel blocks; sequentially calculating the mode of the compensation data in each pixel block to obtain multiple block modes; wherein, each block mode corresponds to a single pixel block; arranging the block modes according to the position distribution of each pixel block in the compensation matrix to obtain mode mapping data; calculating the difference between each pixel compensation value in each pixel block and the corresponding block mode to obtain residual data; quantizing the residual data to obtain quantized residual data; and generating first medium-high grayscale compressed data based on the mode mapping data and the quantized residual data.

[0010] As a further improvement to the embodiments of the present invention, the step of "quantizing the residual data to obtain quantized residual data" specifically includes: quantizing the residual data according to a preset quantization step size, and setting the bit width and limiting range of the quantized residual data according to the dynamic range of the compensation data of the corresponding gray level.

[0011] As a further improvement to the embodiment of the present invention, the step of "recovering the reconstruction compensation data of the corresponding gray level based on the first medium-high gray level compressed data" specifically includes: reading the block mode corresponding to each pixel block; performing inverse quantization processing on the quantization residual data to obtain the recovery residual data; and adding the recovery residual data to the block mode of the corresponding pixel block to obtain the reconstruction compensation matrix of the corresponding medium-high gray level.

[0012] As a further improvement to the embodiment of the present invention, the step of "performing mode representative value compression and quantization residual compression on the medium-high grayscale compensation data to obtain medium-high grayscale compressed data" further includes: performing a global mode compression algorithm on the medium-high grayscale compensation data; wherein, the global mode compression algorithm includes: calculating a global mode for the compensation matrix under each color channel and each medium-high grayscale; calculating the difference between the compensation value of each pixel in the compensation matrix and the global mode to obtain global residual data; performing quantization and amplitude limiting processing on the global residual data to obtain global quantization residual data; and generating second medium-high grayscale compressed data based on the global mode and the global quantization residual data.

[0013] As a further improvement to the embodiment of the present invention, the step of "recovering the corresponding gray level reconstruction compensation data based on the second medium-high gray level compressed data" specifically includes: reading the global mode; performing inverse quantization processing on the global quantization residual data to obtain the recovered global residual data; and adding the recovered global residual data to the global mode to obtain the corresponding medium-high gray level reconstruction compensation matrix.

[0014] As a further improvement to the embodiments of the present invention, the method further includes: performing error evaluation on the reconstructed compensation data to obtain error evaluation results; wherein, the error evaluation results include the root mean square error and / or maximum absolute error between the original compensation data and the reconstructed compensation data; and selecting a compression algorithm for medium and high grayscale compensation data from block mode compression algorithm and global mode compression algorithm according to the error evaluation results, compression ratio and / or storage resource requirements.

[0015] Beneficial Effects: Compared with existing technologies, this invention has the following significant advantages: Based on the data distribution characteristics of different grayscale compensation data, this invention employs different compression and recovery methods for low-grayscale compensation data and medium-to-high-grayscale compensation data. For low-grayscale compensation data, since there is a correlation between the target low-grayscale compensation matrix and its adjacent reference grayscale compensation matrices, this invention uses interpolation from the adjacent reference grayscale compensation matrices to obtain the prediction compensation matrix for the target low-grayscale, and only saves the directional average error of the prediction error in a preset direction, without saving the complete two-dimensional compensation matrix of the target low-grayscale. Since the directional average error is one-dimensional data, its data volume is smaller than that of the complete two-dimensional compensation matrix, thus reducing the storage volume of low-grayscale compensation data. Simultaneously, the directional average error can reflect the overall offset characteristics of the prediction error in the row or column direction. During recovery, by expanding the directional average error and adding it back to the prediction compensation matrix, the interpolation prediction deviation can be corrected, improving the accuracy of the target low-grayscale reconstruction compensation data.

[0016] For mid-to-high grayscale compensation data, since the compensation values ​​typically exhibit a concentrated distribution within local pixel blocks or the entire compensation matrix, this invention uses the block mode or global mode as a representative value to represent the main compensation amount, and only stores the quantization residual of each pixel's compensation value relative to the representative value. Because the dynamic range of the quantization residual is usually smaller than that of the original compensation value, a bit width smaller than the bit width of the original compensation value can be used to represent the quantization residual, thus reducing the storage space occupied by mid-to-high grayscale compensation data. Simultaneously, the quantization residual still retains pixel-level compensation differences; during recovery, by adding the dequantized residual back to the corresponding mode, the spatial details in the compensation matrix can be reconstructed.

[0017] Furthermore, this invention evaluates the reconstructed compensation data using root mean square error and maximum absolute error, and selects between block-based mode compression and global mode compression based on the error evaluation results, compression ratio, and storage resource requirements. When global mode compression meets the reconstruction accuracy requirements, its lower representative value storage overhead can be utilized to further improve the compression ratio; when global mode compression fails to meet the reconstruction accuracy requirements, block-based mode compression can be used to preserve local spatial distribution characteristics. Thus, this invention can reduce the storage space required for compensation data while maintaining the accuracy of reconstructed compensation data and the display compensation effect, and improve the adaptability of the compensation data compression method, based on different gray levels, different color channels, and different compensation data distributions. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall process of compressing compensation data according to the present invention.

[0019] Figure 2 This is a schematic diagram of the low grayscale interpolation and storage preset direction average error compression process of the present invention.

[0020] Figure 3 This is a schematic diagram of the medium-to-high grayscale mode residual compression process of the present invention. Detailed Implementation

[0021] The technical solution of the present invention will be described in detail below with reference to specific embodiments and accompanying drawings. It should be understood that the following embodiments are only used to illustrate the present invention and are not intended to limit the scope of protection of the present invention. Without departing from the concept of the present invention, those skilled in the art can adaptively adjust the grayscale values, pixel block size, quantization step size, residual bit width, and limiting range according to the resolution of the display panel, the number of grayscale levels, the number of color channels, storage resources, and compensation accuracy requirements.

[0022] like Figure 1 As shown, this invention first acquires multi-grayscale compensation data for multiple color channels of the display panel. The multiple color channels may include R, G, and B channels, or compensation data channels corresponding to other color sub-pixels. The multi-grayscale compensation data can be obtained by capturing detection images of the display panel at multiple grayscale levels using a camera, and then calculating them using the Demura algorithm or other brightness uniformity compensation algorithms. Taking several discrete grayscale levels as an example, each color channel can form a three-dimensional compensation matrix, where the first two dimensions correspond to the pixel rows and columns of the display panel, and the third dimension corresponds to the grayscale index.

[0023] In one specific embodiment, the plurality of gray levels may include 16 gray levels, 32 gray levels, 64 gray levels, 128 gray levels, 160 gray levels, 192 gray levels, and 255 gray levels. A gray level threshold Gth can be preset, for example, Gth=64, to classify compensation data with gray level values ​​not greater than Gth as low gray level compensation data, and compensation data with gray level values ​​greater than Gth as medium-high gray level compensation data. For low gray level compensation data, a target gray level located between adjacent reference gray levels can be selected as the gray level to be compressed and restored. For example, 16 gray levels and 64 gray levels can be used as adjacent reference gray levels, and 32 gray levels can be used as the target low gray level. The 32 gray level compensation matrix can be restored by interpolation of adjacent gray levels and storage of the average directional error. For medium-high gray level compensation data, 128 gray levels, 160 gray levels, 192 gray levels, and 255 gray levels can be used as a preset medium-high gray level set, and the compensation matrices of each gray level in this set can be compressed and restored using the mode representative value and quantization residual.

[0024] In other embodiments, low-grayscale compensation data and medium-high-grayscale compensation data can also be determined based on a preset grayscale set or the distribution characteristics of compensation data. For example, when the target grayscale compensation matrix has a strong correlation with the adjacent reference grayscale compensation matrix, and the interpolation prediction error shows a consistent distribution in the row or column direction, the target grayscale can be classified as low-grayscale compensation data; when the compensation value in a grayscale compensation matrix is ​​concentrated within a local pixel block or the entire matrix, and is suitable for representation using block mode or global mode, the grayscale can be classified as medium-high-grayscale compensation data.

[0025] Example 1

[0026] like Figure 2 As shown, this embodiment illustrates the compression and recovery process of low grayscale compensation data. In one specific embodiment, a target grayscale level within the low grayscale range is selected as the grayscale level to be compressed and recovered, and two adjacent grayscale levels located on either side of the target grayscale level are selected as reference grayscale levels. For any color channel, the two adjacent reference grayscale compensation matrices are denoted as the first reference compensation matrix and the second reference compensation matrix, respectively, and the original compensation matrix of the target grayscale level is the target compensation matrix. This embodiment first performs interpolation calculations based on the first and second reference compensation matrices to obtain the target grayscale prediction compensation matrix; then, it calculates the error matrix between the target compensation matrix and the target grayscale prediction compensation matrix.

[0027] Furthermore, since the target grayscale prediction error has a certain consistency along the preset direction, this embodiment averages the error matrix along the preset direction to obtain the directional average error. During compression, the complete original target grayscale compensation matrix is ​​not saved, but the first reference compensation matrix and the second reference compensation matrix are saved or reused, and the directional average error is additionally saved. During decompression and recovery, the directional average error is expanded into an error matrix of the same size as the target grayscale prediction compensation matrix, and the expanded directional average error is added to the target grayscale prediction compensation matrix to obtain the target grayscale reconstruction compensation data.

[0028] In one specific embodiment, the preset direction is the column direction. In this case, the error matrix is ​​averaged along the column direction to obtain a one-dimensional column average error. During recovery, the one-dimensional column average error is copied and extended along the row direction to the same size as the target grayscale prediction compensation matrix. This method corresponds to the low grayscale compression recovery process in this embodiment. In other embodiments, the preset direction can also be the row direction. In this case, the error matrix is ​​averaged along the row direction to obtain a one-dimensional row average error. During recovery, the one-dimensional row average error is copied and extended along the column direction to the same size as the target grayscale prediction compensation matrix.

[0029] In one specific embodiment, the target grayscale can be an intermediate grayscale level located between two reference grayscale levels within a low grayscale range, and the interpolation calculation can employ a linear interpolation method. For example, when the first reference compensation matrix is ​​denoted as A1 and the second reference compensation matrix is ​​denoted as A2, the target grayscale prediction compensation matrix can be calculated using A_est=(A1+A2) / 2; the error matrix can be calculated using err=A_target-A_est; the column average error can be obtained by averaging the error matrix column by column; and the target grayscale reconstruction compensation matrix can be obtained using A_rec=A_est+col_err, where A_target represents the original target grayscale compensation matrix, and col_err represents the expanded column average error.

[0030] In this embodiment, the data length of the column average error is equal to the number of columns in the compensation matrix, which is much smaller than the data size of the complete two-dimensional compensation matrix. Therefore, the storage overhead of the target low grayscale compensation data can be reduced. The restored target grayscale reconstruction compensation data can be recombine with other grayscale compensation data to form a complete multi-grayscale compensation data set for subsequent image compensation or display panel driving compensation.

[0031] Example 2

[0032] like Figure 3As shown, this embodiment illustrates the block mode and quantization residual compression process for high grayscale compensation data. For any grayscale level and any color channel within the high grayscale range, the corresponding compensation matrix is ​​read and divided into multiple pixel blocks. In a preferred embodiment, the pixel blocks can be non-overlapping square pixel blocks, such as 8×8 pixel blocks; in other embodiments, pixel blocks of other sizes or shapes can also be used. When the size of the compensation matrix cannot be divided evenly by the pixel block size, the edge regions can be cropped, padded, or treated as incomplete pixel blocks separately.

[0033] For each pixel block, the mode of all compensation values ​​within that pixel block is calculated, and this mode is used as the block mode of the pixel block. The block mode is used to characterize the most frequently occurring major compensation value within the pixel block. Subsequently, the difference between each pixel compensation value within the pixel block and the block mode is calculated to obtain the residual data corresponding to that pixel block.

[0034] Further, the residual data is quantized according to a preset quantization step size to obtain quantized residual data, and the bit width and limiting range of the quantized residual data are set according to the dynamic range of the compensation data for the corresponding grayscale. The compressed data includes mode mapping data composed of the block mode of each pixel block and the quantized residual data corresponding to each pixel. During decompression and recovery, the block mode of the corresponding pixel block is read, and the quantized residual data is dequantized according to the preset quantization step size, and then the corresponding block mode is added back to obtain the reconstruction compensation data of that pixel block.

[0035] In a specific example, the preset quantization step size can be set to 2. For mid-to-high grayscale compensation data of 128, 160, and 192 grayscale, if the original representation bit width of the compensation data is 6 bits, the bit width of the quantization residual data can be set to 4 bits, and the limiting range of the quantization residual can be set to -8 to 7. Specifically, for any pixel compensation value A and its corresponding block mode M within a pixel block, the residual r=AM is first calculated, and then the quantization residual q is obtained according to q=round(r / 2), and q is limited to the range of -8 to 7; during restoration, the restoration residual is obtained according to r_rec=q×2, and the restoration residual r_rec is added to the block mode M to obtain the reconstruction compensation value of the pixel.

[0036] For 255 grayscale compensation data, if the original representation bit width of the compensation data is 7 bits, the preset quantization step size can be set to 2, the bit width of the quantization residual data can be set to 5 bits, and the limiting range of the quantization residual can be set to -16 to 15. Specifically, the residual between the pixel compensation value and the block mode is calculated first, and then quantization and limiting are performed according to the preset quantization step size. During restoration, the quantization residual is multiplied by the preset quantization step size to obtain the restoration residual, and then added back to the corresponding block mode to obtain the reconstructed compensation value.

[0037] In other embodiments, the preset quantization step size, the bit width of the quantization residual data, and the limiting range can also be adjusted according to the grayscale range, channel type, dynamic range of the compensation data, or error evaluation results. For grayscales with a small dynamic range of compensation data, a lower bit width can be used to represent the quantization residual; for grayscales with a large dynamic range of compensation data, a higher bit width can be used to represent the quantization residual. Thus, this embodiment can reduce the overall storage amount of mid-to-high grayscale compensation data while preserving local spatial differences.

[0038] Example 3

[0039] This embodiment illustrates the global mode and quantization residual compression process for mid-to-high grayscale compensation data. For any grayscale level and any color channel within the mid-to-high grayscale range, the corresponding compensation matrix is ​​read, and the compensation value that appears most frequently in the entire compensation matrix is ​​calculated and taken as the global mode. Subsequently, the difference between the compensation value of each pixel in the compensation matrix and the global mode is calculated to obtain global residual data. The global residual data is then quantized and limited according to a preset quantization step size to obtain global quantized residual data.

[0040] In this embodiment, the global quantization residual data can also use the same quantization step size, bit width, and limiting range as in Embodiment 2. For example, for compensation data of 128 grayscale, 160 grayscale, and 192 grayscale, the preset quantization step size can be set to 2, the bit width of the quantization residual data can be set to 4 bits, and the limiting range of the quantization residual can be set to -8 to 7; for compensation data of 255 grayscale, the preset quantization step size can be set to 2, the bit width of the quantization residual data can be set to 5 bits, and the limiting range of the quantization residual can be set to -16 to 15. The difference is that this embodiment uses the global mode of the entire compensation matrix as the representative value, while Embodiment 2 uses the block mode of each pixel block as the representative value.

[0041] During compression, the global mode and global quantization residual data corresponding to each grayscale level and each color channel are saved. During decompression and recovery, the global mode is read, and the global quantization residual data is dequantized according to a preset quantization step size. Then, it is added to the global mode to obtain the reconstruction compensation matrix of the corresponding grayscale level.

[0042] Compared to Example 2, this example does not require storing the mode mapping data for each pixel block, but only needs to store the global mode and corresponding quantization residual for each channel of each grayscale, thus resulting in lower representative value storage overhead. This method is suitable for scenarios where the overall distribution of compensation data is concentrated, spatial changes are gradual, or hardware storage budgets are more stringent. In practical applications, Example 2 or Example 3 can be selected as the medium-to-high grayscale compression method based on reconstruction error, maximum error, compression ratio, grayscale range, channel type, or storage resource requirements. Alternatively, two compression methods can be used for different grayscale levels or different color channel combinations.

[0043] Example 4

[0044] This embodiment illustrates the process of selecting a medium-to-high grayscale compression method based on error evaluation results. For the same color channel and the same medium-to-high grayscale compensation matrix, compression and restoration can be performed according to the block-based mode compression algorithm described in Embodiment 2 and the global mode compression algorithm described in Embodiment 3, respectively, to obtain the first reconstruction compensation matrix and the second reconstruction compensation matrix.

[0045] Subsequently, the root mean square error (RMSE) and maximum absolute error (MAXE) between the original compensation matrix and each reconstructed compensation matrix are calculated. RMSE can be obtained by taking the square root of the average of the squared differences between the compensation values ​​of each pixel in the original compensation matrix and their corresponding reconstructed compensation values; MAXE can be determined by the maximum absolute difference between the compensation values ​​of each pixel and their corresponding reconstructed compensation values.

[0046] In one specific embodiment, the root mean square error threshold TR and the maximum absolute error threshold TM can be preset, for example, TR can be set to 1.5 and TM can be set to 20. If the RMSE corresponding to a certain compression method is not greater than TR and the MAXE is not greater than TM, then the compression method is considered to meet the reconstruction accuracy requirements. When both the block-based mode compression method and the global mode compression method meet the reconstruction accuracy requirements, a compression method with a higher compression ratio and smaller storage space can be selected; when only one compression method meets the reconstruction accuracy requirements, the compression method that meets the reconstruction accuracy requirements is selected; when neither compression method meets the reconstruction accuracy requirements, the quantization step size, quantization residual bit width, limiting range, or pixel block size can be adjusted, and compression, restoration, and error evaluation can be performed again.

[0047] For example, for a certain medium-to-high grayscale compensation matrix, if the RMSE of the block-based mode compression method is 1.04, the MAXE is 18, and the compression rate is 31.77%, while the RMSE of the global mode compression method is 1.47, the MAXE is 12, and the compression rate is 33.33%, and both meet the preset error threshold, then the global mode compression method with the higher compression rate can be selected. If, for another grayscale or another color channel, the error of the global mode compression method exceeds the preset threshold, while the block-based mode compression method meets the preset threshold, then the block-based mode compression method is selected. Thus, an adaptive balance between reconstruction accuracy and compression rate can be achieved based on different grayscale levels, different color channels, and different compensation data distributions.

Claims

1. A method for compressing and restoring compensation data for a display panel, characterized in that, include: Obtain the original compensation data of multiple color channels of the display panel at multiple gray levels; divide the original compensation data into low gray level compensation data and medium-high gray level compensation data according to the gray level range; The low grayscale compensation data is compressed using adjacent grayscale interpolation and a preset directional average error storage; the medium and high grayscale compensation data is compressed using the mode representative value and quantization residual; and the corresponding grayscale reconstruction compensation data is restored based on the compressed and saved data.

2. The method according to claim 1, characterized in that, The multiple color channels include any color channels that the display panel can support displaying, and the multiple gray levels include any gray level within the gray level range supported by the display panel; the low gray level compensation data and the medium and high gray level compensation data are determined according to a preset gray level threshold, a preset gray level set, or the distribution characteristics of the compensation data.

3. The method according to claim 1, characterized in that, The low grayscale compensation data includes the target low grayscale compensation data to be recovered, and the adjacent grayscale interpolation includes calculating the prediction compensation matrix of the target low grayscale using the adjacent reference grayscale compensation matrices located on both sides of the target low grayscale.

4. The method according to claim 3, characterized in that, The preset direction average error storage includes calculating the error matrix between the original compensation matrix of the target low grayscale and the predicted compensation matrix, and obtaining the direction average error along the preset direction; wherein, the preset direction is the row direction or the column direction; during compression, the adjacent reference grayscale compensation matrix and the direction average error are saved.

5. The method according to claim 4, characterized in that, During recovery, based on the predicted compensation matrix and the average directional error, the average directional error is expanded to the same matrix size as the predicted compensation matrix in the same preset direction as during compression, and then added to the predicted compensation matrix to obtain the reconstruction compensation matrix of the target low grayscale.

6. The method according to claim 1, characterized in that, When compressing the medium-to-high grayscale compensation data using block mode and quantization residual, the compensation matrix under medium-to-high grayscale is divided into multiple non-overlapping pixel blocks. The mode of the compensation data in each pixel block is calculated as the block mode of that pixel block, and the mode mapping data composed of the block modes of each pixel block is saved.

7. The method according to claim 6, characterized in that, The quantization residual is the residual value after quantization of the difference between the compensation value of each pixel in the pixel block and the mode of the corresponding block, with a preset quantization step size. The quantization step size is used to determine the corresponding data interval between adjacent quantization levels. The bit width and limiting range of the quantization residual are set according to the dynamic range of the compensation data of the corresponding gray level.

8. The method according to claim 1, characterized in that, When compressing the mid-to-high grayscale compensation data using the global mode and quantization residual, the global mode is calculated for each color channel and each mid-to-high grayscale compensation matrix, the residual of each pixel compensation value relative to the global mode is calculated, and the global mode and quantized residual data are saved.

9. The method according to claim 8, characterized in that, When recovering the global mode and the compressed quantized residual data, the reconstruction compensation matrix of the corresponding gray level is obtained based on the global mode, the quantized residual data, and the preset quantization step size.

10. The method according to any one of claims 1 to 9, characterized in that, It also includes error evaluation of the reconstruction compensation data, which includes calculating the root mean square error and maximum absolute error between the original compensation data and the reconstruction compensation data, and determining whether to use block mode compression or global mode compression based on the error evaluation results.

Images