A grayscale compensation method, device and storage medium for high-order polynomial fitting

Abstract

The present application discloses a grayscale compensation method, device and storage medium for high-order polynomial fitting, which are used in Demura compensation. The present application method includes: determining a target high-order polynomial, which is a relationship between the grayscale value and the brightness value of a display device; based on the initial iteration point, using Newton's method to iterate the target high-order polynomial, and calculating the next iteration point; after each iteration, judging whether the obtained iteration point meets the preset singularity condition; and\or; detecting the change trend of the first distance value and the first function value; when it is determined that the preset singularity condition is met, adjusting the iteration process through the singularity adjustment mechanism; when it is determined that the change trend meets the preset change condition, adjusting the iteration process through the divergence processing mechanism; judging whether the adjusted iteration process meets the preset convergence condition; if it is determined that it is met, outputting the convergence point; and compensating the grayscale of the display device according to the convergence point.

Description

Technical Field

[0001] The present application relates to the field of screen optical detection technology, and in particular to a grayscale compensation method, device and storage medium for high-order polynomial fitting. Background Art

[0002] With the continuous advancement of display technology, display devices, especially LCDs and OLEDs, have put forward higher requirements on the performance of color and brightness. In order to improve the display effect and eliminate the display deviation caused by display hardware, manufacturing errors and environmental changes, it is usually necessary to calibrate the grayscale of the display to ensure the accurate display of the image. Traditional grayscale compensation methods usually use linear relationship models to fit grayscale and brightness, but this method cannot completely solve the problem of nonlinear changes in display brightness. Especially in high-contrast areas, the display may have uneven brightness and color distortion, which affects the quality of the display effect.

[0003] To address this problem, the Demura compensation method uses a high-order polynomial fitting model to describe the nonlinear relationship between the input grayscale and the actual display brightness, and uses an iterative solution method to gradually correct the grayscale value to obtain more accurate compensation data. The Demura method can effectively improve the brightness uniformity and color consistency of the display device in each grayscale range, thereby achieving a higher quality display effect. This method is not only suitable for the production calibration of new display devices, but can also be used for the later optimization and correction of existing display devices.

[0004] However, in practical applications, due to the differences in specific display parameters and usage environments, traditional demura compensation methods still face some problems, such as: divergence problems in the solution process, numerical instability that may occur in iterative calculations, and a long time required for the solution process. Therefore, it is necessary to further optimize the demura compensation method to improve its accuracy, stability and real-time performance. Summary of the invention

[0005] In order to solve the above technical problems, the present application provides a grayscale compensation method, device and storage medium for high-order polynomial fitting.

[0006] The technical solution provided in this application is described below:

[0007] The first aspect of the present application provides a grayscale compensation method of high-order polynomial fitting, comprising:

[0008] Determine a target high-order polynomial, wherein the high-order polynomial is a relationship between a grayscale value and a brightness value of a display device, wherein a variable value represents a grayscale value, and a function value represents a brightness value;

[0009] Based on the initial iteration point, the target high-order polynomial is iterated using Newton's method, and the next iteration point is calculated;

[0010] After each iteration, it is determined whether the obtained iteration point meets the preset singularity condition;

[0011] and\or;

[0012] Detecting a change trend of a first distance value and a first function value;

[0013] When it is determined that the preset singularity conditions are met, the iterative process is adjusted through the singularity adjustment mechanism;

[0014] When it is determined that the change trend meets the preset change condition, the iterative process is adjusted through a divergent processing mechanism;

[0015] Determine whether the adjusted iteration process meets the preset convergence condition;

[0016] If it is determined that the condition is satisfied, a convergence point is output, where the convergence point represents a compensated grayscale value;

[0017] The grayscale of the display device is compensated according to the convergence point.

[0018] Optionally, the preset singularity condition includes:

[0019] Condition 1: The iterative point distance is less than the first threshold for N consecutive times, and the corresponding function value is greater than the second threshold;

[0020] or;

[0021] Condition 2: The first-order derivative value of the corresponding function value is less than a third threshold value, and the third threshold value is a value tending to zero.

[0022] Optionally, adjusting the iteration process by using a singularity adjustment mechanism includes:

[0023] A step factor is introduced into the iterative formula of the Newton method, and the value of the step factor is adjusted. The value range of the step factor is 0 to 1.

[0024] Optionally, adjusting the iteration process by using a singularity adjustment mechanism includes:

[0025] A random disturbance term is introduced into the iterative formula of the Newton method, and the random disturbance term is used to adjust the current iteration point.

[0026] Optionally, adjusting the iteration process by using a singularity adjustment mechanism includes:

[0027] The initial iteration point is updated.

[0028] Optionally, the preset change condition includes:

[0029] Condition 3: The distance between iteration points gradually increases;

[0030] or;

[0031] Condition 4: The reduction amplitude of the iteration point distance is not greater than a preset amplitude threshold and the current function value is greater than a preset function threshold.

[0032] Optionally, the adjusting the iteration process by using a divergent processing mechanism includes:

[0033] Introducing a step factor into the iterative formula of the Newton method and adjusting the value of the step factor, wherein the value range of the step factor is 0 to 1;

[0034] or;

[0035] Iterate using multiple different initial iteration points.

[0036] Optionally, the preset convergence condition includes:

[0037] Condition 5: The iteration point distance is less than a preset first precision threshold, and the function value is less than a preset second precision threshold.

[0038] Optionally, the compensating the grayscale of the display device according to the convergence point includes:

[0039] According to the convergence point, a mapping table is established between the grayscale value and the actual brightness value, and the compensation value that needs to be adjusted for each grayscale value is recorded;

[0040] For each pixel of the display device, generating a calibrated grayscale value using the convergence point;

[0041] The compensated grayscale value is input into the display device as a driving signal, so that the pixel points of the display device output corresponding brightness values.

[0042] The second aspect of the present application provides a grayscale compensation device for high-order polynomial fitting, comprising:

[0043] A polynomial determination unit, configured to determine a target high-order polynomial, wherein the high-order polynomial is a relationship between a grayscale value and a brightness value of a display device, wherein a variable value represents a grayscale value, and a function value represents a brightness value;

[0044] An iteration unit, used for iterating the target high-order polynomial based on an initial iteration point using Newton's method and calculating a next iteration point;

[0045] A singularity judgment unit, used to judge whether the obtained iteration point meets the preset singularity condition after each iteration;

[0046] and\or;

[0047] A divergence detection unit, used to detect a change trend of the first distance value and the first function value;

[0048] A singularity adjustment unit, used to adjust the iterative process through a singularity adjustment mechanism when it is determined that a preset singularity condition is met;

[0049] A divergence adjustment unit, configured to adjust the iterative process through a divergence processing mechanism when it is determined that the change trend meets a preset change condition;

[0050] A convergence judgment unit, used to judge whether the adjusted iterative process meets the preset convergence condition;

[0051] A convergence output unit, configured to output a convergence point if it is determined that the condition is satisfied, wherein the convergence point represents a compensated grayscale value;

[0052] The grayscale compensation unit is used to compensate the grayscale of the display device according to the convergence point.

[0053] A third aspect of the present application provides a grayscale compensation device for high-order polynomial fitting, the device comprising:

[0054] Processor, memory, input-output unit, and bus;

[0055] The processor is connected to the memory, the input and output unit, and the bus;

[0056] The memory stores a program, and the processor calls the program to execute the first aspect and any optional method in the first aspect.

[0057] A fourth aspect of the present application provides a computer-readable storage medium, on which a program is stored. When the program is executed on a computer, the program executes the first aspect and any optional method in the first aspect.

[0058] It can be seen from the above technical solutions that this application has the following advantages:

[0059] 1. The method of this application adopts a high-order polynomial model to describe the relationship between grayscale value and brightness value, which can more accurately fit complex nonlinear relationships and improve the accuracy of brightness correction. The Newton method is used to quickly solve the accurate grayscale value, which helps to achieve high-precision grayscale compensation.

[0060] 2. By detecting singularities and adaptively adjusting the step size, stagnation, oscillation or divergence caused by singularities during the iteration process can be avoided, ensuring the stability and reliability of the solution process. Real-time detection and adjustment of iterative divergence can effectively suppress the risk of divergence and improve the robustness of the solution process.

[0061] 3. The adjustment mechanism based on real-time trend detection can flexibly respond to changes in different display devices or complex environmental conditions. Applicable to various display devices (such as OLED, LCD, etc.), it can effectively correct the uneven brightness caused by panel process differences or aging.

[0062] 4. By accurately compensating the grayscale of the display device, the brightness uniformity and color consistency of the display can be significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solution in the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0064] Figure 1 A schematic diagram of an embodiment of a grayscale compensation method of high-order polynomial fitting provided in this application;

[0065] Figure 2 A schematic flow chart of another embodiment of the grayscale compensation method of high-order polynomial fitting provided in the present application;

[0066] Figure 3 This is a schematic diagram of the structure of an embodiment of a grayscale compensation device for high-order polynomial fitting provided in this application;

[0067] Figure 4 This is a schematic structural diagram of another embodiment of a grayscale compensation device for high-order polynomial fitting provided in the present application. DETAILED DESCRIPTION

[0068] It should be noted that the grayscale compensation method of high-order polynomial fitting provided in this application can be applied to a terminal or a system, and can also be applied to a server. For example, the terminal can be a smart phone or a computer, a tablet computer, a smart TV, a smart watch, a portable computer terminal, or a fixed terminal such as a desktop computer. For the convenience of explanation, this application uses the terminal as an example for explanation.

[0069] See also Figure 1 The present application first provides an embodiment of a grayscale compensation method of high-order polynomial fitting, which includes:

[0070] S101, determining a target high-order polynomial, wherein the high-order polynomial is a relationship between a grayscale value and a brightness value of a display device, wherein a variable value represents a grayscale value, and a function value represents a brightness value;

[0071] First, a relationship model between the grayscale value and the brightness value of the display device is established. Multiple grayscale values ​​and corresponding brightness value data can be collected through display device testing. A high-order polynomial fitting tool (such as the least squares method) is used to establish a target polynomial equation. In one implementation, the polynomial equation can be:

[0072] ;

[0073] In this formula: a n 、a n-1 , ..., a1, a0 are known coefficients, n is the order of the polynomial equation, and x is the variable to be solved.

[0074] In this formula, the variable value x represents the grayscale value, and the function value Represents the brightness value, and the high-order polynomial equation obtained in this way serves as the input model for subsequent calculations.

[0075] S102, based on the initial iteration point, using Newton's method to iterate the target high-order polynomial, and calculate the next iteration point;

[0076] In this step, the grayscale value that meets the target brightness is solved by Newton's method.

[0077] For a given target brightness value, the high-order polynomial equation is transformed into a zero-point problem, where is the function value of a high-order polynomial, x is the grayscale value, and the function value Indicates the difference between the current brightness and the target brightness.

[0078] First, select an initial grayscale value x0 as the initial iteration point. The selection of x0 can be based on empirical values, historical data, or an intermediate grayscale value (such as 128 for 8-bit grayscale). The selection of the initial iteration point has a great influence on the iteration convergence speed and stability.

[0079] Use Newton's method iteration formula to calculate the next iteration point. The iteration formula of Newton's method is:

[0080] ;

[0081] in, yes The formula updates the solution value by estimating the linear approximation of the current point, iterating continuously and gradually approaching the root of the polynomial equation.

[0082] According to the Newton method formula, update the grayscale value and convert the newly calculated x k+1It is used as the input of the next iteration to determine whether the current iteration result is close to convergence. If the convergence condition is met (such as the function value is close to zero or the error of consecutive iteration points is small enough), the iteration ends. If not, enter the next iteration and repeat the above steps.

[0083] In each iteration, record the current iteration value x k and its corresponding function value and the first-order derivative value ,Based on the recorded historical iteration results, the convergence and change trend of the iteration process are analyzed to determine whether the iteration has abnormal fluctuations or tends to a singularity.

[0084] S103, after each iteration, determining whether the obtained iteration point meets a preset singularity condition;

[0085] After each iteration, it is determined whether the current iteration point is at a singular point position to prevent the algorithm from failing or the calculation from diverging due to the singular point problem, and to ensure the stability and convergence of the iteration process. In this step, it is necessary to detect and process the singular point problem to prevent the algorithm from failing at a special point. If it is confirmed that the preset singular point condition is met, step S105 is executed to enter the singular point processing mechanism.

[0086] In numerical computing, a singularity generally refers to one of the following situations:

[0087] The derivative approaches zero: the derivative of the target polynomial at the current iteration point The value is close to zero. At this time, the denominator of the Newton method update formula will approach infinity, resulting in unstable or divergent iterative results.

[0088] Abnormal function value changes: the target polynomial function value at the current point Sudden changes or abnormal deviations may cause the algorithm to fail.

[0089] After each iteration, the derivative value and function value of the current iteration point are tested to determine whether the preset singularity conditions are met.

[0090] S104, detecting a change trend of the first distance value and the first function value;

[0091] The first distance value represents the distance between the current function value and the function value of the previous iteration, and the first function value is the current function value.

[0092] In each iteration process, the changing trend of the iteration points is dynamically monitored, including the change in the distance between two adjacent iteration points (i.e., the first distance value) and the change in the value of the target high-order polynomial function (i.e., the first function value). By analyzing these trends, it is determined whether the iteration process tends to converge or has the risk of divergence.

[0093] Definition and calculation of the first distance value:

[0094] The first distance value d k Indicates the current iteration point x k The previous iteration point x k+1 The absolute distance between them is calculated as:

[0095] d k =|x k -x k+1 ∣

[0096] In one implementation, if d k If d continues to decrease, it indicates that the iteration point is gradually converging. k If it continues to increase, it indicates that the iteration may have a divergent trend. |Continuously decreases and tends to zero, indicating that the target point gradually approaches the true solution.

[0097] If | |If the fluctuation is obvious or increases, it indicates that the iteration process is unstable.

[0098] According to the trend analysis results, corresponding adjustment measures are taken. For example, if the trend shows convergence, the next iteration is continued. If a divergent trend is detected, the divergence processing mechanism is triggered to correct the iteration process by adjusting the step size, adaptive parameter optimization or multiple initial value solving.

[0099] S105, when it is determined that the preset singularity condition is met, adjusting the iteration process through the singularity adjustment mechanism;

[0100] During the grayscale compensation process, the target high-order polynomial may have singularity regions. Singularity is usually manifested as a point where the function value changes dramatically, the slope approaches infinity or approaches zero, which may cause iteration failure or algorithm failure to converge. Therefore, it is necessary to effectively jump out of the singularity region through an adjustment mechanism to ensure the stability of iteration and calculation accuracy.

[0101] After the singularity is detected, the adjustment mechanism is immediately triggered to execute step S106 to prevent the iteration from stagnating or diverging in the singularity area. The adjustment mechanism includes the following methods:

[0102] According to the current derivative size or function value change trend, adjust the step size factor to reduce the iteration step size:

[0103] For example:

[0104] If the derivative is too small, reduce the step size factor to avoid skipping the solution due to too large a step size. If the derivative is too large, increase the step size factor to avoid repeated iterations in steep areas. When it is difficult to jump out of the singularity area, you can also change the current iteration point, select a new initial point farther away from the singularity area, and restart the iteration. The new initial point can be set by the incremental method. For example:

[0105] x k new =x k ±Δx;

[0106] Among them, x k new represents the new initial iteration point, x k represents the current iteration point, and Δx represents the compensation increment.

[0107] Interpolation can also be used for substitution. For example, a low-order polynomial or linear interpolation can be used to locally substitute the target high-order polynomial in the area near the singular point. The substitution function Q(x) satisfies the compatibility with the high-order polynomial near the singular point, and the formula is as follows:

[0108] Q(x)≈ for x∈[x k -Δ, x k +Δ];

[0109] The substitution function is used for local iteration to recover the original high-order polynomial after escaping from the singularity region.

[0110] After singularity detection, multiple initial values ​​are used for parallel solution, and the convergence of each path is analyzed. The result far away from the singularity and with good convergence is preferentially selected as the final solution. After the adjustment mechanism is completed, the Newton method iteration is continued based on the target high-order polynomial to ensure the global consistency of the results.

[0111] In this step, after effectively jumping out of the singularity region through the adjustment mechanism, the iteration process is resumed to ensure the normal operation of the Newton method and the global convergence of the target high-order polynomial solution. This process improves the adaptability and stability of the grayscale compensation method.

[0112] S106, determining that the change trend meets a preset change condition, and adjusting the iterative process through a divergent processing mechanism;

[0113] In the iterative process of grayscale compensation, the Newton method may diverge due to the selection of the initial point, complex function characteristics (such as drastic changes in certain areas of high-order polynomials), or other numerical instabilities. In order to suppress the divergence, the present application redirects the iterative regression convergence track through an adjustment mechanism.

[0114] After each iteration, some preset conditions are used to determine whether divergence occurs. In the case of divergence, a multi-initial value parallel solution strategy is used to redirect the iteration, for example:

[0115] Multiple initial points are uniformly generated near the current iteration point xk for parallel exploration. Generation methods include using fixed incremental offsets or adding random perturbations. For each initial point, Newton iteration is run independently, and the iteration results of each path are recorded, including the final iteration point and the convergence rate, as well as the convergence stability. Convergence stability refers to whether the amplitude of each iteration point change is stable. The optimal final solution is selected according to some rules, such as selecting a path that meets the preset convergence conditions (such as a function value less than a threshold). If multiple paths converge, the one with the largest convergence rate can be selected. If no path converges, the point closest to the target can be selected, that is, among all paths, the point with the smallest absolute value of the function value is selected as the solution.

[0116] In this step, the multi-initial value parallel solution strategy can effectively avoid the divergence problem caused by a single initial point, improving the robustness and success rate of the algorithm. The final iteration point is not only more stable, but also can further improve the compensation accuracy to meet the grayscale adjustment requirements of the display device.

[0117] It should be noted that after completing the judgment process of step S103 and the detection process of step S104, if it is determined that neither the preset singularity condition is satisfied nor the change trend satisfies the preset change condition, step S110 is executed to continue the iteration process. If one of the conditions is satisfied, such as satisfying the singularity condition or satisfying the change condition, it is necessary to enter the corresponding adjustment mechanism for adjustment, and process based on the adjusted iteration process.

[0118] S107, judging whether the adjusted iteration process satisfies a preset convergence condition;

[0119] This step is used to determine whether the currently adjusted iteration process has met the convergence criteria of the algorithm to determine whether to output the final result (convergence point). If it is determined to be satisfied, step S108 is executed. Reasonable setting of the convergence condition can effectively avoid the iteration from falling into an invalid loop, while ensuring the accuracy and stability of the iteration result.

[0120] The preset convergence condition is the basis for judging whether the iteration is finished, which can include:

[0121] The function value is close to zero, that is, the polynomial value corresponding to the current iteration point is close to zero. Or the difference between the current iteration point and the previous iteration point is less than the preset step threshold, or the maximum number of iterations has been reached. Specifically, it can be to judge the accuracy of the current function value or the amplitude of the change of the iteration point and check whether the current number of iterations exceeds the limit.

[0122] In some special cases, the function value error tolerance threshold and the grayscale value change error tolerance threshold can be dynamically adjusted according to the number of iterations, so that the convergence condition is more suitable for the specific polynomial characteristics. The function value error tolerance threshold is used to measure whether the function value of the target high-order polynomial at the current iteration point is close enough to zero. It represents the allowable error between the brightness value and the target brightness value. Specifically, the function value error tolerance threshold is a small positive number. When the value of the polynomial at the iteration point is less than or equal to the function value error tolerance threshold, it means that the function value is close enough to zero and meets the convergence condition, indicating that the grayscale compensation is accurate enough.

[0123] The grayscale value change error tolerance threshold is used to determine whether the change between the current iteration point and the previous iteration point is small enough. It measures the change of the grayscale value (i.e., the independent variable of the target high-order polynomial). k -x k-1 When ∣≤the grayscale value change error tolerance threshold, it means that the grayscale value change is very small, indicating that the iterative process tends to be stable and can be considered to be converged.

[0124] In an optional implementation, if the singularity condition is met, the system will preferentially enter the singularity adjustment mechanism. The singularity adjustment mechanism is used to adjust the current iteration process to avoid convergence failure or calculation anomaly in the Newton method iteration near the singularity. After the adjustment is completed, the subsequent steps are re-executed based on the new iteration point.

[0125] The singularity adjustment mechanism includes but is not limited to the following operations: introducing a step size factor for step size control, adding a random disturbance term to correct the iteration point, or directly updating the initial iteration point.

[0126] If the singularity condition is not met, but the change condition is met, the system will enter the divergence processing mechanism. This mechanism is used to adjust the current iteration method to avoid divergence problems in the iteration process. After the adjustment, the subsequent iterations are also continued based on the new iteration point.

[0127] The divergence processing mechanism includes: adjusting the size of the step factor to alleviate the divergence trend, or recalculating with multiple different initial iteration points to ensure the stability and convergence of the iteration.

[0128] Regardless of whether it is the singularity adjustment mechanism or the divergence processing mechanism, after the adjustment is completed, the system will perform subsequent calculations based on the adjusted iteration point.

[0129] The adjusted iteration process will continue to execute the judgment process and detection process of S103 and S104 to ensure that the iteration point meets the convergence condition of the system or triggers other adjustment mechanisms, and finally outputs the convergence point.

[0130] In actual operation, if both the singularity condition and the change condition are met at the same time, the system will prioritize the singularity condition. This is because the singularity condition usually leads to large calculation errors or instability, and prioritizing the adjustment of the singularity can more effectively ensure the accuracy and stability of the iteration process. The change condition is processed after the singularity condition is processed.

[0131] S108, if it is determined that the condition is satisfied, then a convergence point is output, where the convergence point represents a compensated grayscale value;

[0132] If the current iteration point is determined to be the final convergence point, its output is used as the compensated grayscale value.

[0133] S109: Compensate the grayscale of the display device according to the convergence point.

[0134] In this step, based on the convergence point (i.e., grayscale compensation value) obtained by iterative solution, the input grayscale of the display device is adjusted so that its output brightness is closer to the target brightness value, thereby achieving high-precision brightness compensation.

[0135] Specifically, a corresponding relationship is established between the convergence point (compensated grayscale value) obtained in the iterative process and the target brightness value. The convergence point represents the corrected grayscale value that needs to be input at the target brightness. A lookup table (LUT) can be maintained in the display device, through which the compensated grayscale value can be quickly found. The compensated grayscale values ​​corresponding to all target brightness values ​​are stored in the correction table for subsequent real-time grayscale adjustment. The compensation data is stored in the driver chip or memory of the display device for the display module to call. In the actual operation process, when the display device needs to display a certain target brightness value, the corresponding compensated grayscale value is found through the lookup table or algorithm calculation. The compensated grayscale value is sent to the display control module as an input value to adjust the actual brightness output. After the compensation data is applied, the actual brightness of the display screen is measured by an optical sensor and compared with the target brightness value. If there is still an error, the compensation data can be iteratively optimized again, and the lookup table can be updated to ensure the accuracy stability in long-term use.

[0136] For example, when the target brightness is 50, after high-order polynomial fitting and Newton method iteration, the convergence point x is calculated. k =128. At this time, the grayscale value 128 is input to the display device, and the actual brightness output will be closer to the target value 50 more accurately.

[0137] This compensation processing can significantly improve the brightness consistency of the display device, especially maintaining high-quality display effects under screen aging or external environmental influences (such as temperature and humidity changes).

[0138] In this embodiment, the nonlinear relationship between the brightness and grayscale of the display device is effectively overcome through compensation processing, the brightness drift problem is reduced, and the grayscale accuracy of the display device is improved, especially in the high brightness range and low brightness range, avoiding the phenomenon of uneven or inaccurate brightness.

[0139] S110, continue the iterative process.

[0140] In the process, S110 controls and continues the entire iteration process. This step is mainly used to continue iterating according to the adjustment results of the previous steps when the convergence condition is not met, until the convergence condition is finally met or the upper limit of the number of iterations is reached.

[0141] The execution of step S110 is based on the premise that the convergence condition is not met, which is triggered by the following two situations:

[0142] The adjusted iteration still does not meet the convergence condition (“No” in S107 ).

[0143] The preset singularity condition or change condition is not met (the judgments of S103 and S104 are both “No”).

[0144] At this point, the process will return to the iteration part and continue to calculate new iteration points according to Newton's method.

[0145] Based on the previous iteration, a new initial iteration point is generated using the aforementioned adjustment mechanism (such as singularity adjustment mechanism or divergence processing mechanism). This initial iteration point will serve as the starting point of the current iteration round.

[0146] After the iteration, it is necessary to recalculate and determine whether it is close to the target (for example, whether it is less than the function threshold). It is also necessary to determine whether the distance between the current point and the previous iteration point meets the convergence accuracy requirements.

[0147] If the function value changes in an abnormal trend (for example, it increases repeatedly or the range of change exceeds expectations), it is necessary to determine whether there is a divergence risk and use the divergence processing mechanism to make adjustments.

[0148] It should be noted that the loop of step S110 will not be performed indefinitely, but the iterative process will be exited based on one of the following termination conditions:

[0149] 1. Meeting the Convergence Condition When the preset convergence condition is met (the determination of S107 is “yes”), the process will enter the subsequent step S109 to perform compensation processing on the grayscale value.

[0150] 2. Reaching the upper limit of iterations

[0151] To prevent an infinite loop, you can set a maximum number of iterations (for example, 100) and force the program to exit when the upper limit is reached, returning a failure message.

[0152] 3. Divergence or calculation anomalies

[0153] If it is detected in a round of iteration that the function value cannot converge or the derivative value is zero (which may cause a division by zero error in Newton's method), the process can be terminated and enter exception handling.

[0154] In an optional embodiment, the preset singularity condition includes:

[0155] Condition 1: The iterative point distance is less than the first threshold for N consecutive times, and the corresponding function value is greater than the second threshold;

[0156] or;

[0157] Condition 2: The first-order derivative value of the corresponding function value is less than a third threshold value, and the third threshold value is a value tending to zero.

[0158] In this embodiment, in order to accurately determine whether the iteration process has entered the singularity region, two preset singularity conditions are defined. By monitoring the iteration point distance and function value characteristics of the iteration process, it is ensured that necessary adjustment measures are taken in the singularity region.

[0159] For condition 1:

[0160] The iteration point distance of the iteration is smaller than the first threshold value for N consecutive times, and the corresponding function value is larger than the second threshold value.

[0161] Iteration point distance |x k -x k-1 ∣ represents the distance between the current iteration point and the previous iteration point.

[0162] If the change is smaller than the preset first threshold value N times in a row, it means that the iteration step is gradually decreasing.

[0163] At the same time, the function value Indicates the function value of the current iteration point on the target polynomial. If the value is still greater than the second threshold, it indicates that the iteration has failed to effectively reduce the function value.

[0164] This situation indicates that the iteration point has "stagnated" in a narrow area and failed to converge quickly to the target solution. The possible reason is that the curvature of the polynomial changes dramatically or the current point is close to a singular point.

[0165] Specifically, after each iteration, ∣x is calculated k -x k-1 ∣ and .

[0166] Record the change in N consecutive iterations if the following conditions are met: |x k -x k-1 | is less than the first threshold and If it is greater than the second threshold, it is determined that the current iteration has entered the singularity region.

[0167] For condition 2: the first-order derivative value of the corresponding function value is less than the third threshold

[0168] First derivative of a function value Indicates the current iteration point x k The gradient size.

[0169] If the derivative value | ∣ is less than the preset third threshold (close to 0), it indicates that the current point may be an inflection point or extreme point of a high-order polynomial.

[0170] This situation usually occurs in the flat area of ​​high-order polynomials or in the area where the gradient changes sharply, causing the step size adjustment of Newton's method to fail. Although the derivative value is close to zero, the objective function value It is still large, indicating that the current point is not a solution.

[0171] Specifically, the first-order derivative of the function is calculated after each iteration .

[0172] If the following conditions are met:

[0173] ∣ |If it is less than the third threshold, it is judged that the current iteration has entered the singularity region.

[0174] If any singularity condition (condition 1 or condition 2) is met, the current point is determined to be a singularity. The judgment of the singularity condition has a certain degree of flexibility, and the sizes of the above three thresholds can be dynamically adjusted according to actual needs. If the singularity condition is met, the singularity area is jumped out through the singularity adjustment mechanism (such as adaptive step size adjustment or multi-initial value parallel solution). After adjustment, the iterative process re-enters the normal calculation.

[0175] In this optional embodiment, multiple conditions are introduced to judge the singularity region, thereby improving the robustness of the algorithm, preventing the iteration from failing or staying too long at a special point, and accelerating the convergence speed. The flexible definition of conditions can adapt to complex scenarios with multiple display devices and high-order polynomial characteristics.

[0176] In an optional embodiment, adjusting the iteration process by a singularity adjustment mechanism includes:

[0177] A step factor is introduced into the iterative formula of the Newton method, and the value of the step factor is adjusted. The value range of the step factor is 0 to 1.

[0178] In this embodiment, in order to effectively jump out of the singularity region, the iteration step length is adjusted by introducing a step length factor into the iteration formula of the Newton method. The step length factor has a value range of 0 to 1 and is used to dynamically shorten or extend the step length to avoid falling into the singularity region.

[0179] After introducing the step size factor λ, the iterative formula of Newton's method is:

[0180] ;

[0181] Where λ is the adaptive step size adjustment factor, and its value is 0<λ<1.

[0182] In another implementation, a disturbance term may be introduced. When the step size adjustment is invalid, a random disturbance term σ is introduced near the current iteration point, which is expressed as:

[0183]

[0184] Where σ is a random disturbance term (small disturbance), which is used to break the iterative stagnation and jump out of the singularity area.

[0185] Furthermore, when the above two strategies still cannot solve the problem, a new initial iteration point x0 is selected and the iteration process is restarted. Especially when solving multiple roots, this initial iteration point restart mechanism can better avoid falling into local extreme values ​​or singularities.

[0186] In an optional embodiment, the preset change condition includes:

[0187] Condition 3: The distance between iteration points gradually increases;

[0188] or;

[0189] Condition 4: The reduction amplitude of the iteration point distance is not greater than a preset amplitude threshold and the current function value is greater than a preset function threshold.

[0190] In this embodiment, the preset change condition is used to detect abnormal trends in the iteration process, helping to timely discover and deal with situations that may lead to divergence or iteration failure. The distance value and function value of the iteration result are judged by conditions 3 and 4 to ensure that the iteration process remains within a controllable range.

[0191] For condition 3: the distance between iteration points gradually increases

[0192] The iteration point distance refers to the distance between the current iteration point and the previous iteration point:

[0193] Condition 3 is used to detect whether a divergent trend occurs during the iteration process, that is, the distance value gradually increases.

[0194] In each iteration, the current distance value is calculated and compared with the previous distance value:

[0195] If multiple times d k >d k-1 , that is, the distance value gradually increases, then condition 3 is satisfied.

[0196] When condition 3 is met, the current iteration can be paused and the divergence processing mechanism can be started, such as adjusting the initial point, reducing the step size factor, or starting multiple initial iteration points in parallel for iteration.

[0197] For condition 4: The resulting distance value is not reduced enough and the function value is too large

[0198] Condition 4 is used to detect whether the decreasing trend is obvious during the iteration process, and to monitor whether the function value is still too large. It specifically includes two judgments:

[0199] The decrease amplitude of the current distance value does not exceed the preset amplitude threshold, and whether the current function value is greater than the preset function threshold.

[0200] Specifically, the current distance value and function value are calculated. Compare whether the decrease in the distance value is lower than a preset threshold. Determine whether the function value is still too large, indicating that the target solution is not yet approached.

[0201] When condition 4 is met, the divergence processing or singularity adjustment mechanism can be triggered. The step size factor is dynamically adjusted to shorten the step size to enhance the adjustment accuracy. If there is still no improvement, the iteration process can be restarted or other optimization methods can be selected (for example, finding a better solution through multi-initial value parallel calculation).

[0202] In this embodiment, the direction and accuracy of the iteration process are ensured to be within the control range through dual monitoring of the distance value and the function value. Condition 3 can effectively detect the divergence trend, and condition 4 can identify the slow convergence speed or the wrong direction, thereby avoiding iteration failure. The response mechanism flexibly selects adjustment methods according to different abnormality types, such as adjusting the step size factor or restarting the iteration, to ensure the reliability and efficiency of the algorithm.

[0203] In a more specific implementation, the adjusting the iterative process by the divergent processing mechanism includes:

[0204] Introducing a step factor into the iterative formula of the Newton method and adjusting the value of the step factor, wherein the value range of the step factor is 0 to 1;

[0205] or;

[0206] Iterate using multiple different initial iteration points.

[0207] In this more specific implementation, two methods are provided to deal with the divergence problem during the iteration process to ensure the convergence and accuracy of the final result.

[0208] One way is to introduce a step size factor into the iterative formula of Newton's method and adjust the step size factor. After introducing the step size factor λ, the iterative formula of Newton's method is:

[0209] ;

[0210] Where λ is the adaptive step size adjustment factor, and its value is 0<λ<1.

[0211] Another way is to use multiple different initial iteration points for iteration. Divergence may be caused by improper selection of initial iteration points. To solve this problem, multiple initial iteration points can be selected at the same time and parallel calculations can be performed separately.

[0212] In an optional embodiment, the preset convergence condition includes:

[0213] Condition 5: The iteration point distance is less than a preset first precision threshold, and the function value is less than a preset second precision threshold.

[0214] In this embodiment, the convergence of the iteration process is determined by setting condition 5, thereby ensuring the stability and accuracy of the grayscale compensation method.

[0215] In condition 5, the first precision threshold is used to measure whether the point distance between two iterations is close enough, and the second precision threshold is used to measure whether the current function value is close to zero. If the convergence condition is met, the current iteration point is defined as the convergence point. If the condition is not met within the preset maximum number of iterations (such as too many iterations but no convergence), the divergence processing mechanism can be triggered or the iteration parameters can be dynamically adjusted. In this embodiment, the two precision thresholds are limited at the same time to ensure that the final solution has both positional accuracy (accurate grayscale values) and functional accuracy (brightness value matching). At the same time, it can also prevent the algorithm from performing meaningless additional iterations due to incorrect judgment of the convergence conditions, thereby improving the calculation efficiency.

[0216] See also Figure 2 The present application also provides another embodiment of a grayscale compensation method of high-order polynomial fitting, which includes:

[0217] S201, determining a target high-order polynomial, wherein the high-order polynomial is a relationship between a grayscale value and a brightness value of a display device, wherein a variable value represents a grayscale value, and a function value represents a brightness value;

[0218] S202, based on the initial iteration point, using Newton's method to iterate the target high-order polynomial, and calculate the next iteration point;

[0219] S203, after each iteration, determining whether the obtained iteration point meets a preset singularity condition;

[0220] S204, detecting a change trend of the first distance value and the first function value;

[0221] S205, determining that a preset singularity condition is met, and adjusting the iteration process through a singularity adjustment mechanism;

[0222] S206, determining that the change trend meets a preset change condition, and adjusting the iterative process through a divergent processing mechanism;

[0223] S207, judging whether the adjusted iteration process satisfies a preset convergence condition;

[0224] S208, if it is determined that the condition is satisfied, then a convergence point is output, where the convergence point represents a compensated grayscale value;

[0225] S209, establishing a mapping table between grayscale values ​​and actual brightness values ​​according to the convergence point, and recording the compensation value that needs to be adjusted for each grayscale value;

[0226] The convergence point is the compensated grayscale value obtained through high-order polynomial fitting and iterative optimization, which can effectively correct the deviation between the grayscale value of the display device and the target brightness value.

[0227] Traverse the grayscale value range supported by the display device (for example, integer values ​​from 0 to 255). For each grayscale value G, obtain the actual output brightness value through measurement or model calculation. Generate the target brightness value according to the requirements or design standards of the display device. For each grayscale value G, calculate the compensated grayscale value Gcompensated. Record the correspondence between G and Gcompensated in a mapping table: MappingTable={(G,Gcompensated)} This table is used in the subsequent steps of grayscale compensation.

[0228] S210, for each pixel of the display device, generating a calibrated grayscale value using the convergence point;

[0229] For the input grayscale value Ginput of each pixel in the display device, the corresponding compensation grayscale value Gcompensated is found through the mapping table. For different pixels, due to different process errors and display characteristics, the compensation value in the mapping table needs to be fine-tuned in combination with the calibration data of the specific pixel. For example, based on factory calibration or point-by-point brightness measurement, a personalized compensation value is generated. Traverse all pixels of the display device:

[0230] For each pixel point P(i,j), obtain its input grayscale value Ginput. Find the corresponding Gcompensated from the mapping table. Replace the input grayscale value with the compensation value Gcompensated to generate a calibrated grayscale signal.

[0231] S211 , inputting the compensated grayscale value as a driving signal into a display device, so that the pixel points of the display device output corresponding brightness values.

[0232] The calibrated grayscale values ​​Gcompensated of all pixels are converted into driving signals, which are the voltage or current signals required by the internal circuit of the display device to drive the pixels. The calibration signal is distributed to the driving circuit of each pixel through the control system of the display device. The pixel outputs the corresponding brightness value according to the input signal. The compensated grayscale value ensures that the brightness distribution of the display device meets the design requirements, improves the image quality, and reduces the display problems caused by grayscale inconsistency.

[0233] In this embodiment, the compensation method based on the convergence point and the mapping table can realize point-by-point calibration and improve the accuracy and consistency of the display device. The mapping table can be flexibly adjusted according to the characteristics of the device and is suitable for display devices with different resolutions and technical parameters. Through the search of the mapping table and the generation of the driving signal, the compensation process can be fully automated, improving efficiency and reducing manual intervention.

[0234] S212, continue the iterative process.

[0235] The above describes an embodiment of the method provided in the present application. The following describes an embodiment of the device provided in the present application.

[0236] See also Figure 3 The present application provides an embodiment of a grayscale compensation device for high-order polynomial fitting, and the device of this embodiment includes:

[0237] A polynomial determination unit 301 is used to determine a target high-order polynomial, wherein the high-order polynomial is a relationship between a grayscale value and a brightness value of a display device, wherein a variable value represents a grayscale value and a function value represents a brightness value;

[0238] An iteration unit 302, configured to iterate the target high-order polynomial based on an initial iteration point using Newton's method and calculate a next iteration point;

[0239] The singularity judgment unit 303 is used to judge whether the obtained iteration point meets the preset singularity condition after each iteration;

[0240] and\or;

[0241] A divergence detection unit 304, used to detect a change trend of the first distance value and the first function value;

[0242] A singularity adjustment unit 305, configured to adjust the iteration process through a singularity adjustment mechanism when it is determined that a preset singularity condition is met;

[0243] A divergence adjustment unit 306, configured to adjust the iterative process through a divergence processing mechanism when it is determined that the change trend meets a preset change condition;

[0244] A convergence judgment unit 307, used to judge whether the adjusted iterative process meets a preset convergence condition;

[0245] A convergence output unit 308, configured to output a convergence point if it is determined that the condition is satisfied, wherein the convergence point represents a compensated grayscale value;

[0246] The grayscale compensation unit 309 is used to compensate the grayscale of the display device according to the convergence point.

[0247] Optionally, the preset singularity condition includes:

[0248] Condition 1: The iterative point distance is less than the first threshold for N consecutive times, and the corresponding function value is greater than the second threshold;

[0249] or;

[0250] Condition 2: The first-order derivative value of the corresponding function value is less than a third threshold value, and the third threshold value is a value tending to zero.

[0251] Optionally, the singularity adjustment unit 305 is specifically used for:

[0252] A step factor is introduced into the iterative formula of the Newton method, and the value of the step factor is adjusted. The value range of the step factor is 0 to 1.

[0253] Optionally, the singularity adjustment unit 305 is specifically used for:

[0254] A random disturbance term is introduced into the iterative formula of the Newton method, and the random disturbance term is used to adjust the current iteration point.

[0255] Optionally, the singularity adjustment unit 305 is specifically used for:

[0256] The initial iteration point is updated.

[0257] Optionally, the preset change condition includes:

[0258] Condition 3: The distance between iteration points gradually increases;

[0259] or;

[0260] Condition 4: The reduction amplitude of the iteration point distance is not greater than a preset amplitude threshold and the current function value is greater than a preset function threshold.

[0261] Optionally, the divergence adjustment unit 306 is specifically configured to:

[0262] Introducing a step factor into the iterative formula of the Newton method and adjusting the value of the step factor, wherein the value range of the step factor is 0 to 1;

[0263] or;

[0264] Iterate using multiple different initial iteration points.

[0265] Optionally, the preset convergence condition includes:

[0266] Condition 5: The iteration point distance is less than a preset first precision threshold, and the function value is less than a preset second precision threshold.

[0267] The preset convergence conditions include:

[0268] Condition 5: The iteration point distance is less than a preset first precision threshold, and the function value is less than a preset second precision threshold.

[0269] Optionally, the grayscale compensation unit 309 is specifically used for:

[0270] According to the convergence point, a mapping table is established between the grayscale value and the actual brightness value, and the compensation value that needs to be adjusted for each grayscale value is recorded;

[0271] For each pixel of the display device, generating a calibrated grayscale value using the convergence point;

[0272] The compensated grayscale value is input into the display device as a driving signal, so that the pixel points of the display device output corresponding brightness values.

[0273] See also Figure 4 The present application also provides a grayscale compensation device for high-order polynomial fitting, comprising:

[0274] Processor 401, memory 402, input and output unit 403, bus 404;

[0275] The processor 401 is connected to the memory 402, the input and output unit 403 and the bus 404;

[0276] The memory 402 stores a program, and the processor 401 calls the program to execute any of the above methods.

[0277] The present application also relates to a computer-readable storage medium on which a program is stored, wherein when the program is run on a computer, the computer is caused to execute any of the above methods.

[0278] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0279] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0280] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0281] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0282] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), disk or optical disk and other media that can store program code.

Claims

1. A grayscale compensation method based on high-order polynomial fitting, characterized in that: The method comprises: Determine a target high-order polynomial, wherein the high-order polynomial is a relationship between a grayscale value and a brightness value of a display device, wherein a variable value represents a grayscale value, and a function value represents a brightness value; Based on the initial iteration point, the target high-order polynomial is iterated using Newton's method, and the next iteration point is calculated; After each iteration, it is determined whether the obtained iteration point meets the preset singularity condition; and\or; Detecting a change trend of a first distance value and a first function value, wherein the first distance value represents an absolute distance between a current iteration point and a previous iteration point; When it is determined that the preset singularity conditions are met, the iterative process is adjusted through the singularity adjustment mechanism; When it is determined that the change trend meets the preset change condition, the iterative process is adjusted through a divergent processing mechanism; Determine whether the adjusted iteration process meets the preset convergence condition; If it is determined that the condition is satisfied, a convergence point is output, where the convergence point represents a compensated grayscale value; The grayscale of the display device is compensated according to the convergence point.

2. The grayscale compensation method of high-order polynomial fitting according to claim 1, characterized in that: The preset singularity conditions include: Condition 1: The iterative point distance is less than the first threshold for N consecutive times, and the corresponding function value is greater than the second threshold; or; Condition 2: The first-order derivative value of the corresponding function value is less than a third threshold value, and the third threshold value is a value tending to zero.

3. The grayscale compensation method of high-order polynomial fitting according to claim 2, characterized in that: The adjusting of the iteration process by the singularity adjustment mechanism includes: A step factor is introduced into the iterative formula of the Newton method, and the value of the step factor is adjusted. The value range of the step factor is 0 to 1.

4. The grayscale compensation method of high-order polynomial fitting according to claim 2, characterized in that: The adjusting of the iteration process by the singularity adjustment mechanism includes: A random disturbance term is introduced into the iterative formula of the Newton method, and the random disturbance term is used to adjust the current iteration point.

5. The grayscale compensation method of high-order polynomial fitting according to claim 2, characterized in that: The adjusting of the iteration process by the singularity adjustment mechanism includes: The initial iteration point is updated.

6. The grayscale compensation method of high-order polynomial fitting according to claim 1, characterized in that: The preset change conditions include: Condition 3: The distance between iteration points gradually increases; or; Condition 4: The reduction amplitude of the iteration point distance is not greater than a preset amplitude threshold and the current function value is greater than a preset function threshold.

7. The grayscale compensation method of high-order polynomial fitting according to claim 6, characterized in that: The adjustment of the iterative process by the divergent processing mechanism includes: Introducing a step factor into the iterative formula of the Newton method and adjusting the value of the step factor, wherein the value range of the step factor is 0 to 1; or; Iterate using multiple different initial iteration points.

8. The grayscale compensation method of high-order polynomial fitting according to claim 1, characterized in that: The preset convergence conditions include: Condition 5: The iteration point distance is less than a preset first precision threshold, and the function value is less than a preset second precision threshold.

9. The grayscale compensation method of high-order polynomial fitting according to claim 1, characterized in that: The grayscale compensation processing of the display device according to the convergence point includes: According to the convergence point, a mapping table is established between the grayscale value and the actual brightness value, and the compensation value that needs to be adjusted for each grayscale value is recorded; For each pixel of the display device, generating a compensated grayscale value using the convergence point; The compensated grayscale value is input into the display device as a driving signal, so that the pixel points of the display device output corresponding brightness values.

10. A grayscale compensation device using high-order polynomial fitting, characterized in that: include: A polynomial determination unit, configured to determine a target high-order polynomial, wherein the high-order polynomial is a relationship between a grayscale value and a brightness value of a display device, wherein a variable value represents a grayscale value, and a function value represents a brightness value; An iteration unit, used for iterating the target high-order polynomial based on an initial iteration point using Newton's method and calculating a next iteration point; A singularity judgment unit, used to judge whether the obtained iteration point meets the preset singularity condition after each iteration; and\or; A divergence detection unit, used to detect a change trend of a first distance value and a first function value, wherein the first distance value represents an absolute distance between a current iteration point and a previous iteration point; A singularity adjustment unit, used to adjust the iterative process through a singularity adjustment mechanism when it is determined that a preset singularity condition is met; A divergence adjustment unit, configured to adjust the iterative process through a divergence processing mechanism when it is determined that the change trend meets a preset change condition; A convergence judgment unit, used to judge whether the adjusted iterative process meets the preset convergence condition; A convergence output unit, configured to output a convergence point if it is determined that the condition is satisfied, wherein the convergence point represents a compensated grayscale value; The grayscale compensation unit is used to compensate the grayscale of the display device according to the convergence point.

11. A grayscale compensation device using high-order polynomial fitting, characterized in that: The device comprises: Processor, memory, input-output unit, and bus; The processor is connected to the memory, the input and output unit, and the bus; The memory stores a program, and the processor calls the program to execute the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a program stored thereon, wherein the program, when executed on a computer, performs the method according to any one of claims 1 to 9.

Images