Full-automatic correction method and device for LED display screen

Abstract

The application provides a full-automatic correction method and device for an LED display screen, and belongs to the technical field of display screen correction; solves the problems of low efficiency, limited precision, and inability to dynamically adjust in the existing LED display screen correction method; and comprises the following steps: starting correction, including: system initialization; user setting target; connecting correction equipment and verifying whether the equipment is successfully connected; initial data acquisition; AI pre-analysis; main correction, including: parameter adjustment: preliminarily adjusting the display screen parameters according to the initial correction coefficient and judging whether the standard is reached; performing multiple iterations on the display screen parameters that do not reach the standard; double-dimension judgment: judging whether the correction coefficient after each iteration can reach the standard, and predicting whether it is possible to exceed the target parameter through a super-solution prediction model; outputting the best parameters after multiple iterations, and generating two schemes; and the application is applied to the LED display screen.

Description

Technical Field

[0001] This application relates to the field of display screen calibration technology, and in particular to a fully automatic calibration method and apparatus for LED displays. Background Technology

[0002] During the installation phase, LED display calibration is essential to ensure display quality. However, current calibration methods primarily rely on manual operation, which is inefficient and prone to errors. Furthermore, existing calibration methods cannot adjust the calibration effect in real time, making it difficult to adapt to changes in brightness and color intensity under different usage environments. Regarding the light decay of LED chips, current methods only adjust the calibration coefficient based on usage time, failing to accurately reflect actual light decay and resulting in unsatisfactory calibration results. Existing calibration methods cannot achieve differentiated adjustment of R, G, and B monochrome attenuation and compensation ratios, making it difficult to meet the demands of high-precision displays.

[0003] In summary, existing LED display calibration methods have the following drawbacks:

[0004] The calibration process is complex and requires professional personnel, which increases labor costs.

[0005] The calibration cycle is long, and it cannot quickly respond to changes in the brightness and color of the display screen;

[0006] The calibration accuracy is limited, making it difficult to achieve a display effect with high uniformity and high consistency. Summary of the Invention

[0007] To address the problems of low efficiency, limited accuracy, and inability to dynamically adjust existing LED display calibration methods, this application proposes a fully automatic calibration method and apparatus for LED displays that can meet the requirements of high-precision display.

[0008] The technical solution adopted in this application is: a fully automatic calibration method for LED displays, comprising the following steps:

[0009] S1: Begin calibration, including:

[0010] S101: System initialization;

[0011] S102: User settings: The user inputs the target parameters and selects the optimization mode according to their needs;

[0012] S103: Connect the calibration device and verify that the device is successfully connected;

[0013] S104: Initial data acquisition: The display screen's color blocks and brightness values ​​are synchronously acquired through a multispectral camera and a brightness meter from the initial unpowered state, the heating process after power-on, to the maximum heat value. The display screen's temperature value is also acquired in real time through a temperature sensor. An initial report is then generated, which includes a brightness uniformity distribution heatmap, a CIE1976 chromaticity scatter plot, and a temperature curve.

[0014] S105: AI Pre-analysis: An AI model is used to identify the luminance uniformity distribution map and the CIE1976 chromaticity scatter plot, and to generate initial correction coefficients;

[0015] S2: Main correction, including:

[0016] S201: Parameter adjustment: Make preliminary adjustments to the display parameters based on the initial correction coefficients; then proceed to step S203 to compare the adjusted display parameters with the user's target parameters to determine whether they meet the standards. If they do not meet the standards, proceed to step S202 for iterative optimization.

[0017] S202: Data Re-acquisition: If the parameters are not met after step S201, step S104 needs to be repeated for data re-acquisition, and AI pre-analysis is performed through step S105 to obtain the correction coefficient of the first iteration. The display parameters are readjusted according to the correction coefficient of the first iteration, the adjusted display parameters are obtained, and a difference report is generated by comparing them with the initial report.

[0018] S203: Two-dimensional judgment: Adjust the display parameters according to the correction coefficient after each iteration, compare the adjusted display parameters with the user-set target parameters, and determine whether they have reached the user's target parameters. If they have reached the target parameters, record the parameters of the current iteration and enter the super-optimal solution detection step. Predict whether it is possible to exceed the target parameters through the super-optimal solution prediction model. If it can exceed the target parameters, start the genetic algorithm optimization and record the final optimized parameters. If it cannot exceed the target parameters, record the parameters of the current iteration as the best parameters. If the user-set target parameters are not reached, execute the gradient descent optimization algorithm to achieve fast convergence.

[0019] S3: Output Decision: Based on the above multiple iterations, output the optimal parameters and generate two schemes.

[0020] Furthermore, in the process of multiple data re-acquisitions in step S202 and multiple iterative judgments in step S203, an anti-dead loop mechanism in step S204 is added.

[0021] Furthermore, the anti-dead-loop mechanism includes an iteration counter, convergence detection, and timeout forced exit triple anti-dead-loop mechanism. The iteration calculation can be forcibly exited when any one of these conditions is met.

[0022] Furthermore, the AI ​​model used in the AI ​​pre-analysis is a convolutional neural network.

[0023] Furthermore, the superoptimal solution prediction model employs an LSTM neural network to predict the hardware limits of the display system in the display screen.

[0024] Furthermore, the output provides two solutions: Solution A, which satisfies the user-defined parameters, and Solution B, which uses AI-recommended optimal parameters. The output of the two solutions can be visualized, and the optimal parameters can be stored.

[0025] Furthermore, the target parameters set by the user include brightness, chromaticity, and optimization control parameters. The optimization modes include standard mode and aggressive mode. The optimization control parameters include the maximum number of iterations and the convergence threshold.

[0026] Furthermore, before or after calibration, users can individually adjust the R, G, B monochrome water content and compensation ratio on the PC.

[0027] Furthermore, the parameter adjustment in step S201 is achieved by calculating the compensation matrix of each pixel using the least squares method, inputting the correction coefficients obtained in each iteration into the lookup table of the display screen, and driving the IC to adjust the PWM duty cycle and RGB mixing ratio by region to achieve the adjustment of the display screen parameters.

[0028] A computer device includes a memory, a processor, and a computer program stored in the memory, characterized in that: the processor executes the computer program to implement the steps of the method.

[0029] The advantages of this application over the prior art are as follows:

[0030] High efficiency: The fully automated calibration process greatly shortens the calibration time and improves work efficiency;

[0031] High precision: Through multiple rounds of optimization and dynamic adjustment, a display effect with high uniformity and high consistency is achieved;

[0032] Flexibility: Supports user-defined correction effect levels to adapt to the needs of different application scenarios;

[0033] Intelligent: It automatically determines the correction effect without manual intervention, reducing the difficulty of operation;

[0034] Reduced labor costs: Reduced reliance on professional correction personnel, thus lowering labor costs;

[0035] Improve equipment utilization: Rapid calibration increases equipment availability and extends equipment lifespan;

[0036] Enhancing market competitiveness: High-precision calibration technology can increase product added value and strengthen market competitiveness;

[0037] Driving industry development: It provides the LED display industry with an efficient and accurate calibration method, thus driving technological progress in the industry;

[0038] Enhance user experience: High-precision calibration can improve display quality and enhance the user's visual experience;

[0039] Energy saving and environmental protection: By optimizing the calibration process, the energy consumption of the equipment is reduced, which is in line with the concept of energy saving and environmental protection. Attached Figure Description

[0040] The following description, in conjunction with the accompanying drawings, further illustrates this application:

[0041] Figure 1 A flowchart illustrating the method provided in this application embodiment;

[0042] Figure 2 The calibration flowchart is provided for the embodiments of this application. Detailed Implementation

[0043] like Figure 1 and 2 As shown, this application provides a fully automatic calibration method for LED displays. By setting calibration effect levels, it achieves self-adjustment of the brightness and chromaticity of each pixel. Through multiple rounds of repeated operations involving uploading calibration coefficients and camera acquisition and judgment, the calibration effect is gradually optimized until the set target effect is achieved. The main steps are as follows: Setting calibration effect levels allows users to select different calibration precision and effects according to actual needs. After connecting to the calibration equipment system, the calibration program automatically starts, achieving self-adjustment of brightness and chromaticity. Multiple rounds of repeated operations involving uploading calibration coefficients and camera acquisition and judgment are used to gradually optimize the calibration effect until the set target effect is achieved. Based on the calibrated effect acquired by the camera, the brightness and chromaticity attenuation and compensation ratio are dynamically adjusted to achieve different adjustments of R, G, and B monochrome attenuation and compensation ratios. When the improvement in the calibration effect after multiple rounds is almost unchanged, the calibration task automatically ends and the final calibration effect is output.

[0044] The following is based on the appendix Figure 1 and 2 The method described in this application is explained in detail, and its specific steps are as follows:

[0045] S1: Start calibration: Includes:

[0046] S101: System Initialization: Start the system calibration program, initialize system parameters, and prepare to enter the setting process.

[0047] S102: User settings: Users input target parameters according to their needs and select optimization mode (correction accuracy level). Different optimization modes have different accuracy levels. The accuracy level will affect the number of subsequent adjustments, parameter step size and allowable error range (e.g., high accuracy requires chromaticity ΔE < 2, standard requires chromaticity ΔE < 5).

[0048] In this embodiment, the input target parameters include luminance, chromaticity ΔE, and optimization control parameters. Luminance mainly considers luminance uniformity and is expressed as a percentage, such as 95%. Luminance-related parameters include the target luminance value and the allowable tolerance range. Chromaticity-related parameters include the target chromaticity coordinates and the ΔE qualification threshold. Optimization control parameters include the maximum number of iterations and the convergence judgment threshold.

[0049] In this embodiment, there are two optimization modes: standard mode (safe mode) and aggressive mode. The aggressive mode has higher requirements for brightness uniformity accuracy and lower chromaticity ΔE than the standard mode, thereby achieving a higher level of accuracy. It can achieve the best correction effect when the hardware of the display control system is at its limit. For example, in the standard mode, it is necessary to ensure that the final brightness and chromaticity are within ±3% of the target parameters, while in the aggressive mode, it is necessary to ensure that the final brightness and chromaticity are within ±0.5% of the target parameters.

[0050] S103: Connect and verify the calibration device: After the user sets the target parameters and optimization mode, the corresponding calibration device is connected, and the device response time is detected by sending a command to determine whether the device is successfully connected; if the device is successfully connected, the device parameters are initialized; if the device fails to connect, an alarm is triggered, prompting the user to check the physical connection of the interface or the driver installation, and an automatic retry mechanism is supported.

[0051] The calibration equipment includes a calibrator (brightness meter). The calibrator is connected to the display screen's control system via USB / Ethernet. The system automatically identifies the device driver and confirms that the communication protocol handshake is successful.

[0052] After the calibrator initializes its parameters, it loads the pre-stored parameters for the display model (such as default brightness and color temperature) by default. Then, it starts acquiring a full-screen uniformity test image of the display (such as full white / full red / full green / full blue) through the default connected multispectral camera. If the calibrator fails to initialize its parameters, it starts dynamic calibration. Dynamic calibration requires the ambient light sensor to be enabled, and the initial value is adjusted in real time to compensate for the influence of ambient light. The ambient light sensor is mainly used to sense the light intensity of the surrounding environment, thereby realizing the parameter compensation optimization settings. The initial value is a certain parameter compensation value based on the surrounding environment sensed by the ambient light sensor to improve the calibration effect.

[0053] The multispectral camera is connected to the display screen's control system via the Camera Link interface. The system automatically identifies the device driver and confirms that the communication protocol handshake is successful. The multispectral camera can be a high-precision CCD camera.

[0054] S104: Initial data acquisition: The display screen's color blocks (RGBW+50% gray) and brightness values ​​are simultaneously acquired using a multispectral camera and a brightness meter from the initial unpowered state, the heating process after power-on, to the maximum heat value. The display screen's temperature value is also acquired in real time using a temperature sensor, and then an initial report is generated.

[0055] The initial report includes a luminance uniformity distribution heatmap, a CIE1976 chromaticity scatter plot, and a temperature profile. The luminance uniformity distribution heatmap is generated by superimposing multiple images of the display screen's heating process captured by a camera over time. The CIE1976 chromaticity scatter plot is generated by superimposing the luminance values ​​of the display screen during the heating process captured by a luminance meter over time. In this process, the display screen temperature values ​​collected by the temperature sensor are converted into a temperature profile.

[0056] S105: AI Pre-analysis: A convolutional neural network is used to identify the brightness uniformity distribution map and the chromaticity CIE1976 scatter plot, identify hot spots and mosaic effects, and generate initial compensation suggestions. The initial compensation suggestions are initial correction coefficients (including RGB channel compensation parameters, pixel-level compensation parameters, and uniformity optimization parameters). The intervention value of the initial correction coefficients can be adjusted according to the temperature curve. This intervention value refers to a specific value used to adjust the initial correction coefficients.

[0057] S2: Main correction: includes:

[0058] S201: Parameter Adjustment: The least squares method is used to calculate the compensation matrix for each pixel. The initial correction coefficients are input into the lookup table of the display screen. The driver IC adjusts the PWM duty cycle and RGB mixing ratio by region to adjust the display screen parameters. Then, in step S203, the adjusted display screen parameters are compared with the user's target parameters. The full-screen uniformity index (e.g., brightness standard deviation σ≤5%) is calculated. If the standard is met, a PDF report is generated, which includes a comparison image before and after correction and the percentage improvement in brightness uniformity. If the standard is not met, it proceeds to step S202 for iterative optimization.

[0059] S202: Data Re-acquisition: If the parameters are not met after adjustment in S201, step S104 needs to be repeated for data re-acquisition. Then, AI pre-analysis is performed in step S105 to obtain the correction coefficient for the first iteration. The display parameters are readjusted based on the correction coefficient for the first iteration, and the adjusted display parameters are obtained and compared with the initial report to generate a difference report.

[0060] S203: Two-Dimensional Judgment: Adjust the display parameters based on the correction coefficients after each iteration, and compare the adjusted display parameters with the user-defined target parameters. If the user-defined target parameters are reached, record the parameters of the current iteration and proceed to the super-optimal solution detection step. Predict whether the target parameters can be exceeded using the super-optimal solution prediction model. If they can be exceeded, start the genetic algorithm optimization and record the final optimized parameters. If they cannot be exceeded, record the parameters of the current iteration as the optimal parameters. If the user-defined target parameters are not reached, execute the gradient descent optimization algorithm to achieve fast convergence. The purpose of executing the gradient descent optimization algorithm is to quickly adjust the correction coefficients and other parameters by calculating the gradient direction of the loss function when the display parameter adjustment does not reach the user-defined target parameters, so that the algorithm iteratively updates along the optimal path, thereby accelerating convergence to the optimal solution close to the target parameters.

[0061] In this embodiment, the super-optimal solution prediction model uses an LSTM neural network to predict hardware potential.

[0062] S204: Anti-dead-loop mechanism: An anti-dead-loop mechanism is added during the multiple data re-acquisitions in step S202 and the multiple iterations in step S203. This mechanism includes a triple anti-dead-loop mechanism: an iteration counter, a convergence detection, and a timeout-forced exit. A forced exit occurs when any one of these conditions is met. In this embodiment, the iteration counter is set to a maximum of 20 iterations; after 20 iterations, a forced exit occurs. The convergence detection is achieved by determining whether the chromaticity ΔE fluctuation in the last 5 iterations is less than 0.2 or whether the luminance standard deviation σ value fluctuation in the last 3 iterations is less than 0.5%. If the ΔE fluctuation is less than 0.2 or the σ fluctuation is less than 0.5%, it is considered as no improvement, indicating non-convergence. In this case, a genetic algorithm is used to optimize the compensation coefficient to avoid local optima. The timeout-forced exit is set to 30 minutes in this embodiment; after 30 minutes, a forced exit occurs.

[0063] S3: Output Decision: Based on the optimal parameters output after multiple iterations, two schemes are generated and visualized. The optimal parameters are stored. The two schemes are: Scheme A, which meets the user-defined parameters, and Scheme B, which uses AI-recommended optimal parameters. The visualization includes a side-by-side view comparing color gamut coverage and a view dynamically demonstrating the differences in Gamma curves. The optimal parameters can be written to the controller's non-volatile memory in the display control system and backed up synchronously to the cloud (using AES-256 encryption algorithm for transmission), supporting fast retrieval.

[0064] Among them, the AI-recommended super-optimal parameter scheme B is a parameter scheme generated by further AI optimization methods such as genetic algorithms after the super-optimal solution prediction model judges that there is potential for surpassing.

[0065] After automatic calibration is completed, disconnect the calibration device, release system resources, and record the operation log (including timestamp, operator ID, and snapshot of relational parameters).

[0066] Furthermore, this application also supports different adjustments to the monochromatic attenuation and compensation ratio of R, G, and B, which can be performed separately before or after the automatic calibration process.

[0067] A typical workflow example for this application is as follows:

[0068] 1. The user sets △E≤3.0 and selects "Standard Mode";

[0069] 2. The system reached ΔE=2.9 after 8 rounds of iteration;

[0070] 3. AI detected that it could be optimized to ΔE=2.3 (requiring 3 more iterations);

[0071] 4. Generation:

[0072] Option A: ΔE = 2.9 (power consumption reduced by 15%);

[0073] Option B: △E = 2.3 (color gamut expanded by 8%);

[0074] 5. Users select the final solution based on their actual needs.

[0075] The following is a portion of the Python code used to implement the automatic correction of this application:

[0076] 1. Target parameter settings:

[0077] .

[0078] 2. Initial data acquisition and AI preprocessing:

[0079] ;

[0080] The simulated data mentioned above refers to virtual data generated during algorithm iteration or model training to verify optimization strategies, test algorithm performance, or pre-adjust auxiliary parameters, rather than real measurement data directly collected by hardware devices such as luminance meters and colorimeters.

[0081] 3. The generated adjustment results:

[0082] ;

[0083] Adjust the brightness and chroma pixel by pixel based on the brightness adjustment curve and chroma adjustment trajectory generated above.

[0084] 4. Multiple iterations:

[0085] .

[0086] This application can achieve the following:

[0087] Automated calibration: Enables fully automated operation from calibration equipment to calibration completion, without manual intervention, thus improving calibration efficiency.

[0088] Dynamic optimization: Through multiple rounds of repeated operations and dynamic adjustments, the correction effect is gradually optimized to ensure correction accuracy.

[0089] Differentiated adjustment: Supports different adjustments to R, G, and B monochrome attenuation and compensation ratio to meet high-precision display requirements.

[0090] Intelligent judgment: The system uses camera data to collect and judge data, and monitors the correction effect in real time to ensure the accuracy and reliability of the correction process.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A fully automatic calibration method for an LED display screen, characterized in that: Includes the following steps: S1: Begin calibration, including: S101: System initialization; S102: User settings: Users input target parameters and select optimization modes according to their needs; the target parameters set by users include brightness, chromaticity and optimization control parameters, the optimization modes include standard mode and aggressive mode, the optimization control parameters include maximum number of iterations and convergence judgment threshold; S103: Connect the calibration device and verify that the device is successfully connected; S104: Initial data acquisition: The display screen's color blocks and brightness values ​​are synchronously acquired through a multispectral camera and a brightness meter from the initial unpowered state, the heating process after power-on, to the maximum heat value. The display screen's temperature value is also acquired in real time through a temperature sensor. An initial report is then generated, which includes a brightness uniformity distribution heatmap, a CIE1976 chromaticity scatter plot, and a temperature curve. S105: AI Pre-analysis: An AI model is used to identify the luminance uniformity distribution map and the CIE1976 chromaticity scatter plot, and to generate initial correction coefficients; Before or after calibration, users can adjust the R, G, and B monochrome attenuation and compensation ratio individually on the PC. S2: Main correction, including: S201: Parameter adjustment: Make preliminary adjustments to the display parameters based on the initial correction coefficients; then proceed to step S203 to compare the adjusted display parameters with the user's target parameters to determine whether they meet the standards. If they do not meet the standards, proceed to step S202 for iterative optimization. The parameter adjustment in step S201 is achieved by calculating the compensation matrix of each pixel using the least squares method, inputting the correction coefficients obtained in each iteration into the lookup table of the display screen, and driving the IC to adjust the PWM duty cycle and RGB mixing ratio by region to achieve the adjustment of the display screen parameters. S202: Data Re-acquisition: If the parameters are not met after step S201, step S104 needs to be repeated for data re-acquisition, and AI pre-analysis is performed through step S105 to obtain the correction coefficient of the first iteration. The display parameters are readjusted according to the correction coefficient of the first iteration, the adjusted display parameters are obtained, and a difference report is generated by comparing them with the initial report. S203: Two-dimensional judgment: Adjust the display parameters according to the correction coefficient after each iteration, compare the adjusted display parameters with the user-set target parameters, and determine whether they have reached the user's target parameters. If they have reached the user's target parameters, record the parameters of the current iteration and enter the super-optimal solution detection step. Predict whether it is possible to exceed the target parameters through the super-optimal solution prediction model. If it can exceed the target parameters, start the genetic algorithm optimization and record the final optimized parameters. If it cannot exceed the target parameters, record the parameters of the current iteration as the best parameters. If the user-set target parameters are not reached, execute the gradient descent optimization algorithm to achieve fast convergence. S3: Output Decision: Output the optimal parameters based on the above multiple iterations, and generate two schemes.

2. The full-automatic correction method of the LED display screen according to claim 1, characterized in that: In step S202, multiple data re-acquisitions are performed, and in step S203, multiple iterations are performed to determine the data. A dead loop prevention mechanism is added in step S204.

3. The full-automatic correction method of the LED display screen according to claim 2, characterized in that: The anti-dead-loop mechanism includes an iteration counter, convergence detection, and timeout forced exit. The iteration calculation can be forcibly exited when any one of these conditions is met.

4. The full-automatic correction method of the LED display screen according to claim 3, characterized in that: The AI ​​model used in the AI ​​pre-analysis is a convolutional neural network.

5. The full-automatic correction method of the LED display screen according to claim 3, characterized in that: The super-optimal solution prediction model uses an LSTM neural network to predict the hardware limits of the display system in the display screen.

6. The full-automatic correction method of the LED display screen according to claim 3, characterized in that: The system outputs two solutions: Solution A, which satisfies the user's set parameters, and Solution B, which uses AI-recommended optimal parameters. The system can visualize the two solutions and store the best parameters.

7. A computer apparatus comprising a memory, a processor, and a computer program stored on the memory, wherein: The processor executes the computer program to implement the steps of the method according to any one of claims 1-6.

Images