LED display quality stability control method and device, equipment and medium

Abstract

The present application relates to the technical field of LED display, solves the problem that the dynamic light decay compensation based on the real-time running state of LED chips cannot be realized in the prior art, it is difficult to guarantee the long-term uniformity and display quality of the display screen, and provides an LED display quality stability control method, device, equipment and medium.The method comprises the following steps: correcting the LED display screen according to an initial calibration coefficient;in the running process of the corrected LED display screen, the running parameters of each LED chip in the LED display screen are collected in real time;according to the running parameters and the chip light decay model constructed in advance, the light decay degree of the LED display screen is obtained;when the light decay degree of the LED display screen meets the preset condition, the initial calibration coefficient is adjusted according to the light decay degree to obtain a target calibration coefficient;the LED display screen is corrected according to the target calibration coefficient.The present application improves the display stability, service life and maintenance convenience of the LED display screen.

Description

Technical Field

[0001] This invention relates to the field of display technology, and in particular to a method, apparatus, device, and storage medium for stable control of LED display quality. Background Technology

[0002] LED displays, with their advantages of high brightness, low energy consumption, and long lifespan, have been widely used in various occasions such as information display, advertising, and stage performances. However, with the continuous increase in display size and the diversification of application scenarios, the stability of display quality has become an increasingly prominent issue. In practical applications, due to the inherent light decay characteristics of LED chips, which are affected by factors such as operating temperature, current, and lighting duration, long-term operation can easily lead to Mura phenomena such as uneven brightness and color distortion, severely impacting the display effect.

[0003] In the prior art, there are already automatic calibration schemes for compensating for light decay in LED displays. For example, patent publication number CN113611242 A discloses an automatic calibration method for LED displays. This method monitors the junction temperature of LED chips and combines it with a preset light decay curve, adjusting the calibration coefficient based on usage time to achieve light decay compensation and uniformity improvement to a certain extent. This method can periodically correct the calibration parameters according to temperature changes and cumulative usage time, thereby improving the display quality degradation problem caused by light decay.

[0004] However, the existing solutions still have significant shortcomings. First, relying solely on average junction temperature and illumination duration as primary references fails to adequately reflect the combined impact of various actual operating conditions, such as current and environmental changes, on light decay. This results in limited compensation accuracy and difficulty in adapting to complex and dynamic operating environments. Second, existing methods are mostly based on periodic testing and batch parameter updates, unable to provide refined, dynamic, and real-time compensation and adjustments according to actual operating conditions. Even after long-term operation, manual intervention or periodic maintenance is still necessary. Third, traditional solutions lack sufficient consideration for the differences at the chip level in large-scale, high-resolution LED displays, making it difficult to achieve single-chip precision correction.

[0005] Therefore, there is an urgent need for a method to stabilize LED display quality by combining real-time operating parameters of LED chips, adapting to dynamic changes in multiple operating conditions, and dynamically adjusting calibration coefficients based on a refined light decay model, so as to further improve the uniformity, color consistency and automation of the display screen during long-term operation. Summary of the Invention

[0006] In view of this, embodiments of the present invention provide a method, apparatus, device and storage medium for stable control of LED display quality, in order to solve the problem that the prior art cannot achieve dynamic light decay compensation based on the real-time operating status of LED chips, and it is difficult to guarantee the long-term uniformity and display quality of the display screen.

[0007] In a first aspect, embodiments of the present invention provide a method for stabilizing LED display quality, the method comprising:

[0008] The LED display screen is calibrated according to an initial calibration coefficient, wherein the initial calibration coefficient is obtained by performing a mura test on the LED display screen;

[0009] During the operation of the calibrated LED display screen, the operating parameters of each LED chip in the LED display screen are collected in real time;

[0010] The light decay level of the LED display screen is obtained based on the operating parameters and the pre-built chip light decay model.

[0011] When the light decay of the LED display screen meets the preset conditions, the initial calibration coefficient is adjusted according to the light decay to obtain the target calibration coefficient;

[0012] The LED display screen is calibrated according to the target calibration coefficient.

[0013] Preferably, during the operation of the calibrated LED display screen, the real-time acquisition of operating parameters of each LED chip in the LED display screen includes:

[0014] The total working time is obtained by statistically analyzing the lighting time of each LED chip based on the lighting records of the LED display screen.

[0015] The operating temperature of each LED chip is obtained based on a pre-built temperature database;

[0016] The operating current of each LED chip is collected according to the preset sampling frequency;

[0017] Based on the operating current, obtain the average current of each LED chip during the operating time;

[0018] The operating parameters are obtained based on the average current, the operating temperature, and the cumulative operating time.

[0019] Preferably, the operating temperature of each LED chip is obtained based on a pre-built temperature database, including:

[0020] Based on the physical structure parameters of the LED display screen and the LED chip arrangement information, establish a mapping relationship between the physical row and column coordinates and the number of the LED chips;

[0021] After the target LED screen reaches thermal equilibrium, acquire a temperature field distribution image of the target LED screen;

[0022] Based on the mapping relationship, obtain the pixel position information of each LED chip in the temperature field distribution image;

[0023] Based on the pixel location information and the preset deep learning model, the temperature field distribution image is divided to obtain the temperature region image of each LED chip.

[0024] Based on the temperature region image, obtain the temperature value of each LED chip;

[0025] Based on the temperature values ​​and the mapping relationship, the temperature database is established;

[0026] The operating temperature of each LED chip is obtained by querying the temperature database based on the chip number of each LED chip.

[0027] Preferably, the temperature field distribution image is divided according to the pixel position information and the trained deep learning model to obtain a temperature region image for each LED chip;

[0028] Based on the pixel position information, a coordinate prior heatmap with the same resolution as the temperature field distribution image is generated, wherein in the coordinate prior heatmap, the center pixel of each LED chip is assigned a value of 1, and the other pixels are assigned a value of 0;

[0029] The infrared temperature field distribution image and the coordinate prior heatmap are stitched together to obtain a dual-channel input tensor;

[0030] The deep learning model trained by the dual-channel input tensor is used to obtain multi-scale mixed features of temperature coordinates. The deep learning model includes a fully convolutional network model, a U-Net model, and a region convolutional neural network model.

[0031] Based on the multi-scale mixing features, the temperature feature components corresponding to each pixel in the infrared temperature field distribution map are obtained to obtain the temperature feature map.

[0032] Based on the temperature feature map, the first-order difference of each pixel in the horizontal and vertical directions is calculated to obtain the temperature gradient map.

[0033] The gradient magnitude of each pixel is calculated based on the temperature gradient map to obtain the temperature difference attention weights;

[0034] The chip region probability map is obtained by weighting and fusing the temperature difference attention weights and the features of the decoding layer of the deep learning model.

[0035] Based on the chip region probability map, connected component filtering is performed to obtain a set of candidate chip regions;

[0036] Based on the candidate chip region set and the coordinate prior heatmap, the connected component closest to the center pixel of each LED chip is obtained, thus obtaining the temperature region image of each LED chip.

[0037] Preferably, the light decay degree of the LED display screen is obtained based on the operating parameters and a pre-built chip light decay model, including:

[0038] Based on the historical operating data of the LED display under different operating conditions, a chip light decay model under different operating conditions is constructed.

[0039] Based on the aforementioned operating parameters, determine the actual operating condition range;

[0040] The operating parameters are input into the chip light decay model corresponding to the actual operating condition range to obtain the light decay degree of the LED display screen.

[0041] Preferably, the step of constructing a chip light decay model under different operating conditions based on historical operating data of the LED display screen under different operating conditions includes:

[0042] Based on the historical operating data of the aforementioned operating condition intervals, an experimental sample set for each operating condition interval is obtained;

[0043] Feature extraction is performed on the test sample set to obtain the attenuation feature dataset for each operating condition range;

[0044] Nonlinear regression analysis was performed based on the attenuation characteristic dataset to obtain the mathematical model of light attenuation under each operating condition range;

[0045] Based on the optical decay model, the attenuation rate under the operating conditions in the specified operating range is analyzed to obtain the attenuation rate coefficient;

[0046] Based on the attenuation rate coefficient and corresponding optical decay mathematical model under different operating conditions, the chip optical decay model under different operating conditions is obtained.

[0047] Preferably, when the light decay level of the LED display screen meets a preset condition, adjusting the initial calibration coefficient according to the light decay level to obtain the target calibration coefficient includes:

[0048] The actual light decay rate of the LED display screen is obtained based on the light decay level and the preset light decay threshold.

[0049] Based on the actual light decay rate of each LED chip and the preset light decay threshold, the LED chips that need brightness compensation are determined, and the chips that need compensation are obtained.

[0050] Calculate the brightness compensation value for each chip based on its actual light decay rate and preset brightness value.

[0051] Based on the brightness compensation value, a target brightness instruction is generated for each chip that needs compensation.

[0052] Based on the target brightness command and the initial calibration coefficient, the calibration coefficient of each chip requiring compensation is adjusted to obtain the target calibration coefficient.

[0053] Secondly, embodiments of the present invention provide an LED display quality stabilization control device, the device comprising:

[0054] The first calibration module is used to calibrate the LED display screen according to an initial calibration coefficient, wherein the initial calibration coefficient is obtained by performing a mura test on the LED display screen;

[0055] The operating parameter acquisition module is used to collect the operating parameters of each LED chip in the LED display screen in real time during the operation of the calibrated LED display screen.

[0056] The light decay degree acquisition module is used to acquire the light decay degree of the LED display screen based on the operating parameters and the pre-built chip light decay model;

[0057] The calibration coefficient adjustment module is used to adjust the initial calibration coefficient according to the light decay level when the light decay level of the LED display screen meets the preset conditions, so as to obtain the target calibration coefficient.

[0058] The second calibration module is used to calibrate the LED display screen according to the target calibration coefficient.

[0059] Thirdly, embodiments of the present invention provide an electronic device, including: at least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method of the first aspect described above.

[0060] Fourthly, embodiments of the present invention provide a storage medium storing computer program instructions, which, when executed by a processor, implement the method of the first aspect described above.

[0061] In summary, the beneficial effects of the present invention are as follows:

[0062] The LED display quality stability control method, apparatus, device, and storage medium provided in this invention obtain initial calibration coefficients by performing Mura testing on the LED display screen. This enables precise correction of the initial non-uniformity of each display screen, improving initial display uniformity and image quality consistency. During display screen operation, the operating parameters of each LED chip are continuously collected, providing a data foundation for subsequent dynamic compensation and achieving comprehensive perception of the chip-level operating status. Utilizing a pre-constructed chip light decay model, combined with real-time collected operating parameters, the light decay degree of each chip can be dynamically calculated and accurately reflected, thereby accurately determining the current state of the display screen. When the light decay degree is detected to meet preset conditions, the calibration coefficients can be dynamically adjusted based on the actual light decay condition to promptly compensate for the decrease in LED chip brightness, ensuring long-term stable display effects. Through continuous feedback and adjustment of the calibration coefficients, dynamic and adaptive closed-loop compensation of the display screen is achieved without frequent manual intervention, effectively improving the long-term uniformity and display quality stability of the LED display screen.

[0063] In summary, this invention enables high-precision automatic uniformity compensation of the display screen from initial manufacturing to long-term operation, effectively solving problems such as uneven brightness and color distortion caused by LED chip light decay and changes in operating conditions, and significantly improving the display quality, lifespan, and maintenance convenience of LED displays. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments of the present invention will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, and these are all within the protection scope of the present invention.

[0065] Figure 1 This is a flowchart illustrating an embodiment of the LED display quality stability control method of the present invention.

[0066] Figure 2 This is another schematic diagram of the LED display quality stability control method according to an embodiment of the present invention.

[0067] Figure 3 This is another flowchart illustrating the LED display quality stability control method according to an embodiment of the present invention.

[0068] Figure 4 This is a schematic diagram of the LED display quality stabilization control device according to an embodiment of the present invention.

[0069] Figure 5 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0070] The features and exemplary embodiments of various aspects of the present invention will now be described in detail. To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention and are not configured to limit the present invention. For those skilled in the art, the present invention can be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present invention by illustrating examples of the invention.

[0071] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element.

[0072] Example 1

[0073] Please see Figure 1-3 This invention provides a method for stabilizing LED display quality, the method comprising:

[0074] S1. The LED display screen is calibrated according to the initial calibration coefficient, wherein the initial calibration coefficient is obtained by performing a mura test on the LED display screen;

[0075] Specifically, Mura testing refers to using high-resolution image acquisition equipment to accurately detect the brightness and color distribution of an LED display screen, quantifying issues such as dark spots, bright spots, and color unevenness on the screen, and representing these non-uniformities as quantifiable data. Based on these test results, corresponding calibration coefficients (i.e., initial calibration coefficients) can be automatically calculated for each pixel or LED chip. These coefficients are used to compensate for brightness or color differences in different areas. In this way, before the LED display screen is officially put into use, the light emission performance of each area can be unified by adjusting the parameters of the driver IC or control system, achieving factory-standard brightness and color uniformity. This method greatly improves the initial non-uniformity caused by chip performance fluctuations and splicing errors, laying a solid foundation for subsequent long-term automated uniformity maintenance.

[0076] S2. During the operation of the calibrated LED display screen, the operating parameters of each LED chip in the LED display screen are collected in real time;

[0077] Specifically, this step is a continuous process during the operational phase after the initial calibration. Its core function is to collect the operating parameters of each LED chip in the LED display screen in real time. Operating parameters refer to various data reflecting the actual working state of the LED chip, commonly including drive current, cumulative lighting time, and operating temperature. The main purpose of this step is to provide data support for subsequent light decay assessment and dynamic compensation, ensuring that subsequent compensation measures respond based on actual operating conditions. In practice, integrated current detection modules, temperature sensors, timers and counters, and other hardware, combined with the software acquisition logic of the display screen control system, can periodically acquire and store relevant operating data for each chip according to a preset sampling period. Continuous data acquisition not only provides detailed evidence for dynamically evaluating changes in LED chip performance but also enables large-scale automated operation and maintenance.

[0078] S3. Based on the operating parameters and the pre-built chip light decay model, obtain the light decay degree of the LED display screen;

[0079] Specifically, this step combines the real-time operating parameters obtained in the previous step with a pre-built chip light decay model to dynamically calculate the degree of light decay of the display screen. The chip light decay model is based on extensive historical experiments and actual operating data, reflecting the quantitative relationship between LED chip brightness and factors such as time, temperature, and current. The main purpose of this step is to accurately assess the current brightness decay level of each chip using the model, thereby promptly grasping the trend of uniformity changes in the display screen. In practice, parameters such as the operating time, operating current, and temperature of each chip are input into the light decay model to obtain the current brightness decay value or decay rate of each chip. Through a model-based, data-driven evaluation method, not only is the accuracy of light decay compensation improved, but a scientific basis is also provided for subsequent compensation threshold determination and automatic adjustment. This technical feature enables objectivity and real-time performance in light decay judgment, effectively improving the intelligent level of display screen uniformity maintenance.

[0080] S4. When the light decay of the LED display screen meets the preset conditions, the initial calibration coefficient is adjusted according to the light decay to obtain the target calibration coefficient;

[0081] When the light decay level is detected to meet preset conditions, the initial calibration coefficient is automatically adjusted according to the specific light decay situation, thereby generating a new target calibration coefficient. The core purpose of this step is to achieve dynamic adaptive compensation for performance changes of each LED chip during operation. During implementation, the actual light decay level is compared with a preset threshold. When the brightness decay of a certain area or a certain chip is found to exceed the allowable range, the required compensation amount is automatically calculated, and the calibration coefficient corresponding to that chip or area is corrected to ensure that its output brightness is restored to the design level or close to the initial state. This process requires no manual intervention and can efficiently respond to and correct changes in display uniformity during display operation, significantly improving the stability and self-maintenance capability of the display system.

[0082] S5. The LED display screen is calibrated according to the target calibration coefficient.

[0083] This step implements the results of the aforementioned dynamic adjustments, performing a second global or local calibration of the display screen based on the target calibration coefficients. The purpose of this step is to apply the newly calculated calibration parameters to the control logic of each chip in the display screen, ensuring that the compensation effect takes effect promptly, thereby maintaining the long-term brightness and color uniformity of the display screen. Specifically, brightness compensation for the relevant LED chips is achieved by updating chip drive parameters or PWM duty cycles. Through a continuous closed-loop dynamic calibration mechanism, the display screen can adaptively cope with performance fluctuations and unevenness that occur during long-term operation, effectively extending product lifespan, reducing maintenance costs, and improving the end-user's display experience.

[0084] Preferably, during the operation of the calibrated LED display screen, the real-time acquisition of operating parameters of each LED chip in the LED display screen includes:

[0085] S21. Based on the lighting records of the LED display screen, calculate the lighting time of each LED chip to obtain the cumulative working time;

[0086] Specifically, the lighting time of each LED chip is calculated by tracking the power-on, power-off, and power-off operation logs of each chip in the display screen control system. The lighting periods of each instance are then summed to obtain the cumulative operating time of each chip since it was put into operation. The cumulative operating time not only provides basic data for subsequent aging analysis but also directly reflects the operating load experienced by the chip, playing a crucial role in chip lifespan management and light decay assessment.

[0087] S22. Obtain the operating temperature of each LED chip based on the pre-built temperature database;

[0088] Specifically, based on a pre-built temperature database, the operating temperature of each LED chip is obtained. This is achieved by utilizing the temperature database established earlier through infrared thermal imaging, spatial coordinate mapping, and chip numbering, combined with the physical location and arrangement information of the chips, to accurately retrieve the temperature value corresponding to each LED chip under actual operating conditions. Operating temperature, as a crucial environmental parameter affecting the chip's luminous efficiency and light decay rate, is of great significance for assessing its long-term stability.

[0089] S23. Collect the operating current of each LED chip according to the preset sampling frequency;

[0090] Specifically, the operating current of each LED chip is collected in real time according to a set sampling frequency. This process uses a built-in current detection module to periodically measure and record the driving current of each chip, enabling a dynamic reflection of the actual power consumption of the chip under different loads and display content. Periodic data collection not only ensures the integrity of the sample but also provides a reliable basis for analyzing the long-term energy consumption and load changes of the chips.

[0091] S24. Based on the operating current, obtain the average current of each LED chip during the operating time;

[0092] Specifically, this step uses the cumulative operating time of each chip as a basis to statistically analyze all collected current data and calculate the average current of the chip throughout its entire operating cycle. The average current, as an indicator reflecting the typical driving strength of the chip, can reveal its long-term power consumption characteristics and changes in load conditions.

[0093] S25. Obtain the operating parameters based on the average current, the operating temperature, and the cumulative operating time.

[0094] Specifically, the average current, operating temperature, and cumulative operating time of each LED chip are summarized to form a set of operating parameters that comprehensively reflect the actual operating status of the chips. This process provides comprehensive and detailed data support for subsequent light decay modeling and dynamic correction strategies based on operating parameters, helping to improve the overall display quality and operational controllability of LED displays.

[0095] Preferably, the operating temperature of each LED chip is obtained based on a pre-built temperature database, including:

[0096] S221. Based on the physical structure parameters of the LED display screen and the LED chip arrangement information, establish the mapping relationship between the physical row and column coordinates and the number of the LED chips.

[0097] Specifically, based on the physical structural parameters of the LED display screen and the chip layout information, a mapping relationship is established between the physical row and column coordinates of each LED chip and its unique number. Specifically, according to the size and resolution of the display module and the arrangement of the chips on the circuit board, the actual physical location of each chip is matched one-to-one with its assigned logical number in the display system, ensuring that the spatial information and data identification of any subsequent chip can be accurately matched. This mapping relationship provides a standardized spatial positioning basis for subsequent thermal distribution analysis and temperature data archiving.

[0098] S222. After the target LED screen reaches thermal equilibrium, acquire a temperature field distribution image of the target LED screen.

[0099] The screen reaching thermal equilibrium refers to a state where, after continuous operation for a certain period, the overall heat distribution stabilizes and no longer fluctuates drastically. At this point, the acquired temperature field distribution image reflects the stable thermal environment of the LED chips under actual operating conditions. The purpose of acquiring the temperature field distribution image is to obtain the current temperature distribution of various areas of the entire screen, providing fundamental data for the subsequent accurate extraction of the operating temperature of individual chips. In practice, after the display screen is powered on and continuously displays a high-brightness or standard image for a set period, a high-precision infrared thermal imager is used to capture images of the entire screen, obtaining a raw temperature image containing the thermal distribution of all chips. This temperature image serves as the base map for subsequent analysis, comprehensively reflecting the screen's thermal load status and significantly improving the representativeness and accuracy of temperature acquisition.

[0100] S223. Based on the mapping relationship, obtain the pixel position information of each LED chip in the temperature field distribution image;

[0101] The key objective of this step is to precisely map the spatial information of the LED chips to the pixel coordinates of the temperature field image. Using the previously established mapping relationship between chip rows, columns, and numbers, the physical spatial location of each chip is mapped to a specific pixel in the thermal imaging image through coordinate transformation. For example, the chip in row 10, column 20 corresponds to pixel coordinates (x, y) in the temperature field image. This step can be implemented using image calibration algorithms (such as perspective transformation and automatic calibration point recognition), combined with the actual module layout for coordinate conversion, ensuring the accurate positioning of each chip in the temperature image. This technical feature helps to seamlessly integrate the actual physical structure and digital image data, laying a precise spatial foundation for subsequent region division and temperature value extraction.

[0102] S224. Based on the pixel position information and the preset deep learning model, the temperature field distribution image is divided to obtain the temperature region image of each LED chip.

[0103] Specifically, a deep learning model is used to automatically segment the temperature field distribution image, obtaining the temperature region image for each LED chip. Deep learning models such as U-Net can perform pixel-level segmentation of the thermal imaging image based on input chip pixel coordinates, thermal distribution patterns, and other information, accurately separating the temperature region of each chip from the overall image. The purpose of this step is to improve the automation and accuracy of chip temperature region extraction, avoiding errors caused by manual segmentation or simple threshold segmentation. In practice, the temperature image and chip coordinates are first input into a pre-trained neural network model. The model automatically outputs a temperature region mask corresponding to each chip, ultimately forming a temperature sub-image corresponding to each chip. Introducing deep learning segmentation greatly improves the automation of data processing and the robustness of segmentation under complex working conditions.

[0104] S225. Obtain the temperature value of each LED chip based on the temperature region image;

[0105] Specifically, representative temperature values ​​are extracted from the temperature region images of each chip. These temperature values ​​are typically based on the average, median, or weighted value of all pixels in that region, objectively reflecting the chip's operating temperature at thermal equilibrium. The goal is to transform the raw, high-dimensional thermal imaging data into structured temperature data that can be directly used for chip-level data analysis. This can be achieved using methods such as region averaging, clustering statistics, or filtering outlier pixels to ensure the extracted temperature values ​​are accurate and stable. This step significantly simplifies the original temperature data structure, providing a foundation for subsequent temperature database establishment and querying.

[0106] S226. Establish the temperature database based on the temperature values ​​and the mapping relationship;

[0107] To further integrate the temperature values ​​of all chips with their corresponding spatial identifiers, a structured temperature database was systematically established. This database uses the chip identifier as the primary key, storing the operating temperature of each chip under its current thermal equilibrium state, and can be linked as needed with metadata such as coordinates, acquisition time, and environmental conditions. The aim is to provide a unified data platform for efficient retrieval, dynamic maintenance, and historical traceability of subsequent chip-level temperature data for the display system. The establishment of this temperature database helps to continuously monitor the temperature evolution trends of chips in various regions, enabling early warning of potential risks such as abnormal temperature rises or uneven heat distribution, and enhancing the overall intelligent perception and self-maintenance capabilities of the system.

[0108] S227. Query the temperature database according to the chip number of each LED chip to obtain the operating temperature of each LED chip.

[0109] Specifically, the real-time operating temperature of the target chip can be efficiently retrieved from the temperature database by chip number, providing accurate environmental input for subsequent temperature-based light decay modeling, dynamic compensation, and fault detection. This query mechanism ensures the timeliness and consistency of temperature data, enabling the system to respond in real time to changes in the thermal environment during display operation, further improving the intelligence and automation level of display uniformity maintenance and anomaly control.

[0110] Preferably, the temperature field distribution image is divided according to the pixel position information and the trained deep learning model to obtain a temperature region image for each LED chip;

[0111] S2241. Based on the pixel position information, generate a coordinate prior heatmap with the same resolution as the temperature field distribution image, wherein in the coordinate prior heatmap, the center pixel of each LED chip is assigned a value of 1, and the other pixels are assigned a value of 0.

[0112] Specifically, the coordinate prior heatmap refers to an auxiliary image with the same resolution as the infrared temperature field distribution image. It assigns a value of 1 to the center pixel of each LED chip and 0 to the remaining pixels. The purpose of generating this heatmap is to explicitly embed the spatial location information of each chip into the subsequent image analysis process, providing clear spatial localization guidance for the deep learning model. In the specific implementation, based on the chip arrangement information, the pixels corresponding to the center of all chips are highlighted on a completely black background image, while other pixels remain at zero. Using this structured prior information as an auxiliary feature not only improves the convergence speed of the model for chip localization but also significantly enhances the model's ability to perceive spatial structure, reducing the risk of misjudgment in segmentation under conditions of dense chip arrangement or partial occlusion.

[0113] S2242. The infrared temperature field distribution image and the coordinate prior heat map are stitched together to obtain a dual-channel input tensor;

[0114] By fusing spatial priors with actual thermal distribution data, a multi-channel input tensor for deep learning segmentation is obtained. In implementation, the infrared temperature field distribution image is concatenated with the aforementioned coordinate prior heatmap along the channel dimension, forming a three-dimensional array with two feature layers (temperature distribution and spatial prior), i.e., a dual-channel input tensor. This approach allows the deep model to automatically focus on chip-related location regions while learning temperature features, improving the context awareness and robustness of the segmentation process.

[0115] S2243. Input the dual-channel input tensor into the trained deep learning model to obtain multi-scale mixed features of temperature coordinates, wherein the deep learning model includes a fully convolutional network model, a U-Net model, and a region convolutional neural network model.

[0116] The dual-channel input tensor is fed into a trained deep learning model. This deep learning model can include segmentation structures such as fully convolutional networks (FCNs), U-Net networks, or region convolutional neural networks (e.g., Mask R-CNN). Utilizing the model's multi-layer convolution and feature fusion capabilities, it automatically extracts and encodes temperature variations and spatial location information from the input tensor, outputting a multi-scale temperature-coordinate hybrid feature. This feature fusion mechanism can simultaneously capture global thermal distribution patterns and local spatial details, providing a highly discriminative information foundation for subsequent chip-level segmentation.

[0117] S2244. Based on the multi-scale mixing features, obtain the temperature feature components corresponding to each pixel in the infrared temperature field distribution map to obtain the temperature feature map.

[0118] The purpose of this step is to transform complex multi-scale mixed features into a temperature feature map that is easy to analyze. In implementation, the feature channels or combinations thereof most relevant to temperature changes are extracted from the feature map output by the deep network, constructing a two-dimensional map reflecting the temperature attributes of each pixel. This temperature feature map makes subsequent gradient and region analysis operable and intuitive, facilitating the identification of minute temperature differences between chips.

[0119] S2245. Based on the temperature feature map, calculate the first-order difference of each pixel in the horizontal and vertical directions to obtain the temperature gradient map.

[0120] Specifically, based on the temperature feature map, the first-order difference of each pixel in the horizontal and vertical directions is calculated to obtain the temperature gradient map. The temperature gradient reflects the drastic degree of temperature change in space, which can help subsequent models accurately identify chip boundaries. By convolving the feature map with the gradient operator, two-dimensional data reflecting the temperature change trend can be quickly generated, providing key support for the attention mechanism.

[0121] S2246. Calculate the gradient magnitude of each pixel based on the temperature gradient map to obtain the temperature difference attention weight;

[0122] Specifically, the amplitude of the temperature gradient map is calculated to obtain the attention weights for temperature differences at each pixel. This operation typically normalizes the gradient vector amplitude at each point, generating a weight map between 0 and 1, reflecting the sensitivity of each region to temperature changes. Regions with higher weights often correspond to chip edges or areas of abrupt changes in the thermal environment, and the subsequent model will pay more attention to these key regions. This mechanism significantly improves the model's ability to identify chip boundaries under complex thermal distributions.

[0123] S2247. The chip region probability map is obtained by weighted fusion based on the temperature difference attention weight and the features of the decoding layer of the deep learning model.

[0124] This method leverages temperature-difference attention weights to guide the weighted fusion of features from the decoding layer of a deep learning model. In implementation, the attention weights are used as masks to weight the multi-channel features output from the decoding layer pixel-by-pixel, improving the segmentation sensitivity of chip region boundaries. The fused output is the chip region probability map, where each pixel's value represents its probability of belonging to the chip region. This approach, through a soft segmentation strategy, avoids boundary jumps caused by hard thresholding, significantly improving segmentation smoothness and boundary accuracy.

[0125] S2248. Perform connected component filtering based on the chip region probability map to obtain a set of candidate chip regions;

[0126] The goal of connectivity component filtering based on the chip region probability map is to extract high-confidence candidate chip regions from the probability map. By setting a probability threshold, regions exceeding the threshold are identified as chip regions. These regions undergo connectivity analysis and labeling to select an independent set of candidate regions. This filtering process eliminates isolated pixel noise that may be output by the model, ensuring the stability and accuracy of subsequent temperature extraction.

[0127] S2249. Based on the candidate chip region set and the coordinate prior heat map, obtain the connected component closest to the center pixel of each LED chip to obtain the temperature region image of each LED chip.

[0128] Based on the selected candidate chip region set and the prior coordinate heatmap, the center pixel of each LED chip is matched with its nearest connected component, achieving precise temperature region assignment for each chip. Finally, temperature region images corresponding one-to-one with the physical chips are output, providing a reliable foundation for subsequent temperature statistics within the region and acquisition of chip operating temperatures. This refined process significantly improves the accuracy of chip-level temperature analysis for high-density LED displays, providing solid technical support for display thermal management and light decay compensation.

[0129] Preferably, obtaining the light decay level of the LED display screen based on the operating parameters and a pre-built chip light decay model includes:

[0130] S31. Based on the historical operating data of the LED display screen under different operating conditions, construct a chip light decay model under different operating conditions.

[0131] A chip light decay model is a mathematical model that reflects the brightness decay of an LED chip over time under different operating conditions (such as temperature, current, and operating duration). Based on historical operating data of the display screen, the data is categorized and statistically analyzed according to different operating condition intervals such as temperature, current, and operating duration. For each operating condition interval, the chip brightness change characteristics are extracted, and algorithms such as nonlinear regression are used for fitting, ultimately establishing an independent light decay curve for each typical operating condition. In this way, the light decay model not only reflects the overall trend but also meticulously distinguishes the influence patterns of various typical operating conditions, providing a high-precision prediction basis for subsequent dynamic compensation. Modeling across multiple operating condition intervals significantly improves the model's adaptability and predictive ability in complex real-world application environments.

[0132] S32. Determine the actual operating condition range based on the operating parameters;

[0133] Specifically, the core of step S32 is to determine the current actual operating condition range of the LED chip based on the collected operating parameters. Operating parameters include, but are not limited to, cumulative operating time, average operating temperature, and typical operating current. The purpose of determining the operating condition range is to select the light decay model that best matches the current actual environment within a multi-model system, thereby obtaining the most accurate decay estimate. In implementation, the operating parameters are compared with pre-defined operating condition thresholds (e.g., temperatures above a certain value are classified as high-temperature range, currents within a certain range are classified as standard range, etc.) to categorize the current chip state into the corresponding operating condition category. In this way, each chip can be matched with its most suitable light decay model, improving the specificity of light decay assessment.

[0134] S33. Input the operating parameters into the chip light decay model corresponding to the actual operating condition range to obtain the light decay degree of the LED display screen.

[0135] Combining the first two steps, the real-time collected operating parameters are input into the corresponding chip light decay model to obtain the light decay level of the LED display under the current operating conditions. This light decay level is typically quantified as brightness attenuation rate, remaining brightness percentage, etc. Specifically, the operating condition range corresponding to the operating parameters is located, the light decay model for that range is retrieved, and real-time parameters (such as running time, current current, and temperature) are used as model input to calculate the current brightness attenuation level of the chip or region. This method not only enables accurate light decay assessment for different usage environments but also provides a reliable foundation for subsequent dynamic brightness compensation and calibration coefficient adjustments, improving the intelligent level of long-term stability and consistent maintenance of the display.

[0136] Preferably, the step of constructing a chip light decay model under different operating conditions based on historical operating data of the LED display screen under different operating conditions includes:

[0137] S311. Based on the historical operating data of the operating condition intervals, obtain the experimental sample set for each operating condition interval;

[0138] The experimental sample set refers to the data collected based on historical operating data of LED displays. All data is categorized according to pre-defined operating condition intervals and archived into multiple experimental sample sets. Operating condition intervals are typically set based on environmental and load conditions such as temperature range, current magnitude, and operating time. For example, they can be divided into high-temperature, high-current intervals, low-temperature, low-current intervals, etc. The purpose of this step is to provide targeted data support for subsequent model building, ensuring that the light decay modeling under each operating condition interval is based on sufficient and effective historical samples. In practice, database retrieval and filtering can automatically aggregate historical operating records belonging to the same interval into independent datasets, laying the foundation for feature analysis and modeling. This approach is beneficial for accurately characterizing the impact of multiple operating conditions on chip light decay, improving the adaptability and generalization ability of subsequent models.

[0139] S312. Extract features from the test sample set to obtain the attenuation feature dataset for each operating condition interval;

[0140] For each experimental sample set, feature extraction was performed to transform the raw time-series data into a feature dataset reflecting the light decay patterns of LED chips. Common features include brightness decay curves corresponding to lighting duration, the relationship between ambient temperature and decay rate, and the impact of current changes on brightness. Feature extraction can employ not only statistical methods, such as calculating brightness change rates, maximum and minimum values, and average trends, but also signal processing or machine learning methods to uncover hidden patterns. The resulting decay feature dataset provides rich and representative input variables for subsequent regression modeling, effectively improving the model's ability to characterize actual operational patterns.

[0141] S313. Perform nonlinear regression analysis based on the attenuation characteristic dataset to obtain the light attenuation mathematical model for each operating condition range;

[0142] Based on the attenuation characteristic dataset for each operating condition interval, nonlinear regression analysis is used to model optical decay. Nonlinear regression, such as exponential, logarithmic, and multinomial regression, can be used to fit the nonlinear trend of chip brightness changing with variables such as time, obtaining a mathematical model of optical decay for each operating condition interval. This mathematical model can be used to quantitatively describe the brightness decay process of the chip under specific operating conditions. Through regression fitting, the model parameters can be kept highly consistent with actual historical data, ensuring the reliability of subsequent predictions. The technical feature of this step is that it uses mathematical fitting tools to automatically find the optimal model form and parameters, improving the scientific nature and engineering practicality of optical decay modeling.

[0143] S314. Analyze the attenuation rate under the operating conditions in the operating range based on the optical attenuation model, and obtain the attenuation rate coefficient;

[0144] The established mathematical model of light decay is analyzed to quantify the decay rate under different operating conditions. The "decay rate coefficient" refers to the factor in the model parameters that directly reflects the rate of light decay. Analysis methods typically include analytical analysis of the model derivative and parameter sensitivity analysis. The purpose of this step is to quantify the specific impact of various environmental and load factors on the decay rate of LED chips, identify the operating conditions under which light decay is most significant, and provide theoretical support for developing differentiated compensation strategies in practical applications. By scientifically decomposing and extracting the decay rate coefficient, more accurate parameter inputs can be provided for subsequent applications such as dynamic brightness correction and lifetime management.

[0145] S315. Based on the attenuation rate coefficient and the corresponding optical decay mathematical model under different operating conditions, obtain the chip optical decay model under different operating conditions.

[0146] By combining the attenuation rate coefficients obtained under each operating condition range with the corresponding mathematical models of light decay, a systematic library of chip light decay models for different operating conditions is formed. Each model not only includes a mathematical expression but also explicitly associates it with the corresponding environment and load range, achieving seamless integration between the model and the actual operating scenario. After obtaining a complete light decay model, the most suitable light decay model can be dynamically selected and invoked based on the real-time monitored chip operating conditions, achieving targeted compensation and accurate prediction. This technical feature significantly improves the display stability and intelligent adaptive level of LED displays under varying operating conditions, providing a solid data and theoretical foundation for achieving high reliability and long lifespan displays.

[0147] Preferably, when the light decay level of the LED display screen meets a preset condition, adjusting the initial calibration coefficient according to the light decay level to obtain the target calibration coefficient includes:

[0148] S41. Obtain the actual light decay rate of the LED display screen based on the light decay level and the preset light decay threshold;

[0149] By comparing actual test results with threshold standards, the light decay level of each chip is quantified, and it is determined whether it has reached the threshold requiring compensation. In practice, the system automatically calculates the actual brightness decay ratio of each LED chip and compares it with the corresponding threshold, thus providing a quantitative basis for compensation decisions. This design helps to accurately identify chips with degraded performance, avoid overcompensation or missing anomalies, and provide a scientific criterion for maintaining display uniformity.

[0150] S42. Based on the actual light decay rate of each LED chip and the preset light decay threshold, determine the LED chips that need brightness compensation, and obtain the chips that need compensation.

[0151] Based on the aforementioned light decay analysis results, it is clearly determined which LED chips require brightness compensation. Chips requiring compensation here refer to those whose actual light decay rate exceeds the system's preset threshold. The main purpose of this step is to filter out the key points that truly affect display uniformity and brightness consistency from all LED chips on the screen, effectively focusing compensation resources. This is achieved by iterating through all chips and automatically adding those with an actual light decay rate greater than the threshold to the compensation list, while the rest remain in their original state. This dynamic filtering significantly improves the accuracy and efficiency of the compensation algorithm, avoids invalid operations, and ensures the stability and consistency of the screen's display performance.

[0152] S43. Calculate the brightness compensation value for each chip based on the actual light decay rate and preset brightness value of the chip to be compensated.

[0153] Brightness compensation value refers to the amount of additional or adjusted compensation required for each chip requiring compensation, based on the difference between its actual light decay rate and the target brightness. For example, if a chip's actual brightness is 85%, while the preset target is 95%, the compensation value is 10%. The purpose of this step is to develop a personalized compensation scheme for each chip requiring compensation, achieving precise and differentiated brightness enhancement. During implementation, the system calculates the numerical difference between the current brightness decay of each chip and the preset target brightness, and outputs the corresponding compensation value. This technical feature can restore and maintain the consistency of chip brightness to the greatest extent, providing a precise foundation for the next step of automatic instruction generation.

[0154] S44. Generate a target brightness instruction for each chip that needs compensation based on the brightness compensation value;

[0155] The numerical brightness compensation requirements are translated into specific hardware execution instructions. The target brightness instruction refers to the adjustment command generated by the system that can directly affect the LED driver IC or control unit, such as current gain adjustment or PWM duty cycle correction. The goal is to achieve automated issuance and hardware-level execution of the compensation strategy. In practice, the system generates a matching target brightness instruction for each chip requiring compensation based on the compensation value and sends it to the corresponding chip control circuit via the data bus, ensuring that the compensation measures can be applied to the display terminal instantly and accurately. This step significantly improves the automation and response speed of the display compensation process, effectively reducing manual intervention and adjustment errors.

[0156] S45. Adjust the calibration coefficient of each chip that needs compensation according to the target brightness command and the initial calibration coefficient to obtain the target calibration coefficient.

[0157] Based on the target brightness command and the original initial calibration coefficients, the calibration coefficients of each chip requiring compensation are updated to obtain the final target calibration coefficients. These calibration coefficients are the core parameters for the drive control system to adjust the actual brightness of each chip. By dynamically adjusting the calibration coefficients according to the latest brightness requirements, the system can continuously correct the chip output, keeping its brightness within the set range. The specific process includes reading the current calibration coefficients, mathematically superimposing or replacing them with the target brightness adjustment amount, and applying the result as the new target calibration coefficients to the chip control process. This closed-loop compensation mechanism significantly improves the brightness stability and uniformity of the LED display during long-term operation, ensuring high-quality image display and a longer, uninterrupted operating life.

[0158] In one embodiment, step S45 further includes:

[0159] S451. Based on the LED display screen and chip layout information, divide the screen area to obtain the partition number of several partitions and the chip set corresponding to each partition.

[0160] Specifically, the process begins by dividing the LED display screen into several regions based on its physical structural parameters and the specific chip layout. This results in several partition numbers and the corresponding chip sets for each region. The physical structural parameters typically include the number of pixel rows and columns, the distribution of unit modules, etc., while the chip layout information refers to the physical coordinates or unique identifier of each LED chip within the entire screen. This scientific division of regions not only facilitates subsequent zoned management of the display status but also provides a physical basis for analyzing regional errors and aging trends. For example, a large display screen can be divided into several rectangular regions, with the number and distribution of chips within each region as uniform as possible to achieve precise compensation and control.

[0161] S452. Statistical analysis is performed on the drift error data of all chips in each partition during the most recent preset number of calibration processes to obtain the drift error time series of each partition.

[0162] Statistical analysis is performed on the drift error data of all chips within each partition during the most recent calibration processes to obtain the drift error time series for each partition. Drift error refers to the deviation of chip parameters such as brightness and chromaticity from the ideal reference value. By collecting and sorting the calibration results of each partition multiple times over time, a data series reflecting the performance evolution trend of each partition can be formed. The purpose of this step is to comprehensively understand the historical variation patterns of each region, providing a sufficient data foundation for the next step of correlation analysis. For example, the average brightness drift of a certain region during the most recent ten calibrations can be plotted as a time series curve, intuitively reflecting its aging process or environmental impact.

[0163] S453. Perform correlation analysis based on the drift error time series of each partition to obtain the correlation coefficient matrix between each partition.

[0164] Specifically, after obtaining the drift error time series for each partition, correlation analysis is performed between each partition and other partitions to obtain a correlation coefficient matrix between the partitions. Correlation analysis often uses methods such as the Pearson correlation coefficient to assess the degree of synchronization of error changes over time in different partitions. By constructing the correlation matrix, it is possible to clearly identify which partitions have strong error linkage and which are relatively independent. The purpose of this step is to provide data for subsequent linkage adjustments during regional compensation, ensuring that compensation decisions are more consistent with the actual situation of the entire screen.

[0165] S454. Determine the associated partitions for each partition based on the correlation coefficient matrix and the preset correlation threshold.

[0166] Specifically, based on step S453, the associated partitions for each partition are determined according to the correlation coefficient matrix and a preset correlation threshold. Associated partitions refer to other partitions that exhibit strong synchronicity with the current partition in terms of error change trends. Typically, partitions with a correlation coefficient greater than a certain threshold (such as 0.5 or 0.7) are selected as associated partitions. This screening process eliminates weakly correlated or irrelevant partitions, focusing on regions that have a real impact on the compensation strategy, greatly improving the effectiveness and computational efficiency of the compensation algorithm.

[0167] S455. Based on the correlation coefficient matrix, perform linkage weight allocation to obtain the linkage weight between each partition and the corresponding associated partition.

[0168] Subsequently, linkage weights are allocated based on the correlation coefficient matrix to obtain the linkage weights (S455) between each partition and its corresponding associated partitions. The linkage weights reflect the degree of impact of a partition's compensation adjustment on other related partitions. Typically, the correlation coefficients are normalized into weight values, ensuring that the weight allocation reflects both the linkage strength between partitions and facilitates the subsequent quantification of compensation parameters. For example, for a region A, assuming its correlation coefficients with regions B and C are 0.8 and 0.6 respectively, the linkage weights for AB and AC can be set to 0.57 and 0.43 (after normalization), respectively.

[0169] S456. Based on the target brightness command and initial calibration coefficient of each chip that needs compensation, perform preliminary adjustment of the chip calibration coefficient to obtain intermediate calibration coefficient.

[0170] Next, based on the target brightness command and initial calibration coefficients of each chip requiring compensation, the chip calibration coefficients are initially adjusted to obtain intermediate calibration coefficients. The target brightness command here is given based on the detected actual brightness attenuation. Each chip requiring compensation, combined with its own initial calibration parameters, yields a preliminary compensation result that does not yet consider the effects of regional linkage. This step lays the foundation for regional compensation, but at this point, the inter-regional linkage effect has not yet been introduced.

[0171] S457. Based on the aforementioned linkage weight, adjust the intermediate calibration coefficient of the associated partitions of each partition to obtain the target calibration coefficient;

[0172] First, for each partition, the system searches for its associated partitions identified in the previous steps (S454, S455). These associated partitions are those that exhibit strong synchronization with the current partition in terms of error change trends, and corresponding linkage weights have been assigned to them. Each linkage weight represents the degree of influence of associated partitions on the compensation result of the current partition during compensation adjustment.

[0173] In actual adjustments, the system does not simply use the intermediate calibration coefficient of the current partition as the final compensation value. Instead, it combines the intermediate calibration coefficient of the current partition with the intermediate calibration coefficients of all its associated partitions. Specifically, the system assigns different levels of importance to the intermediate calibration coefficients of the current partition and its associated partitions according to the linkage weight between each pair of partitions. The intermediate calibration coefficient of the current partition itself usually has a higher weight, while the intermediate calibration coefficients of associated partitions that are highly correlated with the current partition also account for a certain proportion.

[0174] The purpose of this approach is to ensure that the final compensation result for each zone not only fully reflects its own brightness attenuation but also adaptively corrects for changes in surrounding areas. If there is a significant synchronization of brightness changes between a zone and one or more related zones, the compensation amplitude of these related zones will be automatically incorporated when compensating this zone, achieving a smooth transition in brightness adjustment between zones. This effectively prevents the appearance of dividing lines or color blocks on the screen caused by drastic calibration changes in a single area.

[0175] Finally, the system uses the calibration coefficients, adjusted by weighted linkage, as the target calibration coefficients for each zone, and distributes them to each zone or each chip requiring compensation. The entire process is automated, requiring no manual intervention, achieving highly adaptive and dynamically smooth compensation control between regions, significantly improving the overall display uniformity and visual quality of the LED display screen.

[0176] In one embodiment, after step S457, the method further includes:

[0177] S458. Based on the target calibration coefficients of all partitions and the spatial adjacency of the partitions, filter adjacent partition pairs to obtain a set of all spatially directly adjacent partition pairs.

[0178] In this embodiment, after the linkage adjustment of the partition target calibration coefficients is completed, a further region boundary smoothing process is introduced to optimize the display continuity between adjacent partitions. First, in step S458, the system determines the spatial adjacency of all partitions on the display screen based on the target calibration coefficients of all partitions and the known spatial adjacency relationships of the partitions. Through this determination method, the system can filter out all spatially directly adjacent partition pairs to obtain a set of partition pairs for subsequent boundary smoothing processing. The adjacency relationship here usually refers to partitions that share a physical boundary or are the closest in distance, ensuring that the compensation parameter smoothing process focuses on the positions that will actually produce a seamless effect visually.

[0179] S459. Based on the target calibration coefficients of each pair of adjacent partitions, calculate the difference in calibration coefficients to obtain the difference in calibration coefficients between each pair of adjacent partitions.

[0180] Subsequently, in step S459, the system acquires the target calibration coefficient for each pair of adjacent partitions, compares their values, and calculates the difference in calibration coefficients between each pair of adjacent partitions. This difference data reflects the magnitude of the calibration coefficient jump between spatially adjacent areas after compensation adjustment, providing a quantitative basis for whether subsequent smoothing correction is needed.

[0181] S4510. Based on the comparison between the calibration coefficient difference data between each pair of adjacent partitions and the preset boundary smoothing threshold, obtain the set of partition pairs whose calibration coefficient difference exceeds the threshold, and denote it as the partition boundary set to be smoothed.

[0182] In step S4510, the system compares the calibration coefficient difference data of all partition pairs obtained in the previous step with a pre-set boundary smoothing threshold. Any partition pairs with calibration coefficient differences exceeding this threshold are filtered out, forming a set of partition boundaries to be smoothed. This threshold is typically set with reference to the human eye's sensitivity threshold to abrupt changes in brightness or color, ensuring that only areas that may cause visual breaks or color patches are included in the smoothing process.

[0183] S4511. Based on the target calibration coefficients of each pair of partitions in the set of partition boundaries to be smoothed, determine the smoothing correction parameters to obtain the correction magnitude and correction method for each partition to be smoothed.

[0184] For each partition to be smoothed, appropriate smoothing correction parameters are determined based on the actual differences in its target calibration coefficients and the screen's physical layout. Specifically, a correction magnitude, i.e., the amount of compensation to be adjusted, is specified for each pair of partitions, along with the correction method to be used, such as linear interpolation, weighted averaging, or boundary transition layering. This ensures that the smoothing process eliminates abrupt changes without affecting the overall compensation trend of the entire screen.

[0185] Specifically, the process first iterates through the set of all partition boundaries to be smoothed. For each pair of partitions, the target calibration coefficients for these two partitions are extracted, and the actual numerical difference between them is calculated. Then, by combining the magnitude of the difference, the physical location of the boundary on the screen, and the overall calibration strategy, the correction magnitude and specific correction method for that partition are determined.

[0186] The determination of the correction magnitude is primarily based on the difference between the calibration coefficients of the two partitions. For example, based on the actual difference value and according to the hierarchical processing rules, an appropriate correction amount can be selected to eliminate abrupt changes without excessively interfering with the overall compensation trend of each partition. The correction magnitude can be half of the difference between the two, or it can be weighted and adjusted according to factors such as the importance of the partition, the displayed content, and the visual sensitivity of the boundary in the overall screen.

[0187] In terms of correction methods, various algorithm strategies can be adopted. For example, for large-area partitions, linear interpolation can be used to progressively adjust the calibration coefficients of the two partitions at the boundary, creating a smooth transition in the transition area. For smaller or locally sensitive areas, weighted averaging can be used to adjust the calibration coefficients of the two partitions towards the median value according to a set ratio. Furthermore, if the boundaries of certain partition pairs involve complex image details or special display content, the system can also employ more advanced correction methods such as multi-level decreasing / increasing or adaptive transition to achieve the best visual smoothing effect.

[0188] After the correction parameters are determined, the system uses the correction magnitude and selected correction method for each partition to be smoothed as the basis for subsequent synchronous adjustments, ensuring that the adjustment operation can proceed in an orderly manner according to the established strategy. This step lays the foundation for the continuity of parameters between partitions and is a key link in the entire boundary smoothing process, effectively improving the uniformity of the LED display screen and its engineering applicability.

[0189] S4512. Based on the correction magnitude and correction method of each partition to be smoothed, the partition target calibration coefficients are synchronously adjusted to obtain the partition target calibration coefficient set after smoothing correction.

[0190] Based on the correction magnitude and method obtained in the previous step, the target calibration coefficients for each partition to be smoothed are adjusted synchronously. During this synchronous adjustment, the calibration coefficients of two adjacent partitions are brought closer to the median value, or the adjustment amount is distributed according to a certain weight, achieving a smooth progression of the transition zone. The final output is a set of target calibration coefficients for the partitions after boundary smoothing correction, making the spatial distribution of calibration parameters more continuous for all processed partitions.

[0191] S4513. Based on the synchronous adjustment results of all partition calibration coefficients, perform full-screen calibration coefficient consistency and transition detection to obtain the final target calibration coefficient that meets the requirements of boundary continuity and smooth transition.

[0192] Finally, the results of synchronized adjustment of all partition calibration coefficients are tested for full-screen consistency and boundary transition. This testing includes, but is not limited to, re-verifying that the differences in calibration coefficients between all adjacent partitions are all below the smoothing threshold, and determining whether the overall parameter distribution is continuous and smooth. Only when the full-screen calibration coefficients meet the requirements for accurate compensation within the region and achieve a smooth transition at all spatial boundaries are the final target calibration coefficients output. This ensures that the LED display can consistently present a highly uniform visual effect without obvious breaks or color blocks during long-term operation and dynamic compensation. This series of boundary smoothing steps effectively solves the problem of abrupt compensation changes that may occur at partition boundaries in traditional compensation algorithms, significantly improving the image quality and user experience of LED display products.

[0193] Example 2

[0194] Please see Figure 4 This invention provides an LED display quality stabilization control device, the device comprising:

[0195] The first calibration module is used to calibrate the LED display screen according to an initial calibration coefficient, wherein the initial calibration coefficient is obtained by performing a mura test on the LED display screen;

[0196] The operating parameter acquisition module is used to collect the operating parameters of each LED chip in the LED display screen in real time during the operation of the calibrated LED display screen.

[0197] The light decay degree acquisition module is used to acquire the light decay degree of the LED display screen based on the operating parameters and the pre-built chip light decay model;

[0198] The calibration coefficient adjustment module is used to adjust the initial calibration coefficient according to the light decay level when the light decay level of the LED display screen meets the preset conditions, so as to obtain the target calibration coefficient.

[0199] The second calibration module is used to calibrate the LED display screen according to the target calibration coefficient.

[0200] It should be noted that each module and unit in the LED display quality stabilization control device in this embodiment corresponds one-to-one with each step in the LED display quality stabilization control method in the aforementioned embodiment. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned LED display quality stabilization control method, and will not be repeated here.

[0201] Example 3

[0202] In addition, combined Figure 1 The LED display quality stability control method described in this embodiment of the invention can be implemented by an electronic device. Figure 5 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention is shown.

[0203] Electronic devices may include processors and memory storing computer program instructions.

[0204] Specifically, the processor may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement embodiments of the present invention.

[0205] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0206] Computer-readable media include both permanent and non-permanent, removable and non-removable media, which can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient media, such as modulated communication signals and carrier waves.

[0207] The processor reads and executes computer program instructions stored in the memory to implement any of the LED display quality stability control methods in the above embodiments.

[0208] In one example, the electronic device may also include a communication interface and a bus. For example, Figure 5 As shown, the processor 401, memory 402, and communication interface 403 are connected through bus 410 and complete communication with each other.

[0209] The communication interface is mainly used to enable communication between various modules, devices, units and / or equipment in the embodiments of the present invention.

[0210] A bus, including hardware, software, or both, couples components of an electronic device together. For example, and not limitingly, a bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, a bus may include one or more buses. While specific buses are described and illustrated in embodiments of the invention, the invention contemplates any suitable bus or interconnect.

[0211] Example 4

[0212] Furthermore, in conjunction with the LED display quality stabilization control method in the above embodiments, this invention can be implemented using a computer-readable storage medium. This computer-readable storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any one of the LED display quality stabilization control methods in the above embodiments.

[0213] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0214] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0215] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0216] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0217] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0218] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0219] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. An LED display quality stabilization control method, characterized in that, The method comprises: correcting the LED display screen according to initial calibration coefficients, wherein the initial calibration coefficients are obtained by performing a mura test on the LED display screen; in the running process of the corrected LED display screen, collecting running parameters of each LED chip in the LED display screen in real time; obtaining the light decay degree of the LED display screen according to the running parameters and a chip light decay model constructed in advance, comprising: constructing the chip light decay model under different working condition intervals according to historical running data of the LED display screen under different working condition intervals, which specifically comprises: obtaining an experimental sample set of each working condition interval according to the historical running data under the working condition interval; performing feature extraction on the experimental sample set to obtain a decay feature data set under each working condition interval; performing nonlinear regression analysis on the decay feature data set to obtain a light decay mathematical model under each working condition interval; obtaining a decay rate coefficient according to the light decay model to analyze the decay rate under the working condition interval; obtaining the chip light decay model under different working condition intervals according to the decay rate coefficient and the corresponding light decay mathematical model under different working condition intervals; determining an actual working condition interval according to the running parameters; inputting the running parameters into the chip light decay model corresponding to the actual working condition interval to obtain the light decay degree of the LED display screen; when the light decay degree of the LED display screen meets a preset condition, adjusting the initial calibration coefficients according to the light decay degree to obtain target calibration coefficients, comprising: obtaining an actual light decay rate of the LED display screen according to the light decay degree and a preset light decay threshold; determining LED chips that need to be compensated according to the actual light decay rate of each LED chip and the preset light decay threshold to obtain compensation chips; calculating a brightness compensation value of each compensation chip according to the actual light decay rate of the compensation chip and a preset brightness value; generating a target brightness instruction of each compensation chip according to the brightness compensation value; adjusting the calibration coefficients of each compensation chip according to the target brightness instruction and the initial calibration coefficients to obtain the target calibration coefficients, comprising: performing screen region division according to the physical structure parameters and chip arrangement information of the LED display screen to obtain the partition numbers of a plurality of partitions and the region corresponding chip sets; statistically analyzing the drift error data of all chips in each partition in the last preset number of calibration processes to obtain a drift error time sequence of each partition; performing correlation analysis on the drift error time sequence of each partition to obtain a correlation coefficient matrix between the partitions; determining the associated partitions of each partition according to the correlation coefficient matrix and a preset correlation threshold; performing linkage weight distribution according to the correlation coefficient matrix to obtain the linkage weight of each partition and the corresponding associated partition; performing preliminary adjustment of the chip calibration coefficients according to the target brightness instruction of each compensation chip and the initial calibration coefficients to obtain intermediate calibration coefficients; adjusting the intermediate calibration coefficients of the associated partitions of each partition according to the linkage weight to obtain the target calibration coefficients; correcting the LED display screen according to the target calibration coefficients. 2.The LED display quality stabilization control method of claim 1, wherein, The running parameters of each LED chip in the LED display screen are collected in real time during the operation of the corrected LED display screen, including: The lighting time of each LED chip is counted according to the lighting record of the LED display screen, and the cumulative working time is obtained; The working temperature of each LED chip is obtained according to the pre-constructed temperature database; The working current of each LED chip is collected according to the preset sampling frequency; The current mean value of each LED chip within the working time is obtained according to the working current; The running parameters are obtained according to the current mean value, the working temperature and the cumulative working time. 3.The LED display quality stabilization control method of claim 2, wherein, The working temperature of each LED chip is obtained according to the pre-constructed temperature database, including: A mapping relationship between the physical row and column coordinates of the LED chip and the number is established according to the physical structure parameters and the LED chip arrangement information of the LED display screen; After the LED display screen reaches a thermal equilibrium state, a temperature field distribution image of the LED display screen is obtained; According to the mapping relationship, the pixel position information of each LED chip in the temperature field distribution image is obtained; According to the pixel position information and the preset deep learning model, the temperature field distribution image is divided to obtain a temperature region image of each LED chip; According to the temperature region image, the temperature value of each LED chip is obtained; According to the temperature value and the mapping relationship, the temperature database is established; The working temperature of each LED chip is obtained by querying the temperature database according to the chip number of each LED chip. 4.The LED display quality stabilization control method of claim 3, wherein, According to the pixel position information and the trained deep learning model, the temperature field distribution image is divided to obtain a temperature region image of each LED chip; According to the pixel position information, a coordinate prior heat map with the same resolution as the temperature field distribution image is generated, wherein the center pixel of each LED chip in the coordinate prior heat map is assigned a value of 1, and the remaining pixels are assigned a value of 0; The temperature field distribution image and the coordinate prior heat map are spliced to obtain a dual-channel input tensor; The dual-channel input tensor is input into the trained deep learning model to obtain a temperature coordinate multi-scale mixed feature, wherein the deep learning model includes a full convolution network model, a U-Net model and a region convolution neural network model; According to the multi-scale mixed feature, the temperature feature component corresponding to each pixel point in the infrared temperature field distribution image is obtained to obtain a temperature feature map; According to the temperature feature map, the first-order difference of each pixel point in the horizontal and vertical directions is calculated to obtain a temperature gradient map; According to the temperature gradient map, the gradient amplitude of each pixel is calculated to obtain a temperature difference attention weight; According to the temperature difference attention weight and the features of the decoding layer of the deep learning model, weighted fusion is performed to obtain a chip region probability map; According to the chip region probability map, connected domain screening is performed to obtain a candidate chip region set; According to the candidate chip region set and the coordinate prior heat map, the connected domain closest to the center pixel of each LED chip is obtained to obtain a temperature region image of each LED chip.

5. An LED display quality stabilization control device, characterized by, The device comprises: The first correction module is configured to correct the LED display screen according to initial calibration coefficients, wherein the initial calibration coefficients are obtained by performing a mura test on the LED display screen. The running parameter acquisition module is configured to collect running parameters of each LED chip of the LED display screen in real time during operation of the corrected LED display screen. The light decay degree acquisition module is configured to obtain a light decay degree of the LED display screen according to the running parameters and a chip light decay model constructed in advance, including: According to historical running data of the LED display screen under different working condition intervals, a chip light decay model under different working condition intervals is constructed, which specifically includes: According to the historical running data under the working condition intervals, an experimental sample set of each working condition interval is obtained. Feature extraction is performed on the experimental sample set to obtain a decay feature data set under each working condition interval. Nonlinear regression analysis is performed on the decay feature data set to obtain a light decay mathematical model under each working condition interval. According to the light decay model, a decay rate coefficient under the working condition interval is obtained. According to the decay rate coefficients and the corresponding light decay mathematical models under different working condition intervals, a chip light decay model under different working condition intervals is obtained. According to the running parameters, an actual working condition interval is determined. The running parameters are input into the chip light decay model corresponding to the actual working condition interval to obtain the light decay degree of the LED display screen, including: According to the light decay degree and a preset light decay threshold, an actual light decay rate of the LED display screen is obtained. According to the actual light decay rate of each LED chip and the preset light decay threshold, an LED chip that needs to be compensated is determined to obtain a compensation chip. According to the actual light decay rate of the compensation chip and a preset brightness value, a brightness compensation value of each compensation chip is calculated. According to the brightness compensation value, a target brightness instruction of each compensation chip is generated. According to the target brightness instruction and the initial calibration coefficient, a calibration coefficient of each compensation chip is adjusted to obtain a target calibration coefficient, including: According to the physical structure parameters and the chip arrangement information of the LED display screen, screen region division is performed to obtain the partition numbers of a plurality of partitions and the region corresponding chip sets. According to the drift error data of all chips in each partition in the last preset number of calibration processes, statistical analysis is performed to obtain a drift error time sequence of each partition. According to the drift error time sequence of each partition, correlation analysis is performed to obtain a correlation coefficient matrix between the partitions. According to the correlation coefficient matrix and a preset correlation threshold, an associated partition of each partition is determined. According to the correlation coefficient matrix, a linkage weight distribution is performed to obtain a linkage weight of each partition and the corresponding associated partition. According to the target brightness instruction of each compensation chip and the initial calibration coefficient, preliminary adjustment of the chip calibration coefficient is performed to obtain an intermediate calibration coefficient. According to the linkage weight, the intermediate calibration coefficient of the associated partition of each partition is adjusted to obtain the target calibration coefficient. The calibration coefficient adjustment module is configured to adjust the initial calibration coefficient according to the light attenuation degree when the light attenuation degree of the LED display screen meets a preset condition, so as to obtain a target calibration coefficient. The second correction module is configured to correct the LED display screen according to the target calibration coefficient.

6. An electronic device, comprising: The method comprises: at least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method of any one of claims 1-4.

7. A storage medium, characterized by The computer program instructions are stored on the computer readable medium, and when the computer program instructions are executed by the processor, the method of any one of claims 1-4 is implemented.

Images