LED splicing screen cloud correction method, device and equipment and storage medium

Abstract

The application discloses an LED splicing screen cloud correction method and device, equipment and a storage medium, relates to the technical field of data processing, and comprises the following steps: acquiring splicing screen data, performing single-screen basic correction on a screen module based on the splicing screen data; after the single-screen basic correction is completed, whether there is a boundary mismatch is monitored through an edge computing node; the boundary mismatch includes that the brightness and / or chromaticity of the edge pixels of adjacent screen modules deviate; if there is a boundary mismatch, boundary data is uploaded to the cloud through the edge computing node; based on a pre-trained boundary correction model in the cloud, boundary correction parameters are obtained, and cross-screen compensation is performed by receiving the boundary correction parameters. Through cloud intelligent compensation, the splicing joint visible to the naked eye of the splicing screen is eliminated, the display quality is improved, the influence of environmental factors is dynamically compensated, the adaptability of the splicing screen is improved, the resource utilization efficiency is improved by combining edge computing and a cloud model, and the operation and maintenance cost is reduced.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a cloud-based correction method, apparatus, device, and storage medium for LED splicing screens. Background Technology

[0002] In the field of LED display technology, video wall systems are widely used in high-performance scenarios such as performance projection and broadcasting studios, and their display consistency directly affects the visual presentation effect. Existing technologies typically employ a localized single-module basic calibration scheme: collecting module brightness / chromaticity data through edge nodes, combining it with a local aging compensation algorithm to generate calibration parameters, and then burning them into the receiver card.

[0003] However, local edge nodes are limited by hardware resources, resulting in limited pixel-level brightness and color compensation accuracy in the splicing seam area; single-module correction relies only on its own historical data and cannot integrate the spatiotemporal correlation features of cross-screen multi-modules, causing brightness jumps and color discontinuities at the splicing seam; at the same time, the local algorithm lacks the ability to respond to environmental disturbances in real time, and the compensation coefficients are fixed in the hardware memory and cannot be dynamically iterated according to the running status, causing the correction effect to decay rapidly over time.

[0004] Therefore, existing technologies, limited by the computing power constraints of localized architectures, cannot meet the demand for high-precision, long-term elimination of display splicing seams.

[0005] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] The main objective of this application is to provide a cloud-based calibration method, apparatus, device, and storage medium for LED splicing screens, aiming to solve the technical problem of failing to meet the requirements for high-precision and long-term elimination of splicing seams in display screens.

[0007] To achieve the above objectives, this application proposes a cloud-based calibration method for LED video walls, the method comprising:

[0008] Acquire splicing screen data, and perform single-screen basic calibration of the screen module based on the splicing screen data;

[0009] After the single-screen basic calibration is completed, edge computing nodes are used to monitor whether there is boundary mismatch; the boundary mismatch includes deviations in brightness and / or chromaticity of edge pixels of adjacent screen modules;

[0010] If a boundary mismatch exists, the boundary data is uploaded to the cloud via the edge computing node;

[0011] Based on the pre-trained boundary correction model in the cloud, boundary correction parameters are obtained, and cross-screen compensation is performed using the boundary correction parameters.

[0012] In one embodiment, the step of acquiring splicing screen data and performing single-screen basic calibration of the screen module based on the splicing screen data includes:

[0013] Obtain the current brightness value, current chromaticity value, and first working time of each screen module;

[0014] Upload the current brightness value, current chromaticity value, and first working time to the cloud.

[0015] The brightness gain coefficient and / or chromaticity gain coefficient of the corresponding screen module are obtained by using a pixel aging model pre-trained in the cloud.

[0016] The system receives the brightness gain coefficient and / or chromaticity gain coefficient to perform single-screen basic calibration on the corresponding screen module, and stores the calibration data after single-screen basic calibration.

[0017] In one embodiment, the step of monitoring for boundary mismatch via edge computing nodes after the single-screen basic calibration is completed includes:

[0018] Extract the brightness correction value of each pixel within the target area of ​​the splicing screen module;

[0019] Calculate the average corrected brightness of the target area based on the brightness correction value;

[0020] The standard deviation of the brightness of the target area is calculated using the average corrected brightness.

[0021] The brightness standard deviation is compared with a preset brightness standard deviation threshold.

[0022] If the brightness standard deviation is greater than the preset brightness standard deviation threshold, then a boundary mismatch is determined, and the target area is marked as an abnormal area; otherwise, the brightness correction value is maintained.

[0023] In one embodiment, the step of monitoring for boundary mismatch via edge computing nodes after the single-screen basic calibration is completed further includes:

[0024] Extract the chromaticity correction value of each pixel within the target area of ​​the splicing screen module;

[0025] Based on the chromaticity correction value and the pre-stored reference chromaticity value, the color deviation is calculated;

[0026] If the color deviation is greater than a preset color deviation threshold, the target area is marked as an abnormal area; otherwise, the color correction value is maintained.

[0027] In one embodiment, the boundary correction model includes a luminance compensation model and a chrominance compensation model; the step of obtaining boundary correction parameters based on the pre-trained boundary correction model in the cloud includes:

[0028] Receive the boundary data and preprocess the boundary data;

[0029] Extracting the brightness gradient of adjacent pixels and the aging coefficient of the screen module based on preprocessed boundary data;

[0030] The brightness gradient of adjacent pixels and the aging coefficient of the screen module are input into the brightness compensation model to generate a brightness compensation gradient field for the corresponding abnormal region; the brightness compensation gradient field is the brightness correction parameter of the boundary correction parameter.

[0031] In one embodiment, the boundary correction model includes a luminance compensation model and a chrominance compensation model; the step of obtaining boundary correction parameters based on the pre-trained boundary correction model in the cloud further includes:

[0032] Receive the boundary data, and perform three-color separation and feature extraction based on the boundary data;

[0033] The data obtained from feature extraction is input into the color compensation model, and the red / blue / green compensation coefficients are calculated for each channel.

[0034] The red / blue / green compensation coefficients are synergistically optimized based on white balance constraints, color gamut constraints, and visual smoothing constraints to obtain the final gain adjustment coefficient; the final gain adjustment coefficient is the chromaticity correction parameter of the boundary correction parameter.

[0035] In one embodiment, the step of receiving the boundary correction parameters for cross-screen compensation includes:

[0036] Verify whether the standard deviation of brightness and / or color deviation of abnormal areas after cross-screen compensation meets the standards;

[0037] If the brightness standard deviation and / or color deviation do not meet the standards, the non-compliant abnormal areas are marked as secondary abnormal areas, and cloud-based diagnosis or strategy iteration is performed on the secondary abnormal areas.

[0038] Furthermore, to achieve the above objectives, this application also proposes an LED splicing screen cloud calibration device, which includes:

[0039] The data acquisition and processing module is used to acquire splicing screen data and perform single-screen basic calibration of the screen module based on the splicing screen data.

[0040] The boundary monitoring module is used to monitor whether there is boundary mismatch after the single-screen basic calibration is completed through edge computing nodes; the boundary mismatch includes the brightness and / or color deviation of the edge pixels of adjacent screen modules;

[0041] The data communication module is used to upload boundary data to the cloud through the edge computing node if boundary mismatch exists;

[0042] The correction execution module is used to obtain boundary correction parameters based on a pre-trained boundary correction model in the cloud, and to receive the boundary correction parameters for cross-screen compensation.

[0043] In addition, to achieve the above objectives, this application also proposes an LED splicing screen cloud calibration device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the LED splicing screen cloud calibration method as described above.

[0044] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the LED splicing screen cloud correction method described above.

[0045] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the LED splicing screen cloud correction method described above.

[0046] One or more technical solutions proposed in this application have at least the following technical effects:

[0047] This application acquires splicing screen data and performs basic single-screen calibration of the screen modules based on the splicing screen data. After the basic single-screen calibration is completed, edge computing nodes monitor for boundary mismatch. The boundary mismatch includes deviations in brightness and / or color of edge pixels of adjacent screen modules. If boundary mismatch exists, the boundary data is uploaded to the cloud through the edge computing nodes. Based on a pre-trained boundary calibration model in the cloud, boundary calibration parameters are obtained, and cross-screen compensation is performed using the boundary calibration parameters. This application eliminates visible seams in splicing screens through cloud-based intelligent compensation, improves display quality, dynamically compensates for the influence of environmental factors, improves the adaptability of splicing screens, and combines edge computing with cloud models to improve resource utilization efficiency and reduce operation and maintenance costs. Attached Figure Description

[0048] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating an embodiment of the cloud-based calibration method for LED splicing screens in this application.

[0051] Figure 2 This is a flowchart illustrating Embodiment 2 of the LED splicing screen cloud calibration method of this application;

[0052] Figure 3 This is a flowchart illustrating Embodiment 3 of the LED splicing screen cloud calibration method of this application;

[0053] Figure 4 This is a flowchart illustrating Embodiment 4 of the LED splicing screen cloud calibration method of this application;

[0054] Figure 5 This is a flowchart illustrating Embodiment 5 of the LED splicing screen cloud calibration method of this application;

[0055] Figure 6 This is a flowchart illustrating Embodiment Six of the LED splicing screen cloud calibration method of this application;

[0056] Figure 7 This is a schematic diagram of the module structure of the LED splicing screen cloud correction device according to an embodiment of this application;

[0057] Figure 8 This is a schematic diagram of the hardware operating environment involved in the cloud-based calibration method for LED splicing screens in this application embodiment.

[0058] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0059] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0060] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0061] Existing technologies cannot meet the requirements for high-precision, long-term elimination of splicing seams in displays.

[0062] This application provides a solution that acquires splicing screen data and performs basic single-screen calibration of the screen modules based on the splicing screen data. After the basic single-screen calibration is completed, edge computing nodes monitor for boundary mismatch. The boundary mismatch includes deviations in brightness and / or chromaticity of edge pixels of adjacent screen modules. If boundary mismatch exists, the boundary data is uploaded to the cloud through the edge computing nodes. Based on a pre-trained boundary calibration model in the cloud, boundary calibration parameters are obtained, and cross-screen compensation is performed using the boundary calibration parameters. This application eliminates visible seams in splicing screens through cloud-based intelligent compensation, improves display quality, dynamically compensates for the influence of environmental factors, enhances the adaptability of splicing screens, and combines edge computing with cloud models to improve resource utilization efficiency and reduce operation and maintenance costs.

[0063] Based on this, this application provides a cloud-based calibration method for LED splicing screens, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the cloud-based calibration method for LED splicing screens in this application.

[0064] In this embodiment, the LED splicing screen cloud calibration method includes steps S10~S40:

[0065] Step S10: Obtain splicing screen data, and perform single-screen basic calibration of the screen module based on the splicing screen data;

[0066] It should be noted that splicing screen data refers to the collection of optical parameters (brightness, chromaticity), equipment status parameters (cumulative working time, drive current), and environmental parameters (temperature, humidity) collected from each screen module of the LED splicing screen; single-screen basic calibration refers to the process of compensating for pixel-level brightness and chromaticity deviations within a single screen module to bring them to a preset standard value; a screen module refers to the smallest physical unit that constitutes the splicing screen, including LED bead arrays, drive circuits, and control interfaces. Raw data is collected in real time by local optical sensors (such as spectrophotometers) and status monitoring chips, combined with gain coefficients sent from the cloud, to eliminate display deviations within a single screen module caused by LED aging, drive circuit drift, or environmental fluctuations, ensuring the display uniformity of each module and laying the foundation for cross-screen collaborative calibration.

[0067] In one possible implementation, the optical sensor collects the luminance / chromaticity values ​​of the module's center and four corners at a sampling rate of 120Hz, synchronously records the module's surface temperature through an infrared sensor, monitors the power supply voltage based on the ADC, reads the cumulative working time from the receiver card's EEPROM, and uploads it to the cloud via an encrypted link; the cloud-based pixel aging model can generate gain coefficients based on a spatiotemporal convolutional network (STCNN).

[0068] Step S20: After the single-screen basic calibration is completed, the edge computing node is used to monitor whether there is a boundary mismatch; the boundary mismatch includes the brightness and / or color deviation of the edge pixels of adjacent screen modules;

[0069] It should be noted that edge computing nodes refer to embedded processors deployed locally on the splicing screen, capable of analyzing pixel data in real time. Boundary mismatch refers to the visual discontinuity caused by abrupt changes in brightness and color at the seam between adjacent screen modules, quantified by indicators such as brightness standard deviation or color deviation. The target area refers to a rectangular region extending to the left and right of the splicing seam, with several rows / columns of pixels (e.g., 2 columns). After basic calibration on a single screen, the edge nodes extract the calibrated pixel data and detect boundary anomalies through brightness standard deviation and color deviation calculations. The brightness standard deviation threshold and color deviation threshold are dynamically set.

[0070] In one specific implementation: For the third row of the horizontal seam, extract the brightness values ​​of the three columns of pixels in each of the left and right modules, for example, a total of 6×1080 pixels; calculate the brightness standard deviation σ=7.8%, which is greater than the preset brightness standard deviation threshold of 5%, and the color deviation ΔE=4.2, which is greater than the preset color deviation threshold of 3; mark the seam area as an abnormal area and generate a JSON data packet containing the position coordinates and deviation values.

[0071] Step S30: If a boundary mismatch exists, the boundary data is uploaded to the cloud through the edge computing node;

[0072] It should be noted that boundary data refers to the compressed dataset of the abnormal area, including pixel coordinates, corrected luminance / chrominance values, module ID, environmental parameters, and diagnostic indicators; encrypted data packets can be transmitted to the cloud message queue via the 5G / MQTT protocol.

[0073] In one possible implementation, differential coding is used to transmit change amounts and Zstandard compressed pixel data, combined with an encryption algorithm to transmit data to the cloud. When the network is interrupted, local caching is activated, and the interrupted transmission is resumed after the link is restored.

[0074] Step S40: Based on the pre-trained boundary correction model in the cloud, obtain the boundary correction parameters, and receive the boundary correction parameters to perform cross-screen compensation.

[0075] It should be noted that, in this application, the boundary correction model refers to a set of correction models deployed in the cloud, including a brightness compensation network and a chroma compensation network. Cross-screen compensation refers to synchronously distributing compensation coefficients to adjacent modules, and adjusting the PWM (Pulse Width Modulation) driving parameters based on the boundary correction parameters of the brightness compensation gradient field and the final gain coefficient combination to achieve gradual blending of brightness and chroma in the seam area. The pre-trained boundary correction model in the cloud generates compensation parameters optimized by physical constraints, reducing the standard deviation of brightness and color deviation of the splicing seam to within a preset threshold.

[0076] Furthermore, referring to Figure 2 The second embodiment of the data merging and scheduling method of this application provides a flowchart, based on the above. Figure 2 The embodiment shown further refines the step S10, "acquiring splicing screen data and performing single-screen basic calibration of the screen module based on the splicing screen data," including steps A201-A204:

[0077] Step A201: Obtain the current brightness value, current chromaticity value, and first working time of each screen module;

[0078] Step A202: Upload the current brightness value, current chromaticity value, and first working time to the cloud;

[0079] Step A203: Obtain the brightness gain coefficient and / or chromaticity gain coefficient of the corresponding screen module through the pixel aging model pre-trained in the cloud.

[0080] Step A204: Receive the brightness gain coefficient and / or chromaticity gain coefficient to perform single-screen basic calibration on the corresponding screen module, and store the calibration data after single-screen basic calibration.

[0081] It should be noted that in this embodiment, the current brightness value refers to the pixel-level luminous intensity data of the screen module collected in real time by the optical sensor, used to quantify the actual output luminous flux of the LED. The current chromaticity value can be represented as (x, y) coordinates in the CIE 1931 chromaticity coordinate system, which can be obtained by sampling at the center and four corners of the module using a colorimeter, reflecting the accuracy of color reproduction. The first working time refers to the cumulative running time of the module from activation to the present (unit: hours), stored in the encrypted EEPROM of the receiving card, used to quantify the aging degree of the LED chip. The pixel aging model specifically refers to a machine learning model (such as a spatiotemporal convolutional network) deployed in the cloud, which establishes a prediction function for brightness decay and chromaticity drift by analyzing massive amounts of historical aging data. The brightness gain coefficient and chromaticity gain coefficient are scalar parameters in the range of 0.8-1.2, used to compensate for the performance degradation of the LED due to aging, and their application must satisfy the formula: Output value = Input value × Gain coefficient × Temperature compensation factor (T).

[0082] In this embodiment, the highly complex aging modeling is integrated into the cloud GPU cluster, which significantly improves the processing speed compared to the local solution; the pixel aging model is trained by aggregating historical data across modules, which can improve the compensation accuracy; and the gain coefficients processed and distributed from the cloud can dynamically adapt to real-time changes and be updated online.

[0083] In one possible implementation, temperature data is acquired via a temperature sensor, while simultaneously triggering an optical sensor to collect brightness values ​​at positioning points. The first working time is read from the encrypted storage area of ​​the receiver card / storage chip. The edge gateway uploads Zstandard compressed data packets via the 5G network, which are then parsed in the cloud and input into the pixel aging model. Based on historical pixel data stored in the cloud, the gain is predicted: for example, a gain of 1.08 for the red channel to compensate for red light attenuation at high temperatures, a gain of 0.95 for the green channel to suppress color shift over long-term use, and a gain of 1.12 for the blue channel to combat rapid blue light attenuation. During the compensation period, the local receiver card triggers a buffer switching command, seamlessly switching to the new parameters when the display refreshes, ensuring uniform brightness of the compensated screen module. Furthermore, the calibration data storage can adopt a partitioned block strategy: the basic gain coefficient is stored in the receiver card's Flash memory, and the dynamic compensation parameters are temporarily stored in the FPGA cache, supporting power-off protection and fast rollback.

[0084] Furthermore, referring to Figure 3 The third embodiment of the data merging and scheduling method in this application provides a flowchart, based on the above. Figure 3 The embodiment shown further refines the step S20, "after the single-screen basic correction is completed, monitor whether there is boundary mismatch through edge computing nodes," including steps A301-A305:

[0085] Step A301: Extract the brightness correction value of each pixel within the target area of ​​the splicing screen module;

[0086] It should be noted that the target area refers to a rectangular area centered on the seam, extending 2-3 columns of pixels to the left and right of the adjacent modules. This range covers the optical transition zone that is sensitive to the human eye. The brightness correction value is the pixel brightness data (unit: nit) after single-screen basic correction, which is collected in real time by the edge computing node through the optical sensor and stored in the local cache. The edge computing node reads data from the receiver card's dual-buffer memory during the vertical blanking period to ensure that the real-time display refresh rate is not affected. The vertical blanking period refers to the interval between the electron beam returning from the bottom to the top of the screen after the display device completes the scanning of one frame. By extracting full pixel data from the seam area instead of sampled data, the missed detection caused by sampling is avoided, providing complete input for subsequent statistical calculations and solving the problem of insufficient compensation accuracy caused by data silos.

[0087] Step A302: Calculate the average corrected brightness of the target area based on the brightness correction value;

[0088] It should be noted that the average corrected brightness is the arithmetic mean of the brightness correction values ​​of all pixels within the target area, used to quantify the overall brightness level of that target area. Edge computing nodes can employ parallel accumulator hardware units such as FPGAs for rapid calculation. This average corrected brightness value participates in the calculation of the brightness standard deviation, serving as the overall brightness benchmark for the area, eliminating single-point noise interference, and significantly improving detection reliability.

[0089] Step A303: Calculate the standard deviation of the brightness of the target area using the average corrected brightness.

[0090] It should be noted that the standard deviation of brightness is used to quantify the dispersion of pixel brightness, directly reflecting the visual uniformity of the seams. The formula for the standard deviation of brightness is: Where N is the number of samples, Let represent the brightness correction value of the i-th pixel, μ represent the sample mean, and is the average corrected brightness of all pixels within the target area.

[0091] Step A304: Compare the brightness standard deviation with a preset brightness standard deviation threshold;

[0092] It should be noted that the preset brightness standard deviation threshold is a critical value set based on the visual sensitivity of the human eye. When the detected value exceeds this threshold, it is judged as a boundary mismatch. Edge computing nodes support dynamic threshold adjustment: for example, the threshold is set to 8% in high brightness mode (greater than 5000 nits) and 5% in low brightness mode (less than 1000 nits); the preset brightness standard deviation threshold can also be adjusted according to the display scenario. The comparison between the brightness standard deviation and the preset brightness standard deviation threshold can be performed in real time through a hardware comparator.

[0093] Step A305: If the brightness standard deviation is greater than the preset brightness standard deviation threshold, then a boundary mismatch is determined, and the target area is marked as an abnormal area; otherwise, the brightness correction value is maintained.

[0094] It should be noted that when the brightness standard deviation exceeds the standard, the edge computing node triggers anomaly marking and initiates the data compression process; otherwise, the current correction parameters are maintained; at the same time, it ensures that only abnormal data is uploaded to the cloud, improving resource utilization. The marked area coordinates will serve as input for subsequent cloud compensation, forming a closed loop of detection-decision-execution.

[0095] This embodiment quantifies the brightness uniformity of the splicing seam area to objectively determine whether there is boundary mismatch (i.e., brightness deviation of edge pixels of adjacent modules).

[0096] In one specific implementation, the edge computing node extracts the brightness correction values ​​of the three columns of pixels on each side of the third row of the horizontal seam and reads the data from the double buffer memory; calculates the average brightness μ=4850nit, and then calculates the brightness standard deviation σ=7.8%; compares the brightness standard deviation σ with the preset brightness standard deviation threshold of 5%, triggering an over-standard judgment; marks the area "HS3" as abnormal, generates compressed mask data, and the edge computing node uploads the boundary data of the abnormal area (including brightness correction values, average brightness, standard deviation, etc.) to the cloud. The boundary correction model in the cloud generates boundary correction parameters and performs cross-screen compensation to eliminate the brightness deviation of the splicing seam.

[0097] Furthermore, referring to Figure 4 The fourth embodiment of the data merging and scheduling method in this application provides a flowchart, based on the above. Figure 4 The embodiment shown further refines the step S20, "after the single-screen basic calibration is completed, monitor whether there is boundary mismatch through edge computing nodes," and also includes steps A401-403:

[0098] Step A401: Extract the chromaticity correction value of each pixel within the target area of ​​the splicing screen module;

[0099] It should be noted that the chromaticity correction value is the actual chromaticity data of each pixel after the screen module has undergone single-screen basic correction, such as the (x,y) coordinates in the CIE 1931 chromaticity coordinate system, which is collected and stored by the edge computing node through a colorimeter.

[0100] Step A402: Calculate the color deviation based on the color correction value and the pre-stored reference color value;

[0101] It should be noted that the reference chromaticity value refers to a preset standard chromaticity value, stored in the memory of the edge computing node, serving as a benchmark for judging chromaticity consistency. Color deviation is the difference between the chromaticity correction value of a pixel within the target area and the reference chromaticity value. Commonly used quantification metrics include CIE ΔE or the sum of the absolute values ​​of Δx and Δy. The purpose of calculating color deviation is to quantify chromaticity deviation and determine whether boundary mismatch exists. For example, ΔE can be calculated using the CIE 2000 formula: , where ΔL, ΔC, and ΔH are the differences in brightness, chroma, and hue, respectively, ΔL*, ΔC*, and ΔH* are the corresponding thresholds, and R_T is the rotation term.

[0102] Step A403: If the color deviation is greater than a preset color deviation threshold, then the target area is marked as an abnormal area; otherwise, the color correction value is maintained.

[0103] It's important to note that the preset color deviation threshold is a critical value set based on the characteristics of human vision and stored in the edge computing node's memory. Abnormal areas refer to target areas where the color deviation exceeds the threshold; these are marked and uploaded to the cloud for cross-screen compensation. If the color deviation exceeds the preset threshold, it indicates that the color deviation of adjacent module edges has exceeded the acceptable range for the human eye; otherwise, the current color correction value is maintained, indicating good color consistency. For example, when the calculated color deviation ΔE=4 is greater than the preset threshold 3, the edge computing node marks the target area as an abnormal area, stores its coordinates, color deviation value, and other information for uploading to the cloud. The cloud-based color compensation model generates color compensation parameters based on this data. After receiving these parameters, the edge computing node performs cross-screen compensation—adjusting the color correction values ​​of adjacent module edge pixels to make the color deviation less than the preset threshold, thereby eliminating color discontinuities at the seams and ensuring long-term color consistency. Furthermore, maintaining the color correction value does not mean no adjustment at all; rather, it means maintaining the current correction parameters as long as the color deviation does not exceed the limit, reducing cloud load and improving resource utilization efficiency.

[0104] In one possible implementation, the selection of the target area can be adjusted according to the pixel pitch of the splicing screen; the preset color deviation threshold can also be adjusted according to the display scenario.

[0105] Furthermore, referring to Figure 5 The fifth embodiment of the data merging and scheduling method in this application provides a flowchart, based on the above. Figure 5 The embodiment shown includes a luminance compensation model and a chrominance compensation model; the step S40, "obtaining boundary correction parameters based on the pre-trained boundary correction model in the cloud," is further refined into steps A501-A503:

[0106] Step A501: Receive the boundary data and preprocess the boundary data;

[0107] It should be noted that boundary data refers to data related to abnormal areas uploaded by edge computing nodes, including brightness correction values ​​of pixels within the target area, pixel position coordinates, and the first working time information of the screen module. This data is used to determine boundary mismatch and is also the input for generating compensation parameters in the cloud. Preprocessing is the process of cleaning and standardizing boundary data. Its purpose is to remove noise from the data and unify the data format, such as normalizing brightness values ​​to the 0-1 range to ensure the accuracy of feature extraction. For example, a Gaussian filter is used to remove high-frequency noise from the brightness correction values, and min-max normalization is used to map the brightness values ​​to the 0-1 interval, avoiding inaccurate feature extraction due to data noise.

[0108] Step A502: Extract the brightness gradient of adjacent pixels and the aging coefficient of the screen module based on the preprocessed boundary data;

[0109] It should be noted that the adjacent pixel brightness gradient is the rate of change of brightness values ​​of adjacent pixels within the target area, reflecting the degree of brightness jump at the splicing seam. The screen module aging coefficient is a decay coefficient calculated based on the first working time of the screen module, reflecting the degree of brightness decay of the LED beads due to long-term operation.

[0110] Step A503: Input the brightness gradient of adjacent pixels and the screen module aging coefficient into the brightness compensation model to generate the brightness compensation gradient field of the corresponding abnormal area; the brightness compensation gradient field is the brightness correction parameter of the boundary correction parameter.

[0111] It should be noted that the brightness compensation model is a pre-trained machine learning model in the cloud (such as a Convolutional Neural Network (CNN) or a Visual Transformer). Its inputs are the brightness gradients of adjacent pixels and the aging coefficient of the screen module, and its output is the brightness compensation gradient field of the abnormal area. This model learns from a large amount of historical splicing seam correction data to establish a coordinated compensation for brightness jumps and aging decay, solving the problem in existing technologies where the limited computing power of local algorithms prevents the running of complex models. The brightness compensation gradient field is the distribution of brightness compensation parameters for each pixel in the abnormal area. Essentially, it reduces the brightness gradient at the splicing seam to a level imperceptible to the human body by adjusting the brightness driving current of each pixel.

[0112] In this embodiment, by preprocessing and cleaning boundary data, the adjacent pixels reflecting brightness jumps and the screen module aging coefficient reflecting aging decay are extracted. Both are input into a pre-trained brightness compensation model in the cloud to generate an accurate brightness compensation gradient field. This solves the data silo problem and utilizes cloud computing power to run complex boundary correction models, avoiding the constraints of local computing power.

[0113] In one possible implementation, preprocessing includes removing random noise from the brightness correction values ​​using a Gaussian filter, and then mapping the brightness values ​​to the 0-1 range using min-max normalization to ensure consistent data scale. The brightness gradient of adjacent pixels can be calculated using the Sobel operator to obtain the gradient magnitude; the screen module aging coefficient is calculated using an exponential decay model based on the first operating time. The training steps of the brightness compensation model include using a 3-layer convolutional neural network (CNN). The input layer is set to 2 channels, fusing the brightness gradients of adjacent pixels and the aging coefficient of the screen module to form a fused feature map. The intermediate layers consist of 2-3 convolutional layers, each containing 32 3*3 convolutional kernels, with ReLU activation function used to extract spatial features. A batch normalization layer is added after the convolutional layers to accelerate training convergence, and residual connections are added to avoid the gradient vanishing problem in deep networks. The output layer consists of 1 convolutional layer with 1*1 convolutional kernels, using Sigmoid activation function to output a brightness compensation gradient field, such as 1080×4×1, with each pixel corresponding to a compensation coefficient of 0-1. Through inverse normalization, a brightness compensation coefficient of 0.9-1.1 is obtained. To ensure the accuracy (reducing the standard deviation of brightness) and smoothness (avoiding new abrupt changes) of the model's predicted compensation gradient field, a combination of MSE (Mean Squared Error) and gradient loss is used as the loss function: MSE loss measures the difference between the predicted compensation gradient field and the true label; gradient loss adds a visual smoothing constraint, making the changes in compensation coefficients smoother. Furthermore, after model deployment, continuous optimization can be achieved through online learning: for example, collecting compensation effect data from edge computing nodes and periodically fine-tuning the model using new data to improve its timeliness.

[0114] The training process of the brightness compensation model utilizes large-scale labeled data, CNN processing of spatial features, and a loss function-guided approach that balances accuracy and smoothness, ultimately achieving a precise mapping from boundary data to the brightness compensation gradient field. Through cloud computing power and deep learning models, it can quickly adapt to different specifications of splicing screens, demonstrating strong versatility, and achieving high-precision, long-term brightness compensation of splicing seams on local splicing screens.

[0115] In one specific implementation, assuming the training data contains 100,000 abnormal region data points, one of the data points has the following input features: adjacent pixel brightness gradient of 0.2, aging coefficient of 0.3679, and a compensation coefficient of 1.1 as the true label. After model training, inputting this feature results in a predicted compensation coefficient of 1.08. After application, the brightness standard deviation of the abnormal region decreases from 8% to 2%, which is below the brightness standard deviation threshold of 5%, achieving a "visually seamless" effect.

[0116] Furthermore, referring to Figure 6 The sixth embodiment of the data merging and scheduling method of this application provides a flowchart, based on the above. Figure 6 The embodiment shown includes a luminance compensation model and a chrominance compensation model; the step S40, "obtaining boundary correction parameters based on the pre-trained boundary correction model in the cloud," is further refined to include steps A601-A603:

[0117] Step A601: Receive the boundary data, and perform three-color separation and feature extraction based on the boundary data;

[0118] It should be noted that boundary data refers to the abnormal area-related data uploaded by edge computing nodes. HIA includes the chromaticity correction values ​​of pixels within the target area, as well as pixel position coordinates and the first working time of the screen module; three-color separation decomposes the chromaticity data into the chromaticity values ​​of the red (R), green (G), and blue (B) channels, such as the chromaticity coordinates of the R, G, and B components of each pixel. Since the colors displayed by LEDs are composed of three-color LED beads, and because the decay rates of the RGB LEDs are different, channel-specific processing can accurately solve single-channel chromaticity deviations; feature extraction is... Features related to chromaticity deviation are extracted from each channel, including: 1) Chromaticity deviation (the difference between the current chromaticity value and the pre-stored reference chromaticity value, e.g., the current x=0.65 for the red channel, the reference x=0.62, and the deviation is 0.03); 2) Chromaticity gradient (the rate of change of chromaticity values ​​between adjacent pixels, reflecting the degree of chromaticity jump at the seam, e.g., the x-deviation of the red channel between horizontally adjacent pixels is 0.02 / pixel); 3) Channel aging coefficient (the attenuation coefficient of each channel calculated based on the first working time, e.g., the blue channel ages faster, with an aging coefficient of 0.85). The extraction of these features provides accurate input for subsequent channel-specific compensation, solving the chromaticity discontinuity problem caused by the inability of basic correction to integrate cross-screen three-color features.

[0119] Step A602: Input the data extracted from the features into the color compensation model, and calculate the red / blue / green compensation coefficients for each channel.

[0120] It should be noted that the chromaticity compensation model is a pre-trained machine learning model in the cloud, such as a convolutional neural network (CNN). Its input is the features of each channel (chromaticity deviation, gradient, aging coefficient), and its output is the compensation coefficient for the corresponding channel. Channel-by-channel calculation accurately compensates for the independent chromaticity deviation of each channel, so that the overall color returns to the baseline.

[0121] Step A603: Based on white balance constraints, color gamut constraints, and visual smoothing constraints, the red / blue / green compensation coefficients are synergistically optimized to obtain the final gain adjustment coefficient; the final gain adjustment coefficient is the chromaticity correction parameter of the boundary correction parameter.

[0122] It's important to note that white balance constraints mean the compensated RGB color combination must conform to a pre-stored baseline white balance to avoid overall color shift after compensation, such as excessive red channel compensation causing a reddish tint. Color gamut constraints mean the chromaticity value of each compensated channel cannot exceed the color gamut range of the LED module to prevent color distortion. Visual smoothness constraints mean the rate of change of compensation coefficients between adjacent pixels cannot exceed a preset threshold to avoid chromaticity banding after compensation. For example, a sudden change in the blue channel compensation coefficient from 0.98 to 1.05 between adjacent pixels would cause the human eye to perceive a boundary, thus avoiding a "patchwork" effect caused by excessive single-point compensation. Co-optimization, under these constraints, uses nonlinear programming (such as sequential quadratic programming) or genetic algorithms to find the optimal final gain adjustment coefficient, resolving chromaticity deviations in each channel while ensuring normal color and no banding.

[0123] In one possible implementation, the three-color separation is achieved by directly extracting the three-color values ​​from the sensor data; feature extraction is performed by calculating the chromaticity deviation of each channel, the chromaticity gradient of adjacent pixels (e.g., the x-deviation of horizontally adjacent pixels divided by the pixel spacing), and the channel aging coefficient (based on an exponential decay model of the first working time); the chromaticity compensation model can be a 3-layer CNN, with the input being three features (deviation, gradient, and aging coefficient) for each channel, and the output being a chromaticity compensation coefficient; the co-optimization uses a nonlinear programming algorithm, with the objective function being to minimize the sum of squares of the chromaticity deviations of each channel, and the constraints including: 1) white balance constraint; 2) color gamut constraint; 3) visual smoothing constraint.

[0124] Further, after receiving the boundary correction parameters and performing cross-screen compensation, the process includes:

[0125] Verify whether the standard deviation of brightness and / or color deviation of abnormal areas after cross-screen compensation meets the standards;

[0126] If the brightness standard deviation and / or color deviation do not meet the standards, the non-compliant abnormal areas are marked as secondary abnormal areas, and cloud-based diagnosis or strategy iteration is performed on the secondary abnormal areas.

[0127] It should be noted that cross-screen compensation involves edge computing nodes receiving boundary correction parameters (such as brightness compensation gradient field and final gain adjustment coefficient) generated in the cloud. This allows them to adjust the brightness and / or chromaticity values ​​of pixels at the edges of adjacent screen modules, eliminating brightness jumps and chromaticity breaks at the seams and achieving a "visually seamless" display. Verification can be performed by the edge computing nodes. This verification checks whether cross-screen compensation has effectively resolved the boundary mismatch issue. If the brightness standard deviation and / or color deviation still exceed the preset threshold, the compensation is incomplete and requires further processing. This ensures that the display quality of the spliced ​​screen meets the user's requirement for a "seamless" display. Secondary abnormal areas refer to areas where the brightness standard deviation and / or color deviation still do not meet the standards after cross-screen compensation. These areas serve as the target for subsequent cloud-based diagnostics and strategy iterations. Cloud-based diagnostics involves the cloud system analyzing relevant data in secondary anomaly regions, such as the completeness of boundary data, the accuracy of model input features, and the rationality of compensation parameters, to identify the reasons for non-compliance. For example, the lack of "first working time" in boundary data may lead to incorrect calculation of the aging coefficient, or model overfitting may cause compensation parameters to exceed the reasonable range. Strategy iteration involves adjusting the parameters or algorithms of the boundary correction model based on the diagnostic results, such as retraining the model or optimizing co-optimization constraints, to improve the model's adaptability to complex scenarios.

[0128] This application also provides a cloud-based correction device for LED splicing screens; please refer to [reference needed]. Figure 7 The LED splicing screen cloud correction device includes:

[0129] The data acquisition and processing module 10 is used to acquire splicing screen data and perform single-screen basic calibration of the screen module based on the splicing screen data.

[0130] The boundary monitoring module 20 is used to monitor whether there is boundary mismatch through edge computing nodes after the single-screen basic calibration is completed; the boundary mismatch includes the brightness and / or color deviation of the edge pixels of adjacent screen modules;

[0131] The data communication module 30 is used to upload boundary data to the cloud through the edge computing node if boundary mismatch exists;

[0132] The correction execution module 40 is used to obtain boundary correction parameters based on a pre-trained boundary correction model in the cloud, and to receive the boundary correction parameters for cross-screen compensation.

[0133] The LED splicing screen cloud calibration device provided in this application, employing the LED splicing screen cloud calibration method in the above embodiments, can solve the technical problem of failing to meet the requirements for high-precision and long-term display screen splicing seam elimination. Compared with the prior art, the beneficial effects of the LED splicing screen cloud calibration device provided in this application are the same as those of the LED splicing screen cloud calibration method provided in the above embodiments, and other technical features in the LED splicing screen cloud calibration device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0134] This application provides an LED splicing screen cloud calibration device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the LED splicing screen cloud calibration method in the above embodiment 1.

[0135] The following is for reference. Figure 8 This document illustrates a structural schematic diagram of an LED splicing screen cloud calibration device suitable for implementing embodiments of this application. The LED splicing screen cloud calibration device in this application embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 8 The LED splicing screen cloud correction device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0136] like Figure 8As shown, the LED video wall cloud calibration device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the LED video wall cloud calibration device. The processing unit 1001, the read-only memory 1002, and the random access memory 1004 are interconnected via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems can be connected to the input / output interface 1006: input devices 1007 including, for example, touch screens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the LED video wall cloud calibration device to communicate wirelessly or wiredly with other devices to exchange data. Although the figures show LED video wall cloud calibration devices with various systems, it should be understood that implementation or possession of all the systems shown is not required. More or fewer systems may be implemented alternatively.

[0137] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0138] The LED video wall cloud calibration device provided in this application, employing the LED video wall cloud calibration method described in the above embodiments, can solve the technical problem of failing to meet the requirements for high-precision and long-term elimination of display screen seams. Compared with the prior art, the beneficial effects of the LED video wall cloud calibration device provided in this application are the same as those of the LED video wall cloud calibration method provided in the above embodiments, and other technical features of this LED video wall cloud calibration device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0139] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0140] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0141] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the LED splicing screen cloud correction method in the above embodiments.

[0142] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0143] The aforementioned computer-readable storage medium may be included in the LED splicing screen cloud calibration device; or it may exist independently and not be assembled into the LED splicing screen cloud calibration device.

[0144] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the LED splicing screen cloud calibration device, the LED splicing screen cloud calibration device: acquires splicing screen data and performs single-screen basic calibration of the screen modules based on the splicing screen data; after the single-screen basic calibration is completed, it monitors for boundary mismatch through edge computing nodes; the boundary mismatch includes deviations in brightness and / or chromaticity of edge pixels of adjacent screen modules; if boundary mismatch exists, it uploads boundary data to the cloud through the edge computing nodes; based on a pre-trained boundary calibration model in the cloud, it obtains boundary calibration parameters and receives the boundary calibration parameters for cross-screen compensation. This application eliminates visible splicing seams in splicing screens through cloud-based intelligent compensation, improves display quality, dynamically compensates for the influence of environmental factors, improves the adaptability of splicing screens, and combines edge computing with cloud models to improve resource utilization efficiency and reduce operation and maintenance costs.

[0145] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0146] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0147] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0148] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described LED splicing screen cloud calibration method. This solves the technical problem of failing to meet the requirements for high-precision, long-lasting elimination of display screen seams. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the LED splicing screen cloud calibration method provided in the above embodiments, and will not be elaborated upon here.

[0149] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the LED splicing screen cloud correction method described above.

[0150] The computer program product provided in this application can solve the technical problem of failing to meet the requirements for high-precision and long-term elimination of splicing seams in displays. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the cloud-based calibration method for LED splicing screens provided in the above embodiments, and will not be repeated here.

[0151] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A cloud-based calibration method for LED splicing screens, characterized in that, The LED splicing screen cloud calibration method includes: Acquire splicing screen data, and perform single-screen basic calibration of screen modules based on the splicing screen data; wherein, the single-screen basic calibration refers to the process of compensating for pixel-level brightness and color deviation within a single screen module to make it reach a preset standard value. After the single-screen basic calibration is completed, edge computing nodes are used to monitor whether there is boundary mismatch; the boundary mismatch includes deviations in brightness and / or chromaticity of edge pixels of adjacent screen modules; If a boundary mismatch exists, the boundary data is uploaded to the cloud via the edge computing node; Based on a pre-trained boundary correction model in the cloud, boundary correction parameters are obtained, and cross-screen compensation is performed using the boundary correction parameters; wherein, the boundary correction model includes a brightness compensation model and a chromaticity compensation model; The steps for obtaining boundary correction parameters based on the pre-trained boundary correction model in the cloud include: Receive the boundary data and preprocess the boundary data; Extracting the brightness gradient of adjacent pixels and the aging coefficient of the screen module based on preprocessed boundary data; The brightness gradient of adjacent pixels and the aging coefficient of the screen module are input into the brightness compensation model to generate a brightness compensation gradient field for the corresponding abnormal region; the brightness compensation gradient field is the brightness correction parameter of the boundary correction parameter.

2. The LED splicing screen cloud calibration method as described in claim 1, characterized in that, The steps of acquiring splicing screen data and performing single-screen basic calibration of the screen module based on the splicing screen data include: Obtain the current brightness value, current chromaticity value, and first working time of each screen module; Upload the current brightness value, current chromaticity value, and first working time to the cloud. The brightness gain coefficient and / or chromaticity gain coefficient of the corresponding screen module are obtained by using a pixel aging model pre-trained in the cloud. The system receives the brightness gain coefficient and / or chromaticity gain coefficient to perform single-screen basic calibration on the corresponding screen module, and stores the calibration data after single-screen basic calibration.

3. The LED splicing screen cloud calibration method as described in claim 2, characterized in that, The step of monitoring for boundary mismatch through edge computing nodes after the single-screen basic calibration is completed includes: Extract the brightness correction value of each pixel within the target area of ​​the splicing screen module; Calculate the average corrected brightness of the target area based on the brightness correction value; The standard deviation of the brightness of the target area is calculated using the average corrected brightness. The brightness standard deviation is compared with a preset brightness standard deviation threshold. If the brightness standard deviation is greater than the preset brightness standard deviation threshold, then a boundary mismatch is determined, and the target area is marked as an abnormal area; otherwise, the brightness correction value is maintained.

4. The LED splicing screen cloud calibration method as described in claim 2, characterized in that, The step of monitoring for boundary mismatch via edge computing nodes after the single-screen basic calibration is completed also includes: Extract the chromaticity correction value of each pixel within the target area of ​​the splicing screen module; Based on the chromaticity correction value and the pre-stored reference chromaticity value, the color deviation is calculated; If the color deviation is greater than a preset color deviation threshold, the target area is marked as an abnormal area; otherwise, the color correction value is maintained.

5. The LED splicing screen cloud calibration method as described in claim 4, characterized in that, The boundary correction model includes a luminance compensation model and a chrominance compensation model; the step of obtaining the boundary correction parameters based on the pre-trained boundary correction model in the cloud further includes: Receive the boundary data, and perform three-color separation and feature extraction based on the boundary data; The data obtained from feature extraction is input into the color compensation model, and the red / blue / green compensation coefficients are calculated for each channel. The red / blue / green compensation coefficients are synergistically optimized based on white balance constraints, color gamut constraints, and visual smoothing constraints to obtain the final gain adjustment coefficient; the final gain adjustment coefficient is the chromaticity correction parameter of the boundary correction parameter.

6. The LED splicing screen cloud calibration method as described in claim 5, characterized in that, After receiving the boundary correction parameters and performing cross-screen compensation, the following steps are included: Verify whether the standard deviation of brightness and / or color deviation of abnormal areas after cross-screen compensation meets the standards; If the brightness standard deviation and / or color deviation do not meet the standards, the non-compliant abnormal areas are marked as secondary abnormal areas, and cloud-based diagnosis or strategy iteration is performed on the secondary abnormal areas.

7. A cloud-based calibration device for LED splicing screens, characterized in that, The LED splicing screen cloud correction device includes: The data acquisition and processing module is used to acquire splicing screen data and perform single-screen basic calibration of the screen module based on the splicing screen data; wherein, the single-screen basic calibration refers to the process of compensating for pixel-level brightness and color deviation within a single screen module to make it reach a preset standard value. The boundary monitoring module is used to monitor whether there is boundary mismatch after the single-screen basic calibration is completed through edge computing nodes; the boundary mismatch includes the brightness and / or color deviation of the edge pixels of adjacent screen modules; The data communication module is used to upload boundary data to the cloud through the edge computing node if boundary mismatch exists; The correction execution module is used to obtain boundary correction parameters based on a pre-trained boundary correction model in the cloud, and to receive the boundary correction parameters for cross-screen compensation; wherein, the boundary correction model includes a brightness compensation model and a chromaticity compensation model; The steps for obtaining boundary correction parameters based on the pre-trained boundary correction model in the cloud include: Receive the boundary data and preprocess the boundary data; Extracting the brightness gradient of adjacent pixels and the aging coefficient of the screen module based on preprocessed boundary data; The brightness gradient of adjacent pixels and the aging coefficient of the screen module are input into the brightness compensation model to generate a brightness compensation gradient field for the corresponding abnormal region; the brightness compensation gradient field is the brightness correction parameter of the boundary correction parameter.

8. A cloud-based calibration device for LED splicing screens, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the LED splicing screen cloud correction method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the LED splicing screen cloud correction method as described in any one of claims 1 to 6.

Images