LED display screen temperature compensation method

Abstract

The invention discloses an LED display screen temperature compensation method, and relates to the technical field of LED display control. In order to solve the technical defects in the prior art that the existing temperature compensation technology lacks a pixel-level temperature compensation strategy and cannot effectively process the brightness bulge at the edge of a screen, the technical scheme provided by the invention is as follows: the method comprises the following steps of: acquiring a cold-state gray matrix and a hot-state gray matrix of the screen to form the cold-state gray matrix and the hot-state gray matrix; on the basis of the cold-state gray matrix and the hot-state gray matrix, calculating the brightness attenuation ratio of the lamp beads pixel by pixel, and generating a ratio matrix; a correction matrix is constructed and iteratively calculated, so that the thermal-state brightness matrix approaches the cold-state brightness matrix after being processed; constructing a temperature compensation curved surface of the screen, and performing normalization processing; calculating a screen edge brightness ratio, and performing mirror image filling on edge border-crossing pixels; in combination with a temperature compensation matrix, the driving current is dynamically adjusted, and the influence of temperature change on color expression is compensated.

Description

Technical Field

[0001] It relates to the field of LED display control technology, and specifically relates to a method for temperature compensation of an LED display screen. Background Art

[0002] As an important part of modern display technology, LED (light-emitting diode) display screens are widely used in fields such as advertising, sports events, stage performances, command and dispatch, etc. With the development of LED display screens towards ultra-high definition, high brightness, low power consumption, and long lifespan, the influence of temperature on the brightness, chromaticity, and uniformity of LED display screens has become increasingly prominent. During long-term operation, due to its own heat generation and changes in ambient temperature, LED chips will exhibit phenomena such as brightness attenuation, color temperature drift, and uneven brightness between modules, directly affecting the visual experience and product lifespan.

[0003] 1. Research Status of Existing Technologies Currently, for the temperature compensation technology of LED display screens, there are mainly the following methods: (1) Compensation method based on temperature sensors This method installs temperature sensors on the LED module or drive circuit to collect junction temperature data in real time and adjusts the LED drive current to compensate for brightness attenuation. For example, some high-end LED display screens use NTC thermistors or PTC resistors to measure temperature and dynamically adjust the LED drive current through PWM (pulse width modulation). However, this method has the following problems: Sensor response lag: The measurement accuracy of the temperature sensor is affected by the environment and it is difficult to reflect the temperature changes at the LED pixel level in real time.

[0004] High cost and complex installation: Each LED module needs to be equipped with a sensor, increasing the manufacturing cost and system complexity, and it is difficult to apply to large-scale LED display screens.

[0005] Difficulty in compensating for spatial temperature gradients: Even if the sensor can measure the temperature, it is difficult to accurately reflect the uneven local temperature of the LED and the brightness mutation at the module splicing location.

[0006] (2) Correction method based on global brightness compensation This method adjusts the drive current of the entire LED screen or a large area, and sets a unified compensation coefficient based on the brightness measurement results. Common implementation methods include: Adopting a full-screen brightness attenuation model, experimentally determining the change curve of the LED luminous efficiency at different temperatures, and calculating the compensation coefficient based on the curve.

[0007] Analyzing the captured image of the LED screen through an image processing algorithm, calculating the brightness change trend, and then adjusting the PWM duty cycle to compensate for brightness attenuation.

[0008] However, the following problems mainly exist in this method: It is impossible to compensate at the single module or pixel level, which easily leads to uneven brightness in local areas of the screen.

[0009] There is a lack of correction for color temperature drift. Especially, the red LED is significantly affected by temperature, and red light attenuation is likely to occur, resulting in color distortion.

[0010] The dynamic compensation ability is limited, and it is difficult to respond to temperature changes in real time, resulting in compensation lag.

[0011] (3) Temperature compensation method based on software correction Some LED display control systems use software algorithms to optimize temperature compensation. For example: Through artificial neural network (ANN) or machine learning models, based on a large amount of experimental data, train the law of LED brightness attenuation, and establish a temperature-brightness compensation model.

[0012] Adopt a correction method based on LUT (look-up table), pre-store the brightness compensation coefficients at different temperatures, and perform dynamic interpolation calculations during display.

[0013] Although these methods improve the compensation accuracy to a certain extent, there are still problems such as high dependence on training data, large computational amount, and poor real-time performance. Especially in large-size LED display screens, the computational cost is relatively high, and it is difficult to respond to temperature changes in real time.

[0014] 2. Main technical problems of the existing technology Although the existing temperature compensation technology alleviates the problems of brightness attenuation and chromaticity shift of LED display screens to a certain extent, the following unsolved core technical problems still exist: There is a lack of a temperature compensation strategy at the pixel level, resulting in uneven brightness in local areas of the LED screen and making it difficult to meet the requirements of high-precision display.

[0015] It is impossible to effectively handle the sudden change of brightness at the edge of the screen, resulting in unnatural brightness transition at the splicing part and affecting the overall display effect.

[0016] The problem of chromaticity shift has not been effectively solved. Especially, the wavelength of the red LED redshifts at high temperatures, resulting in color distortion. And the existing methods mainly focus on brightness compensation and ignore color stability.

[0017] The global temperature compensation method is difficult to adapt to the individual differences of different LED modules, resulting in inconsistent compensation effects in different areas of the same LED screen.

[0018] The existing compensation algorithms have high computational complexity and insufficient real-time performance, and cannot adjust the compensation coefficients in real time during the dynamic operation of the LED display screen. Summary of the Invention

[0019] To solve the technical deficiencies in the prior art, namely the lack of a temperature compensation strategy at the pixel level in existing temperature compensation technologies and the inability to effectively handle the brightness mutation at the screen edge, the technical solution provided by the present invention is as follows: A temperature compensation method for an LED display screen, comprising: The step of collecting the cold-state and hot-state grayscale matrices of the LED screen to form a cold-state grayscale matrix C and a hot-state grayscale matrix H; The step of calculating the brightness attenuation ratio of each LED lamp bead pixel-by-pixel based on the cold-state grayscale matrix C and the hot-state grayscale matrix H to generate a ratio matrix R; The step of constructing and iteratively calculating a correction matrix K to make the hot-state brightness matrix H approximate the cold-state brightness matrix C after being processed by K; The step of constructing a temperature compensation surface for the LED screen based on the ratio matrix R and performing normalization processing; The step of calculating the brightness ratio at the edge of the LED screen and performing mirror filling on the edge out-of-bounds pixels; The step of constructing an RGB chromaticity correction matrix in the CIE-XYZ color space and dynamically adjusting the LED drive current in combination with the temperature compensation matrix to compensate for the influence of temperature changes on the LED color performance.

[0020] Based on the same inventive concept, the present invention also provides a temperature compensation device for an LED display screen, comprising: A module for collecting the cold-state and hot-state grayscale matrices of the LED screen to form a cold-state grayscale matrix C and a hot-state grayscale matrix H; A module for calculating the brightness attenuation ratio of each LED lamp bead pixel-by-pixel based on the cold-state grayscale matrix C and the hot-state grayscale matrix H to generate a ratio matrix R; A module for constructing and iteratively calculating a correction matrix K to make the hot-state brightness matrix H approximate the cold-state brightness matrix C after being processed by K; A module for constructing a temperature compensation surface for the LED screen based on the ratio matrix R and performing normalization processing; A module for calculating the brightness ratio at the edge of the LED screen and performing mirror filling on the edge out-of-bounds pixels; A module for constructing an RGB chromaticity correction matrix in the CIE-XYZ color space and dynamically adjusting the LED drive current in combination with the temperature compensation matrix to compensate for the influence of temperature changes on the LED color performance.

[0021] Based on the same inventive concept, the present invention also provides a computer storage medium for storing a computer program, and when the computer program is read by a computer, the computer executes the above method.

[0022] Based on the same inventive concept, the present invention also provides a computer, including a processor and a storage medium. When the processor reads the computer program stored in the storage medium, the computer executes the method described above.

[0023] Based on the same inventive concept, the present invention also provides a computer program product. As a computer program, when the computer program is executed, the method described above is implemented.

[0024] Compared with the prior art, the beneficial effects of the technical solution provided by the present invention are as follows: This solution adopts the method of dual-state grayscale matrix acquisition and ratio analysis, uses a high-resolution industrial camera to collect the brightness matrix of the LED display screen in the cold state and the hot state, and calculates the ratio matrix to quantify the temperature influence. Compared with the traditional global brightness compensation, this method can achieve pixel-level brightness attenuation measurement, accurately depict the local temperature characteristics of the LED screen, make the compensation more accurate, and avoid the local unevenness phenomenon caused by global compensation.

[0025] This solution optimizes and corrects the matrix through gradient descent, adjusts the temperature compensation coefficient based on the Nesterov accelerated optimization algorithm, so that the brightness matrix of the LED screen in the hot state is as close as possible to the cold state brightness distribution. Compared with the LUT (look-up table) correction method, this method can adapt to the dynamic changes under different temperature gradients, ensure the brightness uniformity of the LED screen after long-term operation, and improve the real-time performance and adaptability of the compensation.

[0026] This solution uses Makima interpolation to smooth the compensation matrix, performs interpolation calculation for the temperature peak region of each LED module, and generates a smooth temperature compensation surface. Compared with the traditional cubic spline interpolation, this method avoids the overshoot phenomenon, can better fit the temperature change curve of the LED display screen, ensures the smoothness of the compensation coefficient distribution, and improves the brightness stability of the LED screen in different temperature environments.

[0027] This solution adopts the method of logarithmic ratio analysis and exponential interpolation in the processing of edge pixels, and performs non-linear smoothing processing on the brightness change at the splicing edge of the LED screen. Compared with the traditional linear interpolation method, this method can more naturally eliminate the brightness mutation at the module splicing, make the overall brightness transition of the LED display screen more uniform, avoid the visual defects caused by edge brightness mismatch, and improve the splicing display effect.

[0028] This solution constructs an RGB adjustment coefficient matrix in the CIE-XYZ color space through chromaticity correction fusion and performs color temperature consistency adjustment in combination with a temperature compensation matrix. Compared with the existing method that only performs brightness compensation, this method can simultaneously consider the influence of temperature on the color of LEDs. Especially for the problem of wavelength red shift of red LEDs at high temperatures, it compensates for their color drift, enabling the LED display screen to maintain color stability under different temperature conditions and improving the display quality.

[0029] This solution uses FPGA hardware optimization for calculation, and in the compensation algorithm, it uses SVD (Singular Value Decomposition) noise reduction and low-rank matrix approximation, significantly reducing the computational complexity. Compared with the temperature compensation methods based on deep learning or complex neural network models, this solution has higher computational efficiency, can adjust the compensation coefficients in real time during the dynamic operation of the LED screen, avoid the lag problem caused by large computational amounts of traditional algorithms, and improve the real-time performance and engineering applicability of the compensation system.

[0030] It can be used for brightness uniformity correction, chromaticity compensation, and splicing optimization of large LED display screens, and is widely applicable to high-precision display applications such as advertising displays, stage performances, command and dispatch, and sports events. Brief Description of the Drawings

[0031] Figure 1 It is a schematic flow diagram of a temperature compensation method for an LED display screen.

[0032] Figure 2 It is a schematic diagram of the dedicated temperature compensation coefficient for the entire screen.

[0033] Figure 3 It is a schematic diagram of the method for extracting the module temperature peak value.

[0034] Figure 4 It is a schematic diagram of the interpolation of the general compensation coefficient surface.

[0035] Figure 5 It is a schematic diagram for comparing the similarity of the standard temperature compensation coefficients of the horizontal single module and the vertical single module.

[0036] Figure 6 It is a comparison diagram of the cold screen brightness (left) and the hot equilibrium temperature compensation brightness (right). Detailed Embodiments

[0037] To make the advantages and beneficial effects of the technical solution provided by the present invention more clearly manifested, the technical solution provided by the present invention will now be further described in detail with reference to the accompanying drawings. Specifically: Embodiment 1. In this embodiment, the technical solution will be further explained and described in detail through specific steps: A temperature compensation method for an LED box display screen, and the method is as follows: Step 1: Start the temperature compensation process and initialize system parameters, including the screen resolution , the pixel size of the module , and load the camera calibration parameters.

[0038] Step 2: Use a high-resolution industrial camera (resolution ≥ 1440×810) to collect data for the entire M×N box screen. The matrix elements are calculated through the grayscale integration formula: Place the entire screen vertically in a constant-temperature dark room (25 ± 0.5°C). After cooling to room temperature, immediately light it up and perform skip-point acquisition through a telecentric lens to generate a cold screen grayscale matrix C; maintain the white field lit until the screen temperature stabilizes (thermal equilibrium time ≥ 2 hours) to generate a hot screen grayscale matrix , calculate the ratio matrix (element-by-element division), which serves as the dedicated temperature compensation coefficient matrix (before correction) for the entire M×N box screen.

[0039] Among them, represents the pixel grayscale covered by the th LED on the camera when the display screen is at the initial or lower temperature (cold screen), represents integration, represents the grayscale function representing the "cold screen" state, represents the pixel grayscale of the same LED when the screen continuously works under a 600 nit white field display and heats up to a certain thermal equilibrium state (hot screen), represents the integral value of the grayscale function representing the "hot screen" state.

[0040] Step 3: Collect the grayscale matrix that reaches thermal equilibrium after loading the compensation coefficient . Ideally, and should be exactly the same. However, due to multiple reasons, there is an error between the two. Therefore, it is necessary to construct and optimize the correction matrix to adjust the grayscale distribution to be as close as possible to the cold-state brightness distribution . The gradient descent iteration of the correction matrix : until the convergence condition ( ) is satisfied. The gradient descent algorithm adopts Nesterov acceleration optimization, and the learning rate is dynamically adjusted: The initial learning rate , the maximum number of iterations . As the dedicated temperature compensation coefficient matrix for the M×N box full screen (after correction); Among them, denotes element-wise multiplication, is the Frobenius norm, is the adjusted balance brightness matrix, is the cold state brightness matrix, represents adjacent sub-matrix pairs, and are respectively the correction factor matrices of adjacent sub-matrices, The subscripts i and j in the lower right corner respectively represent rows and columns, and represent regularization parameters, represents updating the correction matrix to minimize the error .

[0041] Step 4: In the dedicated temperature compensation coefficient matrix for the full screen, calibrate the temperature peak region of each module (i.e., the 9×9 neighborhood mean near the peak of the ratio matrix), extract the compensation coefficients of all module peak points, and divide the array into matrix peak regions of X×Y (such as 2×2) according to the dense layout characteristics of the back driving IC. Its coordinate set is:

[0042] where X represents the number of rows with concentrated layout of driving ICs, Y represents the number of columns with concentrated layout, and in a typical scenario, the distance between adjacent peak regions is Δx = ⌊M / X⌋, Δy = ⌊N / Y⌋.

[0043] Perform a×a (such as 9×9) neighborhood averaging on each peak region. Define the compensation coefficient matrix as R, and the neighborhood coordinate set of module (m',n') in the (i,j) peak region is Ω_{m',n'}^{(i,j)}, then the compensation coefficient is calculated as:

[0044] represents (element-wise division).

[0045] Among them, the neighborhood window satisfies:

[0046] Each module (m',n') generates X×Y temperature feature points in the X×Y peak region:

[0047] Based on each module generated temperature feature points For the points with the same peak position index in all modules global interpolation is performed, and the specific process is as follows: traverse all modules and extract the temperature feature points at the same peak region index to form a data set corresponding to this position:

[0048] For example, if the total number of modules is then contains data points, and each point corresponds to the compensation coefficient value at the original screen physical position . Associate each data point inwith the physical coordinates of its module (such as pixel or driver IC coordinates) to construct a two-dimensional interpolation input:

[0049] Through the Makima interpolation algorithm, generate a continuous compensation function covering the full screen that satisfies: For peak region indices perform the above operations separately, and finally obtain independent interpolation functions:

[0050] Each characterizes the temperature compensation distribution characteristics of the full screen in the peak region .

[0051] Take the mean of all , and the calculation formula is:

[0052] Finally, normalize the interpolation surface and use it as the general compensation coefficient for the vertical M×N boxes . The Makima interpolation uses extrapolation constraints in the boundary region and defines the slope weight:

[0053] To avoid oscillations in the interpolation surface at the module splicing position, divide by , and divide the result according to the actual module division and sum each module to take the mean, then the standard temperature compensation coefficient for the vertical single module can be obtained for comparison and verification.

[0054] Step 5: Horizontally place two independent boxes in a vertical dark box (ambient light <1 lux). After cooling to room temperature, light them up synchronously, and use a line array CCD camera mounted on the top (sampling rate ≥1 kHz) to perform point-by-point scanning to record the initial cold-state brightness distribution. After reaching thermal equilibrium, repeat the acquisition and calculate the horizontal dedicated single-module compensation coefficient matrix. , and take the average value of the middle 4 modules (index range ) to generate the horizontal single-module standard compensation matrix. .

[0055] Step 6: Verify the compensation effect. Verify and optimize the verification of the similarity between the vertical single-module standard temperature compensation coefficient and the horizontal single-module standard compensation matrix . Verification index: root mean square error of brightness , Step 7: Edge pixel processing Perform non-linear smoothing on the screen edge area (outer 2 rows of pixels): Calculate the left and right edge brightness ratios to generate the exponential adjustment coefficient ( ). Gradually adjust the left / right half-row pixels:

[0056] represents the processed left half-row pixels, represents the left half-row range: column index , and its pixels are: ; represents the adjustment coefficient. For the left half-row, the adjustment coefficient increases from to ; for the right half-row, it decreases from to .

[0057] represents the processed right half-row pixels, represents the right half-row range: column index , and its pixels are: .

[0058] represents the transpose.

[0059] Step 8: Temperature compensation coefficient expansion and fusion Expand the horizontal standard compensation matrix to the full screen size according to the module period: Multiply element by element with the vertical general compensation coefficient to generate the final compensation matrix:

[0060] Step Nine: Verification of Compensation Effect Load the compensation coefficient into the display control system, specifically including: Construct a complete temperature compensation matrix . This matrix is a composite matrix composed of sub-matrices, where each sub-matrix is a matrix. The specific expression is as follows. The subscript is the row and column labels of this pixel, and the meaning of the subscript X is the XYZ component corresponding to the correction coefficient applied by this compensation coefficient. The component means that the X (red), Y (green), and Z (blue) components of the spectrum of this kind of light source are collected. For example, collecting the blue (B) component in red (R): (1) Matrix

[0061]

[0062] (2) Matrix

[0063]

[0064] (3) Matrix

[0065]

[0066] (3) Matrix

[0067]

[0068] Load the correction matrix of the temperature compensation coefficient

[0069] Define as the chromaticity correction coefficient matrix calculated by collecting the CIE-XYZ color space components corresponding to each RGB color, where the subscript is the row and column labels of this pixel and its position. The subscript RGB represents the displayed color, and the subscript XYZ represents the corresponding component, corresponding to X (red), Y (green), and Z (blue) respectively. For example means the correction coefficient of the X (red) component when displaying red, means the correction coefficient of the Y (green) component when displaying blue. is a matrix, composed of and multiply the corresponding elements point by point, and the formula is:

[0070] The specific form is as follows:

[0071]

[0072] The construction of the cold screen gray matrix C and the hot screen gray matrix H includes: collecting the pixel gray values covered by each LED lamp bead in the cold screen and hot screen states through a camera, and respectively accumulating to obtain the corresponding to each LED, and arranging them in the form of an m×n matrix, where the cold screen gray function and the hot screen gray function realize gray accumulation through double integral or discrete summation.

[0073] The construction of the ratio matrix R adopts element-by-element division operation, defined as R = C H, and its matrix element is used to quantify the brightness attenuation rate of the LED in the cold and hot states, and intuitively reflect the full-screen temperature sensitivity distribution through the matrix form.

[0074] The optimization objective function of the correction matrix K includes three constraints: the brightness error minimization term , the sub-matrix consistency term and the correction amount constraint term , where L is the Laplacian matrix, and are tuned through grid search to balance the correction accuracy and smoothness.

[0075] The sub-matrix consistency constraint is realized through the Frobenius norm difference between the adjacent module correction factors K_i,j and K_k,l, and the Laplacian matrix L is used to encode the spatial adjacency relationship between modules to ensure that there is no mutation in the brightness distribution at the splicing.

[0076] Truncate the small singular values to remove noise, and retain the first r principal components to construct a low-rank matrix X, satisfying rank(X) ≤ r, to improve the computational efficiency and robustness of the optimization process.

[0077] The temperature peak extraction adopts a modular partition strategy, divides the screen into M_x×M_y modules according to the module splicing attributes, extracts the adjacent neighborhood gray mean value as a feature point in the driving IC aggregation area of each module, and generates a continuous compensation surface of the whole screen pixels through Makima interpolation.

[0078] The Makima interpolation adopts an improved cubic spline algorithm. After establishing interpolation surfaces at different peak points of the module respectively, weighted fusion is carried out, where the weights are dynamically adjusted according to the symmetry of the module's heat distribution.

[0079] The edge pixel processing adopts a method combining logarithmic domain ratio analysis and exponential interpolation: calculate the natural logarithm r = ln(M / m) of the left - right / up - down edge brightness ratio, generate the exponential adjustment coefficient α_k = exp(kr / w), and perform non - linear gradient adjustment on half - rows / half - columns of pixels to eliminate the optical mutation at the splicing boundary.

[0080] When processing the edge, mirror filling is used to handle out - of - bounds pixels. Define G(p + i,q + j)=G(-p - i,q + j) when p + i < 0, maintaining the periodic boundary characteristics to avoid edge artifacts.

[0081] The chromaticity correction generates the final compensation matrix G by dividing the dedicated compensation matrix R_c by the general compensation surface S element - by - element, where , K is the correction matrix iteratively optimized by the gradient descent method to ensure differential compensation for the wavelength redshift of the three - primary - color LEDs caused by temperature.

[0082] The gradient descent iteration adopts an adaptive learning rate η, and the update formula is , where the gradient term includes the brightness error gradient , the adjacent sub - matrix difference gradient and the correction amount constraint gradient .

[0083] The single - module standard compensation matrix Ĝ is obtained by calculating the per - position mean of the remaining (M_x - 2)×(M_y - 2) intermediate modules after excluding the outer - ring modules, effectively eliminating the compensation deviation introduced by abnormal heat dissipation at the screen border.

[0084] The structural similarity index (SSIM) is used to evaluate the consistency of the vertical and horizontal acquisition modules. Set SSIM > 0.93 as the acceptance threshold. By comparing with the theoretical matrix , the compensation effectiveness is verified by the similarity of the brightness structure and contrast.

[0085] The temperature compensation coefficient back - calculation is realized by multiplying the single - module matrix periodically extended to the full - screen size by the general compensation curve point - by - point. The extension formula is , maintaining the periodic repetition characteristics of the module compensation mode.

[0086] The final compensation matrix It is loaded into each RGB channel through chromaticity space conversion, and a higher compensation gain is set for the red LED in the compensation matrix result to compensate for the higher temperature sensitivity of its AlGaInP material, thereby achieving color temperature consistency correction.

[0087] Embodiment 3. In combination with Figure 1 To describe this embodiment, this embodiment further describes the above-provided technical solution in detail through specific examples. Specifically: The method is as follows: First, collect and calculate the loading correction coefficients for the entire M*N box screen to provide basic data for subsequent brightness uniformity and temperature compensation; after loading the correction coefficients, place the entire screen vertically, and after it has completely cooled to room temperature, immediately perform pixel-by-pixel acquisition of each component of different colors at a long distance; after the white field is lit for a long time until thermal equilibrium, perform n*n pixel-by-pixel acquisition at a long distance; calculate the ratio of the cold-state acquisition to the hot-state acquisition to obtain the temperature compensation coefficient specific to the M*N box; calibrate the temperature peak point at x of a single module, and select the temperature compensation coefficients of all modules in the area of the same temperature peak position. Interpolate into a smooth surface according to the display pixels. Since there are peak points at x, x smooth surfaces can be obtained. Finally, calculate the average value of the x smooth surfaces as the general compensation coefficient for the M*N box. Divide the temperature compensation coefficient specific to the M*N box by the general compensation coefficient of the M*N box, and divide the result according to the actual module division and take the average value to obtain the standard temperature compensation coefficient for a single vertical module for comparison and verification.

[0088] The preparatory work for horizontal acquisition is to separately collect data for two independent boxes, calculate and load the correction coefficients. Place the screens of the two boxes horizontally inside the vertical dark box; completely cool to room temperature. Suddenly light up the two boxes and immediately perform point-by-point acquisition with the top camera; after the white field is lit for a long time until thermal equilibrium, perform point-by-point acquisition with the top camera. The ratio of the cold acquisition to the hot acquisition is the temperature compensation coefficient specific to a single horizontal module; take the average value of the middle four modules as the standard temperature compensation coefficient for a single horizontal module.

[0089] Check whether the currently calculated standard temperature compensation coefficient for a single horizontal module and the standard temperature compensation coefficient for a single vertical module meet the requirements of brightness uniformity and temperature compensation. If not, further adjustment or data re-acquisition is required. And calculate and correct the matrix K by optimizing the objective function so that the corrected hot-state brightness matrix approximates the cold-state brightness matrix C; perform separate processing on the special pixels at the edge of the screen to optimize screen splicing; Periodically extend the horizontal single-module standard temperature compensation coefficient to the horizontal single-module class compensation coefficient of the entire screen; use the vertical general compensation coefficient of the M*N box, multiply it by the horizontal single-module class standard temperature compensation coefficients of different models, and the dedicated temperature compensation coefficient for the entire vertical screen can be deduced inversely; combine the inversely deduced temperature compensation coefficient with the existing correction coefficient; verify whether the final compensation effect meets the design requirements.

[0090] Specifically, it includes: Step 1: Bimodal gray matrix acquisition and ratio analysis. At an initial temperature of 25°C, light up the screen with a 600 nit white field, use a area array CCD camera (resolution ≥ 4K) in conjunction with 650nm / 530nm / 450nm filters to collect the RGB components three times, calculate the gray integral value of the pixels covered by each LED lamp bead, and construct a cold-state matrix , satisfying: ; Keep working until the thermal equilibrium state with a temperature rise < 0.5°C / min is reached, and repeat the acquisition process to construct a hot-state matrix ; Step 2: Ratio matrix construction and correction matrix optimization calculation Calculate the cold and hot state ratios element by element , and identify the attenuation rate of each pixel. Establish the objective function: where is the Laplacian matrix, which constrains the consistency of the correction coefficients of adjacent modules; use SVD decomposition to denoise the matrix and retain the top 10% singular values; use gradient descent iteration to solve the optimal correction matrix , with the termination condition .

[0091] Step 3: Peak extraction and Makima interpolation: Divide the screen into modules, extract the average value of 9×9 pixels at the same peak position of each module as the feature points; construct a piecewise cubic polynomial along the x / y axis, satisfying: The slope calculation uses the weighted method of adjacent four points to avoid the overshoot phenomenon of traditional cubic splines; generate a smooth compensation surface with a resolution of 1440×810 .

[0092] Step 4: Chromaticity correction fusion After calculating the temperature compensation matrix , construct the chromaticity correction matrix , and each element corresponds to the RGB adjustment coefficient in the CIE-XYZ space. Multiply the temperature compensation matrix and element by element to generate a composite correction matrix: Real-time adjust the RGB drive current through PWM modulation to synchronously compensate the brightness and chromaticity.

[0093] Step 3 Optimization effect content space resolution improvement: Sub-pixel level temperature compensation is achieved through a 1440×810 compensation surface, with local brightness difference <1%; Edge transition optimization: The exponential smoothing algorithm makes the SSIM of the splicing area ≥0.95, eliminating visible mutations to the naked eye; Chromaticity consistency improvement: The ΔE color difference is reduced from 5.2 to 2.3, meeting professional display requirements.

[0094] Cold state and hot state gray level acquisition: The corresponding gray values of each LED lamp bead on the camera sensor are obtained respectively under cold state and hot state conditions of the LED display screen, and a cold state gray matrix and a hot state gray matrix are formed.

[0095] Temperature compensation coefficient calculation: Based on the cold state and hot state gray matrices, a temperature compensation coefficient matrix is calculated. This step includes calculating correction coefficients, applying Gaussian filtering, and the Makima interpolation method to optimize the distribution and accuracy of the compensation coefficients.

[0096] Temperature compensation application: By loading the temperature compensation coefficient matrix, temperature compensation is performed on the LED display screen to achieve brightness and chromaticity uniformity.

[0097] Camera imaging and exposure In a digital camera, the image sensor is responsible for converting the received optical signal into an electrical signal and obtaining digital gray values through analog-to-digital conversion. These digital gray values then enter the image signal processing pipeline and are finally stored in the form of an image file. During this process, there is an approximately linear response relationship between the output of each photosensitive pixel unit and the light flux accumulated within a certain exposure time.

[0098] Relationship between exposure and brightness: Under ideal conditions, exposure is the product of illuminance per unit area and exposure time. The true brightness of an object, the transmission coefficient of the lens, and the aperture coefficient jointly determine the exposure. By reasonably controlling the exposure time and aperture settings, it can be ensured that the camera operates within the linear effective range of the sensor, making the digital gray values and the true brightness show an approximately linear relationship.

[0099] Calculation of gray values: In the detection scenario of the LED display screen, the brightness of each LED lamp bead forms a brightness aggregation of several pixel points on the camera sensor. To evaluate the comprehensive brightness of a single LED in a digital image, it is usually necessary to superimpose or integrate the output values of all pixels within the pixel area covered by it on the sensor. In this way, the comprehensive gray values of each LED lamp bead under cold state and hot state conditions can be obtained, which can be further used for the calculation of temperature compensation coefficients.

[0100] Calculation of the whole screen dedicated temperature compensation coefficient In practical applications, the temperature of an LED display screen increases due to the working environment or its own heat generation, which in turn affects its luminous efficiency. To perform effective temperature compensation, it is necessary to construct grayscale matrices under cold and hot states and calculate the temperature compensation coefficient based on this.

[0101] Cold acquisition and hot acquisition: First, under cold state conditions (such as room temperature), capture the full-screen image of the LED display screen to form a cold-state grayscale matrix. Subsequently, under the thermal equilibrium state (such as after the white field is lit for a long time), capture the full-screen image again to form a hot-state grayscale matrix. These two grayscale matrices respectively reflect the brightness distribution of the LED display screen under different temperature conditions.

[0102] Construction of the ratio matrix: By dividing the hot-state grayscale matrix element by element with the cold-state grayscale matrix, construct the ratio matrix. This ratio matrix reflects the brightness change of the LED under hot-state conditions relative to cold-state conditions.

[0103] Calculation of the correction coefficient: After constructing the ratio matrix, it is necessary to further calculate the correction coefficient to achieve precise temperature compensation. This step involves optimizing the objective function, considering factors such as minimizing the brightness error, consistency between sub-matrices, and proximity of the correction matrix. Through optimization algorithms such as the gradient descent method, the optimal correction matrix can be iteratively calculated to optimize the temperature compensation effect.

[0104] Gaussian filtering and low-rank approximation: To improve the robustness and efficiency of the algorithm, perform Gaussian filtering on the ratio matrix to remove noise and extract the main brightness patterns. Through singular value decomposition and low-rank approximation, the matrix structure can be further simplified, reducing the computational amount and accelerating the optimization process.

[0105] Single-module temperature compensation coefficient To ensure the consistency and uniformity of the temperature compensation coefficient across the entire display screen, it is necessary to modularize the display screen.

[0106] Module division: Divide the entire display screen into multiple modules according to rows and columns. After removing the outermost modules, the number of remaining middle modules decreases, facilitating further analysis and processing. By averaging the temperature compensation coefficients of each middle module, a single-module standard temperature compensation ratio matrix can be obtained. This matrix represents the average brightness change of the entire display screen under cold and hot states.

[0107] Calculation of the single-module standard temperature compensation ratio matrix: After adding the compensation ratio matrices of all middle modules point by point at the corresponding positions and then dividing by the number of modules, the single-module standard temperature compensation ratio matrix is obtained. This matrix reflects the average brightness change at each module position of the display screen, ensuring the unity and accuracy of the compensation coefficient.

[0108] Gaussian filtering process: Apply Gaussian filtering to the single-module standard temperature compensation ratio matrix to further smooth the distribution of compensation coefficients and ensure the continuity and consistency of the compensation effect. During the filtering process, mirror padding or extrapolation methods are used to handle the boundaries to avoid distortion of edge pixels.

[0109] Surface similarity evaluation: Evaluate the similarity between the filtered compensation matrix and the theoretical reference matrix by calculating the root mean square error and the structural similarity index. If the root mean square error is lower than the preset threshold and the structural similarity index is higher than the preset standard, the temperature compensation matrix is considered to meet the design requirements.

[0110] Edge pixel processing In the scenarios of large-size images or tiled screens, the inconsistency of edge pixels may lead to poor visual effects. To ensure a natural brightness transition at the module splicing parts, effective smoothing and adjustment of edge pixels are required.

[0111] Definition and division of edge pixels: Divide the edge area of the image into the left and right half rows and the upper and lower half columns, which helps to make targeted adjustments according to the differences in brightness or color in subsequent processing. By subdividing the edge area, the adjustment range of edge pixels can be controlled more precisely.

[0112] Nonlinear edge smoothing method: Use the method of logarithmic transformation and exponential interpolation to perform nonlinear smoothing adjustment on edge pixels. First, map the edge pixel values to the logarithmic domain to enhance the stability of brightness ratio calculation. Subsequently, generate adjustment coefficients based on the logarithmic ratio and achieve a natural transition of brightness differences through exponential interpolation, avoiding the abruptness that may be brought by linear adjustment.

[0113] Determination and application of the adjustment direction: Determine the adjustment direction according to the brightness relationship of the left and right edge and upper and lower edge pixels. For the side with lower brightness, gradually increase the brightness; for the side with higher brightness, gradually decrease the brightness. In this way, ensure a smooth brightness transition at the module splicing parts and achieve the uniformity of overall brightness and chromaticity.

[0114] Embodiment 1 In this embodiment, a 12-row and 6-column LED display screen is used as the experimental object. First, the entire screen is placed vertically and cooled to room temperature (25°C). Then, the white field is lit and pixel acquisition is performed at a long distance for each pixel to form a cold-state grayscale matrix. Subsequently, the white field is kept lit for a long time until the thermal equilibrium state (about 50°C) is reached, and then every other pixel is collected to form a hot-state grayscale matrix. By calculating the ratio matrix, that is, dividing the hot-state grayscale matrix by the cold-state grayscale matrix point by point, the ratio matrix is obtained. Then, using the Gaussian filtering and Makima interpolation methods, the ratio matrix is optimized to generate a temperature compensation coefficient matrix. After loading the temperature compensation coefficient matrix, the brightness and chromaticity of the LED display screen under different temperature conditions have been significantly improved in uniformity, verifying the effectiveness of the present invention.

[0115] Embodiment 2 In this embodiment, the processing of the edge pixels of the horizontal single-module standard temperature compensation coefficient is further optimized. By performing logarithmic transformation and exponential interpolation on the edge pixels, the brightness transition at the edge of the entire screen is ensured to be smooth, avoiding the problem of sudden brightness change caused by inconsistent compensation coefficients. The specific steps include: Logarithmic ratio calculation: Map the brightness values of the edge pixels to the logarithmic domain to enhance the stability of the brightness ratio calculation.

[0116] Adjustment coefficient generation: Based on the logarithmic ratio, generate an adjustment coefficient and adopt the exponential interpolation method to achieve a natural transition of the brightness difference.

[0117] Application of the adjustment coefficient: According to the adjustment direction, apply the adjustment coefficient to the left and right edge pixels and the top and bottom edge pixels respectively to ensure a smooth brightness transition at the module splicing part.

[0118] The experimental results show that the LED display screen after the edge pixel processing has a more consistent visual effect and a more stable display effect under different temperature conditions, further improving the overall visual quality of the display screen.

[0119] Embodiment 3 In this embodiment, the application and verification of the temperature compensation method for high-resolution large-size LED display screens are carried out. The specific steps are as follows: Modular division: Divide the large-size display screen into multiple modules, and each module contains multiple LED lamp beads. After removing the outermost modules, the remaining middle modules are used for the calculation of the temperature compensation coefficient.

[0120] Temperature compensation coefficient calculation: Under cold-state and hot-state conditions, collect the grayscale values of the middle modules respectively to construct a ratio matrix. Through Gaussian filtering, generate a single-module standard temperature compensation ratio matrix.

[0121] Full - screen compensation coefficient generation: Periodically expand the single - module compensation ratio matrix row - by - row and column - by - column to generate the temperature compensation coefficient matrix for the entire display screen.

[0122] Chromaticity correction application: Load the temperature compensation coefficient matrix into the chromaticity correction matrix, and achieve full - screen temperature compensation through point - by - point multiplication.

[0123] After the above - mentioned steps of processing, the brightness and chromaticity uniformity of the high - resolution large - size LED display screen under different temperature conditions have been significantly improved, verifying the effectiveness and scalability of the method of the present invention in practical applications.

[0124] Embodiment 4 This embodiment applies the temperature compensation method to an outdoor large - size LED display screen. The outdoor environmental temperature changes greatly, and the display screen is large in size with uneven temperature distribution. This embodiment performs temperature compensation through the following steps: Multi - point acquisition: Set multiple acquisition points in different areas of the display screen, and respectively acquire the gray - scale values in the cold state and the hot state to form multiple cold - state gray - scale matrices and hot - state gray - scale matrices.

[0125] Regional compensation coefficient calculation: Calculate the corresponding temperature compensation coefficient matrix for the gray - scale value changes in different regions, considering the temperature distribution differences in each region.

[0126] Local optimization: Locally optimize the temperature compensation coefficient matrix for each region to ensure the accuracy and consistency of the compensation effect in each region.

[0127] Full - screen application: Integrate the compensation coefficient matrices of each region and apply them to the entire display screen to achieve the improvement of the overall brightness and chromaticity uniformity.

[0128] The experimental results show that the brightness and chromaticity uniformity of the outdoor large - size LED display screen after regional temperature compensation processing have been significantly improved under different environmental temperatures, the display effect is stable, and it meets the high - standard visual requirements.

[0129] Embodiment 5 This embodiment studies and applies the temperature compensation for an LED display screen in a high - humidity environment. A high - humidity environment may cause the performance of the circuit components of the LED display screen to decline, thereby affecting the brightness and chromaticity performance. This embodiment performs temperature compensation through the following steps: Environmental simulation and gray - scale acquisition: Simulate high - humidity conditions in the laboratory environment, and respectively acquire the gray - scale values of the LED display screen in the cold state and the hot state to form the corresponding cold - state gray - scale matrix and hot - state gray - scale matrix.

[0130] Compensation coefficient calculation and optimization: Based on the acquired gray - scale matrices, calculate the ratio matrix, and optimize the compensation coefficient matrix through Gaussian filtering and Makima interpolation methods.

[0131] Application and effect evaluation of compensation coefficient: Apply the optimized compensation coefficient matrix to the LED display screen and operate it under high humidity conditions to evaluate the improvement effect of brightness and chromaticity uniformity.

[0132] The experimental results show that the brightness and chromaticity uniformity of the LED display screen after temperature compensation treatment are significantly improved under high humidity environment, and the display effect is stable, further verifying the applicability and effectiveness of the present invention under complex environmental conditions.

[0133] The technical solutions provided by the present invention are further described in detail through several specific embodiments above to highlight the advantages and beneficial effects of the technical solutions provided by the present invention. However, the several specific embodiments described above are not used as limitations to the present invention. Any reasonable modifications and improvements to the present invention, combinations of embodiments, equivalent replacements, etc. within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A temperature compensation method for an LED display screen, characterized in that: include: The step of collecting cold state and hot state grayscale matrices of the LED screen to form a cold state grayscale matrix C and a hot state grayscale matrix H; Based on the cold grayscale matrix C and the hot grayscale matrix H, the LED lamp bead brightness attenuation ratio is calculated pixel by pixel to generate a ratio matrix R; The step of constructing and iteratively calculating a correction matrix K so that the hot state brightness matrix H approaches the cold state brightness matrix C after being processed by K; Based on the ratio matrix R, a temperature compensation surface of the LED screen is constructed and normalized; The steps of calculating the brightness ratio of the edge of the LED screen and performing mirror filling on the pixels that cross the edge boundary; The steps of constructing the RGB chromaticity correction matrix of the CIE-XYZ color space and combining it with the temperature compensation matrix to dynamically adjust the LED drive current and compensate for the effect of temperature changes on the LED color performance.

2. A temperature compensation method for an LED display screen according to claim 1, characterized in that: The brightness data of the LED display screen in a stable ambient temperature state and in a thermal equilibrium state after long-term lighting are obtained respectively to form a cold grayscale matrix C and a hot grayscale matrix H.

3. A temperature compensation method for an LED display screen according to claim 1, characterized in that: The Nesterov accelerated gradient descent optimization algorithm is used to construct and iteratively calculate the correction matrix K.

4. A temperature compensation method for an LED display screen according to claim 1, characterized in that: The temperature compensation surface of LED screen is constructed using Makima interpolation method.

5. A temperature compensation method for an LED display screen according to claim 4, characterized in that: The specific method for constructing the temperature compensation surface of the LED screen is as follows: the screen is divided into Modules, extract the 9×9 pixel mean of the same peak position of each module as the feature point; Construct a piecewise cubic polynomial along the x / y axis that satisfies: , a is a constant, and the slope is calculated using the adjacent four-point weighted method to generate a smooth compensation surface with a resolution of 1440×810.

6. A temperature compensation device for an LED display screen, characterized in that: include: A module for collecting cold and hot grayscale matrices of the LED screen to form a cold grayscale matrix C and a hot grayscale matrix H; A module for calculating the LED lamp bead brightness attenuation ratio pixel by pixel based on the cold grayscale matrix C and the hot grayscale matrix H to generate a ratio matrix R; A module for constructing and iteratively calculating a correction matrix K so that the hot state brightness matrix H approaches the cold state brightness matrix C after being processed by K; Based on the ratio matrix R, a module is built to normalize the temperature compensation surface of the LED screen. A module that calculates the brightness ratio of the LED screen edge and performs mirror filling on the pixels that cross the edge boundary; A module that constructs the RGB chromaticity correction matrix of the CIE-XYZ color space and combines it with the temperature compensation matrix to dynamically adjust the LED drive current and compensate for the impact of temperature changes on LED color performance.

7. A computer storage medium for storing a computer program, characterized in that: When the computer program is read by a computer, the computer executes the method of claim 1 .

8. A computer, comprising a processor and a storage medium, characterized in that: When the processor reads the computer program stored in the storage medium, the computer executes the method of claim 1 .

9. A computer program product, being a computer program, characterized in that When the computer program is executed, the method of claim 1 is implemented.

Images