Intelligent dot matrix electronic display screen control system

Abstract

The invention discloses an intelligent dot matrix electronic display screen control system, which belongs to the technical field of electronic display screen control, and comprises a data acquisition module, a spectrum dynamic compensation module, a fault prediction module, a pre-compensation control module, an FPGA (Field Programmable Gate Array) driving module and a feedback iteration module. Data interaction of all the modules forms a complete control closed loop, and the high-quality display effect of the display screen in a complex environment is achieved. According to the method, the problem of color shift of a traditional display screen when the environment changes is effectively solved by dynamically adjusting the spectral compensation coefficient, pixel aging characteristics are learned, fault risks are predicted and quantified in advance, fault pixel defects are masked by adjusting the voltage and lighting time sequence of adjacent pixels and utilizing the persistence of vision effect of human eyes, and the display quality of the display screen is improved. The reliability and stability of the display screen are improved, and powerful support is provided for wide application of the intelligent dot matrix electronic display screen in multiple scenes.

Description

Technical Field

[0001] This invention relates to the field of electronic display screen control technology, and in particular to an intelligent dot matrix electronic display screen control system. Background Technology

[0002] The intelligent dot matrix electronic display screen is composed of a large number of small light-emitting units (pixels) arranged in a matrix. Each pixel can independently control its brightness and color, thereby combining various complex images and text information. Combined with the intelligent control system, it can automatically adjust the display parameters according to environmental changes and display needs, providing a better and more reliable display effect.

[0003] As the core carrier of modern information display, intelligent dot matrix electronic display screens have been widely used in outdoor advertising, medical imaging, traffic control and other fields. Their technological development focuses on two core directions: color consistency and reliability.

[0004] In recent years, the introduction of quantum dot light-emitting technology has brought significant improvements in color gamut coverage and brightness performance to intelligent dot matrix electronic displays. However, factors such as frequent fluctuations in ambient light, pixel temperature drift, and unstable driving voltage still pose a serious challenge to the stability of display effects. Especially in scenarios with high color requirements, any slight color shift may affect the accurate transmission of information and visual experience. At the same time, with the continuous expansion of display screen size and the continuous increase in pixel density, problems such as dark spots and uneven brightness caused by pixel aging are becoming increasingly prominent. The traditional fault handling mode based on regular maintenance is no longer able to meet the needs of high reliability scenarios due to its lack of real-time and predictability, which increases the operation and maintenance costs and the risk of display interruption. Summary of the Invention

[0005] To address the issues raised in the background section, such as color shifts due to environmental changes, lack of real-time performance and predictability, leading to low stability and reliability of the display screen during use, we propose an intelligent dot matrix electronic display screen control system.

[0006] The main technical solution is: an intelligent dot matrix electronic display screen control system, including a data acquisition module, a spectral dynamic compensation module, a fault prediction module, a pre-compensation control module, an FPGA driver module, and a feedback iteration module;

[0007] The data acquisition module is used to collect real-time data on ambient light spectral shift, pixel temperature, driving voltage, and current, providing data input for each module;

[0008] The spectral dynamic compensation module normalizes the ambient light spectral offset, driving voltage, pixel temperature and reference temperature, eliminates the dimensional differences between parameters, calculates the spectral compensation coefficient, and adjusts the driving parameters of the quantum dot light-emitting unit.

[0009] The fault prediction module includes a historical data storage unit, an LSTM model unit, and a fault probability output unit. The historical data storage unit stores pixel time-series running data to form a training dataset. The LSTM model unit learns pixel aging features through a long short-term memory network. The fault probability output unit calculates the fault probability value to provide a decision basis for the pre-compensation control module.

[0010] The pre-compensation control module includes a fault threshold judgment unit, a voltage adjustment calculation unit, and a timing adjustment calculation unit. The fault threshold judgment unit compares the fault probability value with a preset threshold to trigger the compensation logic. The voltage adjustment calculation unit calculates the voltage adjustment amount of adjacent pixels based on the fault probability value and the reference driving parameters. The timing adjustment calculation unit calculates the lighting timing adjustment amount by combining the voltage adjustment amount and the timing adjustment ratio coefficient.

[0011] The FPGA driver module is used to execute the hardware output of the spectral dynamic compensation module and the pre-compensation control module, and to adjust the quantum dot driving voltage and lighting timing.

[0012] The feedback iteration module iteratively optimizes the spectral compensation coefficient and fault prediction model based on the effect data after hardware execution, thereby achieving system closed loop.

[0013] Preferably, the data acquisition module includes a spectral sensing unit, a temperature sensing unit, a voltage acquisition unit, and a current acquisition unit;

[0014] The spectral sensing unit is used to monitor the ambient light spectral shift in real time and output spectral shift data to provide data support for the dynamic spectral compensation module.

[0015] The temperature sensing unit is used to collect the real-time temperature of each pixel, covering the operating temperature range of the display screen, and providing data support for the spectral dynamic compensation module and the fault prediction module.

[0016] The voltage acquisition unit is used to acquire the quantum dot driving voltage output by the FPGA and output voltage fluctuation data to provide data support for the spectral dynamic compensation module and the fault prediction module.

[0017] The current acquisition unit is used to acquire the driving current of the pixels, reflect the changes in light intensity, and provide data support for the fault prediction module.

[0018] Preferably, the spectral dynamic compensation module includes a compensation coefficient calculation unit and a parameter calibration unit;

[0019] The compensation coefficient calculation unit is configured to calculate the spectral compensation coefficient based on the normalized ambient light spectral offset, driving voltage, pixel temperature and reference temperature parameters, combined with quantum dot voltage and wavelength sensitivity and sensor measurement error.

[0020] The parameter calibration unit is used to dynamically adjust the calculation weight of the spectral compensation coefficient based on the feedback spectral shift data, thereby improving the accuracy of color correction.

[0021] Preferably, in the fault prediction module:

[0022] Historical data storage unit is used to store driving data such as historical voltage, historical current, and historical temperature of pixels, as well as spectral compensation coefficients, to form a training dataset;

[0023] The LSTM model unit is used to analyze historical data through a long short-term memory network, learn the time-series features of pixel aging and faults, and output them to the fault probability output unit.

[0024] The fault probability output unit receives the output of the LSTM model unit, processes the historical driving data and spectral compensation coefficients using a normalization method, and then inputs them into the model for analysis and calculation. It quantifies the pixel fault risk within a preset time period and outputs the fault probability value to provide a decision basis for the pre-compensation control module.

[0025] Preferably, in the pre-compensation control module:

[0026] The fault threshold judgment unit receives the fault probability value from the fault probability output unit, compares the fault probability value with the preset threshold, and triggers the pre-compensation control logic.

[0027] The voltage adjustment calculation unit is used to obtain the normalized fault probability value, reference voltage value and driving voltage value. It calculates the pre-compensation voltage adjustment amount of adjacent pixels by multiplying the fault probability value and the reference voltage value, and then combining the calculation with the compensation coefficient of adjacent pixels for weighted adjustment.

[0028] The timing adjustment calculation unit obtains the voltage adjustment amount and the timing adjustment ratio coefficient. By calculating the ratio of the voltage adjustment amount to the reference voltage and then combining it with the timing adjustment ratio coefficient to perform reverse adjustment calculation, the lighting timing adjustment amount of adjacent pixels is obtained.

[0029] Preferably, the FPGA driver module includes a PWM signal generation unit, a voltage output unit, and a timing control unit;

[0030] The PWM signal generation unit is used to convert the pre-compensated voltage adjustment amount output by the voltage adjustment calculation unit into a digital PWM signal.

[0031] The voltage output unit is used to output the adjusted drive voltage to adjacent pixels according to the PWM signal;

[0032] The timing control unit is used to generate a new pixel lighting timing sequence based on the timing adjustment amount output by the timing adjustment calculation unit, ensuring that the lighting timing of adjacent pixels and faulty pixels is synchronized.

[0033] Preferably, the feedback iteration module includes an effect monitoring unit, a model optimization unit, and a compensation coefficient update unit;

[0034] The effect monitoring unit is used to monitor the brightness uniformity and color consistency after pre-compensation and output effect evaluation data.

[0035] The model optimization unit retrains the LSTM model based on the performance evaluation data output by the performance monitoring unit, adjusts the weight parameters for fault prediction, and feeds back the optimized model parameters to the LSTM model unit of the fault prediction module.

[0036] The compensation coefficient update unit is used to iteratively optimize the spectral compensation coefficient based on the temperature and voltage data output by the effect monitoring unit, thereby improving the real-time performance of color correction.

[0037] The feedback iteration module outputs the optimized spectral compensation coefficients to the spectral dynamic compensation module, and adjusts the pixel driving parameters through the FPGA driver module until the iteration stops.

[0038] Preferably, the conditions for stopping iteration include:

[0039] Condition 1: The error of the spectral dynamic compensation coefficient is less than the preset threshold and is continuously stable;

[0040] Condition 2: The accuracy of the LSTM model in the fault prediction module reaches the preset standard and remains consistently without significant fluctuations;

[0041] Condition 3: The brightness uniformity after pre-compensation reaches the level where there is no difference in human visual perception;

[0042] The iteration can stop if any one of the conditions is met, with condition three taking priority.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] In this invention, environmental and pixel operation data are collected in real time by multiple units of the data acquisition module. Combined with the normalization processing and parameter calibration of the spectral dynamic compensation module, color consistency control in complex environments is achieved. This module analyzes the influence of ambient light, temperature, and voltage on quantum dot luminescence and dynamically adjusts the compensation coefficient weights, solving the problem of color shift in traditional displays when the environment changes. This ensures color accuracy in different scenarios and ultimately controls the color deviation of the display within the acceptable range for the human eye, improving the applicability of the display in scenarios with high color requirements.

[0045] In this invention, a fault prediction module stores historical time-series data and trains an LSTM model. Combined with threshold judgment and voltage and timing adjustments by the pre-compensation control module, it achieves early prediction and imperceptible masking of faulty pixels. The fault prediction module learns pixel aging characteristics and outputs the fault probability, while the pre-compensation module adjusts the voltage and illumination timing of adjacent pixels, utilizing the persistence of vision to mask dark spot defects. This avoids display interruptions caused by sudden faults, ultimately enabling the display screen to maintain visual integrity in the early stages of pixel aging through pre-compensation. This reduces maintenance costs and improves the stability of the display screen in scenarios with high reliability requirements.

[0046] In this invention, the FPGA driver module executes compensation instructions and feeds back hardware data. Combined with the effect monitoring and parameter optimization of the feedback iteration module, the system achieves dynamic closed-loop iteration. The FPGA translates algorithm instructions into hardware operations, while the feedback module optimizes the LSTM model and compensation coefficient weights by monitoring indicators such as spectral deviation and brightness uniformity. This continuously improves system performance and avoids over-iteration, ultimately enabling the display screen to maintain color consistency and fault prediction accuracy during long-term operation, balancing display effect and resource consumption. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the dynamic compensation process in this invention.

[0048] Figure 2 This is a flowchart illustrating the fault prediction and pre-compensation process in this invention.

[0049] Figure 3 This is a flowchart illustrating the feedback iteration process in this invention.

[0050] Figure 4 This is a flowchart of the operation of the FPGA driver module in this invention. Detailed Implementation

[0051] The present invention will now be described in further detail with reference to the accompanying drawings and preferred embodiments.

[0052] Example 1, refer to Figures 1-4 As shown, an intelligent dot matrix electronic display control system includes a data acquisition module, a spectral dynamic compensation module, a fault prediction module, a pre-compensation control module, an FPGA driver module, and a feedback iteration module.

[0053] FPGA stands for Field Programmable Gate Array;

[0054] Specifically, the data acquisition module includes a spectral sensing unit for monitoring the ambient light spectral shift, a temperature sensing unit for acquiring the real-time temperature of each pixel, a voltage acquisition unit for recording the driving voltage fluctuations output by the FPGA, and a current acquisition unit for capturing pixel current changes. The acquired data is output to the spectral dynamic compensation unit and the fault prediction module to provide data support for subsequent operation.

[0055] After receiving the data output from the data acquisition module, the spectral dynamic compensation module analyzes the effects of ambient light, voltage, and temperature on the emission wavelength of quantum dots, calculates the spectral compensation coefficient, and dynamically adjusts the calculation weights based on the effect data from the feedback iteration module to ensure color consistency under different environments.

[0056] The fault prediction module stores the pixel time-series data (current, voltage, temperature, spectral compensation coefficient) of the past 7 days through the historical data storage unit to form a training dataset. The LSTM model unit learns pixel aging features based on the time-series data. The fault probability output unit quantizes the model output into a fault probability value and outputs it to the pre-compensation control module.

[0057] LSTM stands for Long Short-Term Memory Network;

[0058] After the fault threshold judgment unit of the pre-compensation control module triggers the compensation logic, the voltage adjustment calculation unit calculates the voltage adjustment amount of adjacent pixels, and the timing adjustment calculation unit combines the voltage adjustment amount to generate the lighting timing adjustment amount. The adjustment data is transmitted to the FPGA driver module through the PWM (pulse width modulation) interface, and the dark spot defects of the faulty pixels are covered by the persistence of vision of the human eye.

[0059] The FPGA driver module executes the hardware output of the spectral dynamic compensation module and the pre-compensation control module to adjust the quantum dot driving voltage and lighting timing. After the hardware execution, the brightness of the adjacent area of ​​the faulty pixel is improved, and there is no obvious defect in the user's vision.

[0060] The feedback iteration module collects the display effect data after pre-compensation and retrains the LSTM model. The compensation coefficient update unit optimizes the calculation weight of the spectral compensation coefficient. The optimized parameters are fed back to the spectral dynamic compensation module and the fault prediction module to realize the closed-loop iteration of the system.

[0061] In the above implementation scheme, through a closed-loop link of data acquisition, dynamic compensation, fault prediction, pre-compensation control, hardware execution, and feedback iteration, the modules work together to achieve real-time optimization of display effects and fault masking. The modules interact with each other through hardware interfaces to form a complete control closed loop, thus achieving high reliability and excellent display effects of the intelligent dot matrix electronic display screen in complex environments.

[0062] Example 2, refer to Figures 1-4 As shown, the spectral dynamic compensation module, as the core unit of the system's color correction, interacts in real time with the spectral sensing unit, temperature sensing unit, and voltage acquisition unit of the data acquisition module. It receives ambient light spectral offset, pixel temperature, and driving voltage data, and uses a maximum-minimum normalization method to map the data to the [0,1] interval to eliminate dimensional differences. Combined with preset quantum dot voltage and wavelength sensitivity coefficients and sensor error parameters, it calculates the spectral compensation coefficient to adjust the quantum dot driving parameters. The formula for calculating the spectral compensation coefficient is as follows:

[0063] ;

[0064] In the formula, G(t) is the spectral compensation coefficient, and Δλ en (t) represents the normalized ambient light spectral shift, reflecting the impact of ambient light wavelength changes on displayed color. dri (t) represents the normalized quantum dot driving voltage, is the real-time voltage value output by the FPGA, and k V The preset quantum dot voltage and wavelength sensitivity coefficient, determined experimentally based on the quantum dot material properties, reflects the degree of influence of voltage changes on wavelength. σ se The mean square error of the spectral sensor is determined by taking the standard deviation from multiple measurements. This is used to reduce the interference of hardware errors on the calculation. pi (t) represents the normalized real-time pixel temperature, T ref T is the normalized reference temperature, and T is the ideal operating temperature of the display screen. pi (t) + T ref The superposition effect factor representing the real-time temperature of the pixel and the reference temperature is used to quantify the comprehensive effect of temperature on the emission wavelength of quantum dots. The superposition of the reference temperature can more accurately reflect the temperature drift.

[0065] Furthermore, after calculating the spectral compensation coefficient, the parameter calibration unit calculates the error contribution of each parameter (ambient light spectral offset, driving voltage, pixel temperature, reference temperature) in the spectral compensation coefficient calculation formula and analyzes its correlation with color deviation.

[0066] The formula for calculating the error contribution is as follows:

[0067] ;

[0068] In the formula, W i The error contribution of parameter i, with a value ranging from 0 to 1, r i Let p be the correlation coefficient between parameter i and systematic error, used to measure the degree of linear correlation between the two, with a value ranging from -1 to 1. iLet be the error percentage of parameter i, representing the proportion of the measurement error of this parameter in the total error, with a value ranging from 0 to 1, where n is the sum of the parameters involved in the calculation, and r... j p is the correlation coefficient between parameter j and system error, ranging from -1 to 1. j The error percentage of parameter j, with a value ranging from 0 to 1;

[0069] By calculating using the above formula, the correlation between the parameters and the error and the error proportion of the parameters themselves are quantified, and the weight allocation of each parameter is realized, providing a scientific basis for the subsequent dynamic adjustment of the spectral compensation coefficient.

[0070] in, ;

[0071] In the formula, x k Let y be the parameter value of the kth sample (such as pixel temperature, ambient light spectral shift). k Let be the systematic error value corresponding to the k-th sampling, which is the deviation between the measured parameter value and the true value at the k-th sampling. The average value of the parameters. The average value of the system error is denoted by , m is the number of samples, and r takes values ​​from -1 to 1. The larger the absolute value, the stronger the linear correlation between the parameter and the error.

[0072] Furthermore, c i The measurement error of parameter i is determined by the accuracy of the hardware sensor, c z This represents the total error.

[0073] and, ;

[0074] In the above implementation scheme, the spectral sensing unit collects the ambient light spectral offset in real time (such as when indoor light changes from cool white to warm yellow, the offset increases), the temperature sensing unit collects the temperature of each pixel synchronously, the voltage acquisition unit records the driving voltage fluctuations output by the FPGA, the compensation coefficient calculation unit normalizes these parameters and substitutes them into the formula to calculate the spectral compensation coefficient, and the parameter calibration unit dynamically adjusts the weight of each parameter in the formula based on the brightness uniformity data returned by the feedback iteration module. Finally, this module controls the color deviation of the display screen under different ambient light and temperature within the range acceptable to the human eye, and greatly improves color consistency.

[0075] Example 3, referring to Figures 1-4As shown, the fault prediction module, as the core unit for ensuring system reliability, interacts with each unit in the data acquisition module in real time, receives historical current, temperature, and driving voltage data of pixels, and obtains historical spectral compensation coefficients from the spectral dynamic compensation module. The historical data storage unit continuously stores the time-series operation data of each pixel to form a training dataset. The LSTM model unit takes the normalized historical data of the past period as input and learns the time-series features of pixel aging through the long short-term memory network. The fault probability output unit quantizes the model output into a fault probability value and outputs it to the pre-compensation control module.

[0076] The formula for calculating the failure probability value is as follows:

[0077] ;

[0078] In the formula, P fb (t) represents the pixel failure probability at time t, with an output range of 0 to 1. The closer the probability is to 1, the higher the failure risk. LSTM is a Long Short-Term Memory network model used to capture long-term dependencies in time series data. pix (t—τ) represents the normalized historical pixel current, V dri (t—τ) represents the normalized historical driving voltage, T pix (t—τ) is the normalized historical pixel temperature, G(t—τ) is the normalized historical spectral compensation coefficient, where τ is the time window, such as the current data of the past ten minutes, and t—τ represents the historical data starting point τ time backward from the current time t.

[0079] Specifically, the historical data storage unit continuously stores the time-series operational data of each pixel (such as current, voltage, and temperature collected every 100ms, combined with the compensation coefficients output by the spectral dynamic compensation module to form a training dataset). The LSTM model unit takes the normalized historical data of the past 10 minutes as input to learn the time-series features of pixel aging. The fault probability output unit quantifies the model output into a fault probability value. The spectral dynamic compensation module can predict pixel faults in advance, providing accurate decision-making basis for the pre-compensation control module and avoiding display defects caused by sudden faults.

[0080] Example 4, refer to Figures 1-4 As shown, the pre-compensation control module, as the core unit of the system fault masking, receives the fault probability value output by the fault prediction module, combines it with the reference voltage and driving voltage data provided by the voltage acquisition unit of the data acquisition module, performs normalization processing, and calculates the voltage adjustment amount and lighting timing adjustment amount of adjacent pixels. The pre-compensation control module includes a fault threshold judgment unit, a voltage adjustment calculation unit, and a timing adjustment calculation unit. The three work together to transmit the calculation results to the FPGA driver module to achieve imperceptible compensation of faulty pixels.

[0081] In the above implementation scheme, the voltage adjustment calculation unit obtains the normalized fault probability value, the reference voltage value, and the driving voltage value. It calculates the pre-compensation voltage adjustment amount for adjacent pixels by multiplying the fault probability value by the reference voltage value and then combining this with the adjacent pixel compensation coefficient for weighted adjustment. The specific calculation formula is as follows:

[0082] ;

[0083] In the formula, ΔV adj (i, j, t) represents the voltage adjustment amount of the neighboring pixels of pixel (i, j) at time t. Positive values ​​indicate an increase in voltage, and negative values ​​indicate a decrease in voltage. P fb (i, j, t) represents the normalized fault probability, output by the fault prediction module, P th The normalized fault probability threshold, V, is determined based on engineering experience and historical fault data statistics. ref V is the normalized reference voltage. dri α is the normalized current driving voltage. nei The compensation coefficients for adjacent pixels are dynamically adjusted based on the location of the faulty pixel.

[0084] By quantifying the deviation between the fault probability and the reference threshold, the difference between the current driving voltage and the ideal reference voltage, and combining the dynamic weight of the pixel position, precise and differentiated voltage compensation for faulty pixels is achieved. The abstract fault risk is transformed into an executable voltage adjustment command, which can actively boost or deboost the voltage for high-risk pixels to reduce the fault probability, and optimize the allocation of compensation resources according to the pixel position. While ensuring system stability, it avoids overcompensation and effectively balances the efficiency of fault prevention and the consistency of display effect.

[0085] Furthermore, the timing adjustment calculation unit obtains the voltage adjustment amount and the timing adjustment ratio coefficient. By calculating the ratio of the voltage adjustment amount to the reference voltage, and then combining it with the timing adjustment ratio coefficient to perform a reverse adjustment calculation, the lighting timing adjustment amount of adjacent pixels is obtained. The persistence of vision effect of the human eye is used to cover up the defects of faulty pixels. The calculation formula for the lighting timing adjustment amount is as follows:

[0086] ;

[0087] In the formula, ΔT tim (i, j, t) represents the lighting timing adjustment amount, where negative values ​​represent early lighting and positive values ​​represent delayed lighting, and k tim The timing adjustment scaling factor is a preset constant, which is determined through experimental verification.

[0088] By inversely linking with the voltage adjustment (lighting up earlier when the voltage increases and lighting up later when the voltage decreases), and controlling the adjustment range with a timing ratio coefficient, a two-dimensional compensation closed loop of voltage and timing is formed. At the same time, it assists in voltage adjustment to optimize the display effect, which not only avoids abnormal brightness caused by voltage changes, but also prevents secondary faults such as ghosting and flickering caused by overcompensation by quantifying the timing adjustment range. Ultimately, it improves the system's adaptability to complex display scenarios and ensures the uniformity and synchronization of brightness across the entire screen.

[0089] The fault threshold judgment unit compares the fault probability value with the threshold and triggers the pre-compensation logic. The voltage adjustment calculation unit takes the normalized fault probability, the reference voltage and the current driving voltage as inputs to calculate the voltage adjustment amount of the adjacent pixels. The timing adjustment calculation unit calculates the lighting timing adjustment amount based on the voltage adjustment amount. After the FPGA driving module performs the adjustment, the brightness of the adjacent pixels will be slightly improved and the lighting timing will be synchronized with the faulty pixels. By utilizing the persistence of vision effect of the human eye, the dark spot defects of the faulty pixels are covered up.

[0090] Example 5, refer to Figures 1-4 As shown, the FPGA driver module includes a PWM signal generation unit, a voltage output unit, and a timing control unit. The FPGA driver module obtains the spectral compensation coefficient G(t) from the spectral dynamic compensation module and the voltage adjustment amount ΔV from the pre-compensation module. adj (i, j, t) and the lighting timing adjustment ΔT tim (i, j, t);

[0091] The PWM signal generation unit will adjust the voltage ΔV adj (i, j, t) is converted into a digital PWM signal. The voltage adjustment is achieved by adjusting the duty cycle. The voltage output unit outputs the adjusted drive voltage to adjacent pixels according to the PWM signal. The timing control unit adjusts the lighting timing ΔT according to the signal. tim (i, j, t) generates a new pixel illumination timing sequence;

[0092] Furthermore, the driving voltage and lighting timing output by the FPGA driver module are monitored in real time by the voltage acquisition unit and current acquisition unit in the data acquisition module, providing current fluctuation data for the fault prediction module.

[0093] Meanwhile, the voltage adjustment performed by the FPGA changes the current consumption of the pixel, and the historical current data is stored in the historical data storage unit for training the LSTM model.

[0094] The feedback iteration module continuously optimizes system parameters by monitoring the effect data after FPGA execution, forming a closed-loop control.

[0095] Specifically, the effect monitoring unit collects effect data, calculates the deviation between the spectral offset of each pixel and the target spectrum to reflect color consistency, statistically analyzes the brightness difference between adjacent pixels to reflect brightness uniformity, and evaluates the visual repair success rate of faulty pixels through user feedback or image recognition technology.

[0096] The formulas for calculating the spectral shift and the target spectral deviation are as follows:

[0097] ;

[0098] In the formula, Δλ en (t) represents the normalized ambient light spectral shift, Δλ max The maximum deviation is Δλ en (t)×Δλ max G(t) is the original offset, and G(t) is the spectral compensation coefficient.

[0099] The formula for calculating the brightness difference between adjacent pixels is as follows:

[0100] ;

[0101] In the formula, ΔL a (i, j, t) represents the average brightness difference between pixel (i, j) and its neighboring pixels at time t, n is the number of effective pixels, L(i, j, t) represents the actual brightness value of pixel (i, j) at time t, which is collected by the sensor, and L(i+k, j, t) represents the brightness value of the neighboring pixels of pixel (i, j) at time t, where k represents the offset in the adjacent direction.

[0102] The formula for calculating the success rate of visual restoration is as follows:

[0103] ;

[0104] In the formula, N su (t) represents the number of faulty pixels that were successfully repaired visually at time t, N to (t) represents the total number of faulty pixels detected at time t;

[0105] If the fault prediction rate is lower than the preset threshold, the model optimization unit will retrain the LSTM model using the newly collected current, voltage, and temperature data, and feed the optimized model parameters back to the LSTM model unit of the fault prediction module to replace the old model.

[0106] If the color consistency deviation is large, the compensation coefficient update unit will update the compensation coefficient based on the calculated Δλ. err (t) Adjusting the voltage and wavelength sensitivity coefficient k of the spectral compensation coefficient V The updated coefficients are then fed back to the compensation coefficient calculation unit of the spectral dynamic compensation module;

[0107] Furthermore, to avoid excessive iteration leading to wasted system resources, the feedback iteration module sets the following stop iteration conditions (any one of which will trigger the stop):

[0108] Condition 1: The error of the spectral dynamic compensation coefficient is less than the preset threshold and is continuous and stable, that is, the color correction accuracy meets the standard.

[0109] Condition 2: The fault prediction model runs stably, the accuracy of the LSTM model reaches the preset standard, and there are no significant fluctuations.

[0110] Condition 3: The brightness uniformity after pre-compensation reaches the level of no difference in human visual perception;

[0111] Among them, the preset threshold and preset standard are determined by experimental data, and the specific values ​​can be adjusted according to the actual application scenario;

[0112] If, after stopping the iteration, the ambient light spectral offset exceeds the threshold, the pixel temperature exceeds the design threshold, or the fault prediction accuracy falls below the threshold, the system will automatically resume the iteration.

[0113] It should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should also be within the scope of protection of this invention.

Claims

1. A control system for an intelligent dot matrix electronic display screen, characterized in that: It includes a data acquisition module, a dynamic spectral compensation module, a fault prediction module, a pre-compensation control module, an FPGA driver module, and a feedback iteration module; The data acquisition module is used to collect real-time data on ambient light spectral shift, pixel temperature, driving voltage, and current, providing data input for each module; The spectral dynamic compensation module normalizes the ambient light spectral offset, driving voltage, pixel temperature and reference temperature, eliminates the dimensional differences between parameters, calculates the spectral compensation coefficient, and adjusts the driving parameters of the quantum dot light-emitting unit. The fault prediction module includes a historical data storage unit, an LSTM model unit, and a fault probability output unit. The historical data storage unit stores pixel time-series running data to form a training dataset. The LSTM model unit learns pixel aging features through a long short-term memory network. The fault probability output unit calculates the fault probability value to provide a decision basis for the pre-compensation control module. The pre-compensation control module includes a fault threshold judgment unit, a voltage adjustment calculation unit, and a timing adjustment calculation unit. The fault threshold judgment unit compares the fault probability value with a preset threshold to trigger compensation logic. The voltage adjustment calculation unit calculates the voltage adjustment amount of adjacent pixels based on the fault probability value and the reference driving parameters. The timing adjustment calculation unit calculates the lighting timing adjustment amount by combining the voltage adjustment amount and the timing adjustment ratio coefficient. The FPGA driver module is used to execute the hardware output of the spectral dynamic compensation module and the pre-compensation control module, and adjust the quantum dot driving voltage and lighting timing. The feedback iteration module iteratively optimizes the spectral compensation coefficient and fault prediction model based on the effect data after hardware execution, thereby achieving system closed loop.

2. The intelligent dot matrix electronic display screen control system according to claim 1, characterized in that: The data acquisition module includes a spectral sensing unit, a temperature sensing unit, a voltage acquisition unit, and a current acquisition unit. The spectral sensing unit is used to monitor the ambient light spectral shift in real time and output spectral shift data to provide data support for the dynamic spectral compensation module. The temperature sensing unit is used to collect the real-time temperature of each pixel, covering the operating temperature range of the display screen, and providing data support for the spectral dynamic compensation module and the fault prediction module. The voltage acquisition unit is used to acquire the quantum dot driving voltage output by the FPGA and output voltage fluctuation data to provide data support for the spectral dynamic compensation module and the fault prediction module. The current acquisition unit is used to acquire the driving current of the pixels, reflect the changes in light intensity, and provide data support for the fault prediction module.

3. The intelligent dot matrix electronic display screen control system according to claim 2, characterized in that: The spectral dynamic compensation module includes a compensation coefficient calculation unit and a parameter calibration unit; The compensation coefficient calculation unit is configured to calculate the spectral compensation coefficient based on the normalized ambient light spectral offset, driving voltage, pixel temperature and reference temperature parameters, combined with quantum dot voltage and wavelength sensitivity and sensor measurement error. The parameter calibration unit is used to dynamically adjust the calculation weight of the spectral compensation coefficient based on the feedback spectral shift data, thereby improving the accuracy of color correction.

4. The intelligent dot matrix electronic display screen control system according to claim 1, characterized in that: In the fault prediction module: Historical data storage unit is used to store driving data such as historical voltage, historical current, and historical temperature of pixels, as well as spectral compensation coefficients, to form a training dataset; The LSTM model unit is used to analyze historical data through a long short-term memory network, learn the time-series features of pixel aging and faults, and output them to the fault probability output unit. The fault probability output unit receives the output of the LSTM model unit, processes the historical driving data and spectral compensation coefficients using a normalization method, and then inputs them into the model for analysis and calculation. It quantifies the pixel fault risk within a preset time period and outputs the fault probability value to provide a decision basis for the pre-compensation control module.

5. The intelligent dot matrix electronic display screen control system according to claim 4, characterized in that: In the pre-compensation control module: The fault threshold judgment unit receives the fault probability value from the fault probability output unit, compares the fault probability value with a preset threshold, and triggers the pre-compensation control logic. The voltage adjustment calculation unit is used to obtain the normalized fault probability value, reference voltage value and driving voltage value. It calculates the pre-compensation voltage adjustment amount of adjacent pixels by multiplying the fault probability value and the reference voltage value, and then combining the calculation with the compensation coefficient of adjacent pixels for weighted adjustment. The timing adjustment calculation unit obtains the voltage adjustment amount and the timing adjustment ratio coefficient. By calculating the ratio of the voltage adjustment amount to the reference voltage and then combining it with the timing adjustment ratio coefficient to perform reverse adjustment calculation, the lighting timing adjustment amount of adjacent pixels is obtained.

6. The intelligent dot matrix electronic display screen control system according to claim 5, characterized in that: The FPGA driver module includes a PWM signal generation unit, a voltage output unit, and a timing control unit; The PWM signal generation unit is used to convert the pre-compensated voltage adjustment amount output by the voltage adjustment calculation unit into a digital PWM signal. The voltage output unit is used to output the adjusted drive voltage to adjacent pixels according to the PWM signal; The timing control unit is used to generate a new pixel lighting timing sequence based on the timing adjustment amount output by the timing adjustment calculation unit, ensuring that the lighting timing of adjacent pixels and faulty pixels is synchronized.

7. The intelligent dot matrix electronic display screen control system according to claim 6, characterized in that: The feedback iteration module includes an effect monitoring unit, a model optimization unit, and a compensation coefficient update unit; The effect monitoring unit is used to monitor the brightness uniformity and color consistency after pre-compensation and output effect evaluation data. The model optimization unit retrains the LSTM model based on the performance evaluation data output by the performance monitoring unit, adjusts the weight parameters for fault prediction, and feeds back the optimized model parameters to the LSTM model unit of the fault prediction module. The compensation coefficient update unit is used to iteratively optimize the spectral compensation coefficient based on the temperature and voltage data output by the effect monitoring unit, thereby improving the real-time performance of color correction. The feedback iteration module outputs the optimized spectral compensation coefficients to the spectral dynamic compensation module, and adjusts the pixel driving parameters through the FPGA driving module until the iteration stop condition is met.

8. The intelligent dot matrix electronic display screen control system according to claim 7, characterized in that: The conditions for stopping the iteration include: Condition 1: The error of the spectral dynamic compensation coefficient is less than the preset threshold and is continuously stable; Condition 2: The accuracy of the LSTM model in the fault prediction module reaches the preset standard and remains consistently without significant fluctuations; Condition 3: The brightness uniformity after pre-compensation reaches the level where there is no difference in human visual perception; The iteration can stop if any one of the conditions is met, with condition three taking priority.

Images