Method and system for intelligently and dynamically adjusting brightness of LED backlight source

Abstract

The invention relates to the technical field of LED light source adjustment, and discloses an intelligent dynamic adjustment method and system for the brightness of an LED backlight source. The method comprises the following steps: acquiring an image frame sequence and real-time brightness; a brightness distribution histogram and inter-frame brightness variation are extracted based on the image frame sequence, image content change feature extraction is carried out, and an image monitoring result is generated; calculating target brightness and adjustment parameters of backlight partitions according to a monitoring result, performing deviation analysis and frequency amplitude weight matching in combination with real-time brightness, and generating partition adjustment demand data; smoothing the demand data to obtain a partition brightness curve and a smooth real-time index; calculating an intermediate compensation coefficient by correlating the fusion curve and the index to obtain a compensation coefficient set and a curve driving basis; judging the matching degree of the compensation demand and the amplitude weight to obtain a matching evaluation result; and finally, adjusting the brightness of the partitions, compensating inter-frame brightness variation, and determining a display output state. According to the method, the brightness adjustment precision and response speed under the dynamic content are improved.

Description

Technical Field

[0001] This invention relates to the field of LED light source adjustment technology, and in particular to a method and system for intelligent dynamic adjustment of LED backlight brightness. Background Technology

[0002] Display technology plays a central role in modern electronic devices, directly determining users' perceptions of image quality, energy consumption, and visual comfort. Backlight brightness adjustment, a crucial element, requires intelligent adjustment based on dynamic changes in image content to enhance image detail and reduce energy consumption. As the core component of backlight technology, the brightness adjustment mechanism of LEDs needs to be closely integrated with image processing algorithms. This involves dynamically sensing content characteristics to adjust the brightness parameters of each zone, thereby maintaining balance in scenarios with large or frequent changes.

[0003] In one existing technology, the system first captures a complete frame of the currently displayed image at fixed time intervals. Next, it calculates the average brightness value of all pixels in that frame. Then, it compares this average brightness value with a preset brightness curve, which defines the target backlight brightness level corresponding to different average brightness levels. Finally, the system applies the calculated target brightness level as a unified instruction to the backlight of the entire screen, adjusting it to the new brightness. The entire process has a fixed calculation and response cycle. Traditional technology relies on the average brightness of the entire frame for global adjustment, which cannot respond to differences in the internal partitions of the image, and the fixed-cycle adjustment mechanism leads to response delays.

[0004] Therefore, existing technologies cannot improve the brightness adjustment accuracy and response speed of display devices under dynamic content. Summary of the Invention

[0005] This invention provides a method and system for intelligent dynamic adjustment of LED backlight brightness, so as to improve the brightness adjustment accuracy and response speed of display devices under dynamic content.

[0006] Firstly, in order to solve the above-mentioned technical problems, the present invention provides an intelligent dynamic adjustment of LED backlight brightness, comprising: Acquire image frame sequences and real-time brightness from the display device; Based on the image frame sequence, a brightness distribution histogram is extracted and inter-frame brightness variation is calculated to extract image content change features and generate image monitoring results. Based on the image monitoring results, the target brightness and zoning adjustment parameters of the backlight zone are calculated, and the deviation is calculated and the frequency amplitude weight matching degree is analyzed in combination with the real-time brightness to generate zoning adjustment requirement data. Based on the frequency amplitude weight matching degree and the partition adjustment requirement data, smoothing is performed to obtain the partition brightness curve and smoothing real-time index. Based on the correlation and fusion of the partition brightness curve and the smooth real-time index, the intermediate compensation coefficient for partition adjustment is calculated to obtain the compensation coefficient set and the curve driving basis. Based on the compensation coefficient set and the curve driving basis, combined with the partition brightness curve and the smoothing real-time index, the matching degree between the compensation demand and the weighted service is determined, and the matching evaluation result is obtained. Based on the matching evaluation results, adjust the brightness level of the partition and compensate for inter-frame brightness variations to determine the display output status and the range of changes in stable output content.

[0007] In one optional implementation, the step of extracting a brightness distribution histogram and calculating inter-frame brightness variations based on the image frame sequence, extracting image content change features, and generating image monitoring results includes: Based on the image frame sequence, a weighted average algorithm is used to calculate the brightness value of each pixel to obtain the brightness component data of each frame image; Based on the brightness component data, the pixel number distribution is statistically analyzed using the histogram statistical method to obtain the brightness distribution histogram for each frame. Based on the brightness distribution histogram, the brightness similarity of the brightness distribution histograms of two adjacent frames is calculated using the histogram intersection algorithm. When the brightness similarity is lower than a preset difference threshold, it is determined that there is a significant change between the two corresponding frames, and inter-frame brightness variation data is obtained. Based on the inter-frame brightness variation data, a pre-trained convolutional neural network model is used to extract image content change features and generate image monitoring results that include the magnitude of content change and the degree of inter-frame brightness variation.

[0008] In one optional implementation, the step of calculating the target brightness and zoning adjustment parameters of the backlight zone based on the image monitoring results, and combining the real-time brightness to perform deviation calculation and frequency amplitude weight matching degree analysis, and generating zoning adjustment requirement data, includes: When the image monitoring result exceeds the preset inspection threshold, the number of variations of the inter-frame brightness variation data within a preset unit time is counted using the sliding window method to obtain the number of content changes. Based on the brightness component data corresponding to the number of content changes, the frequency domain feature analysis of the brightness change signal is performed using the Fast Fourier Transform algorithm to extract the main frequency component and obtain the content change frequency. Based on the number of content changes and the frequency of content changes, a weighted average algorithm is used to calculate the weight of the content change magnitude. Based on the content change frequency, the content change magnitude weight, and the brightness component data, a weighted average algorithm is used to perform a fusion calculation to obtain the target brightness value of the backlight zone. Based on the target brightness value, a PID algorithm is used to dynamically calculate the adjustment parameters and generate brightness adjustment parameters, resulting in partitioned adjustment parameters that include the weight of content change amplitude, brightness component data, brightness distribution histogram, target brightness and brightness adjustment parameters. The deviation between the target brightness and the real-time brightness is calculated to obtain the brightness deviation; Based on the brightness deviation and the content change magnitude weight, a weighted average algorithm is used to perform a fusion calculation to obtain the frequency amplitude weight matching degree. When the frequency amplitude weight matching degree is lower than the preset weight matching threshold, it is determined that there is an adjustment need, and partition adjustment need data containing partition adjustment need identifier and partition adjustment parameters is generated.

[0009] In one optional implementation, the smoothing process based on the partition adjustment demand data to obtain the partition brightness curve and real-time smoothing index includes: Based on the partition adjustment requirement data, a Gaussian filtering algorithm is used to suppress noise and smooth the brightness component data to generate a smooth brightness sequence. The statistical features of the brightness distribution histogram are extracted, and the statistical features of the brightness distribution histogram are fused into the smooth brightness sequence using a linear interpolation algorithm to obtain the fused adjustment sequence. Based on the fusion adjustment sequence, the least squares method is used to perform curve fitting to construct the partition brightness curve; Based on the partition brightness curve, a smoothness index is analyzed using a pre-trained convolutional neural network model to generate a smoothness real-time index.

[0010] In one optional implementation, the step of correlating and fusing the partition brightness curve and the smoothing real-time index to calculate the intermediate compensation coefficients for partition adjustment, thereby obtaining a set of compensation coefficients and a curve-driving basis, includes: Based on the partition brightness curve and the smoothing real-time index, combined with the content change magnitude weight and the brightness adjustment parameter, the content change magnitude weight and the brightness adjustment parameter are fused using a weighted average algorithm to obtain the adjustment fusion parameter; Based on the adjustment and fusion parameters, the proportional relationship between the brightness adjustment parameters and the content change magnitude weight is analyzed by linear regression algorithm, and the intermediate compensation coefficient corresponding to each partition is calculated. Based on the intermediate compensation coefficient and the partition brightness curve, a pre-trained convolutional neural network model is used to extract curve-driven features and generate a set of curve-driven criteria and compensation coefficients to guide backlight adjustment.

[0011] In one optional implementation, the step of determining the matching degree between compensation requirements and weighted services based on the compensation coefficient set and the curve-driven criteria, combined with the partition brightness curve and the smoothing real-time index, to obtain a matching evaluation result includes: Based on the aforementioned smoothing real-time index, a smoothing process is performed using a mean filtering algorithm to obtain a smoothed change sequence; Based on the smoothed change sequence, the mutation response features are extracted using the gradient descent algorithm to obtain the mutation response set; Based on the set of variant responses and the set of compensation coefficients, a weighted average algorithm is used to perform a fusion calculation to obtain the fusion weight parameters; Based on the fusion weight parameters and the curve-driven criteria, a convolutional neural network algorithm is used to analyze the business matching degree and compensation requirements to obtain the matching evaluation results.

[0012] In one optional implementation, adjusting the brightness level of the partition and compensating for inter-frame brightness variations based on the matching evaluation result, and determining the display output state and the magnitude of changes in stable output content, includes: Based on the matching evaluation results and the inter-frame brightness variation data, a backlight adjustment instruction set is generated through the Drools rule engine; According to the backlight adjustment instruction set, pulse width modulation is used to drive the backlight module to adjust the brightness level of the partition, and exponential smoothing algorithm is used to process and compensate for the brightness variation between frames, and finally display the output status and the range of changes in stable output content.

[0013] Secondly, the present invention provides an intelligent dynamic adjustment system for LED backlight brightness, comprising: The data acquisition module is used to acquire image frame sequences and real-time brightness from the display device; The image monitoring module is used to extract the brightness distribution histogram and calculate the inter-frame brightness variation based on the image frame sequence, extract image content change features, and generate image monitoring results; The adjustment demand analysis module is used to calculate the target brightness and adjustment parameters of the backlight zone based on the image monitoring results, and to perform deviation calculation and frequency amplitude weight matching degree analysis in combination with the real-time brightness to generate zone adjustment demand data. The smoothing index analysis module is used to perform smoothing processing based on the frequency amplitude weight matching degree and the partition adjustment demand data to obtain the partition brightness curve and the real-time smoothing index. The curve driving module is used to correlate and fuse the partition brightness curve and the smooth real-time index to calculate the intermediate compensation coefficient for partition adjustment, and obtain the compensation coefficient set and curve driving basis. The matching evaluation module is used to determine the matching degree between compensation requirements and weighted services based on the compensation coefficient set and the curve driving basis, combined with the partition brightness curve and the smoothing real-time index, and obtain the matching evaluation result. The output module is used to adjust the brightness level of the partition and compensate for inter-frame brightness variations based on the matching evaluation results, and to determine the display output status and the range of changes in stable output content.

[0014] Thirdly, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the LED backlight brightness intelligent dynamic adjustment method described in any one of the above.

[0015] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the LED backlight brightness intelligent dynamic adjustment method described in any one of the above.

[0016] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention collects the image frame sequence and real-time brightness of the display device, extracts the image content change features based on the brightness distribution histogram and inter-frame brightness variation, generates image monitoring results, realizes real-time perception and quantitative analysis of dynamic content changes, provides a precise data basis for backlight adjustment, and improves the response speed and adaptability of brightness adjustment.

[0017] (2) The present invention calculates the target brightness and partition adjustment parameters of the backlight partition based on the image monitoring results, and performs deviation calculation and frequency amplitude weight matching in combination with real-time brightness to generate partition adjustment requirement data, thereby realizing fine brightness control of different areas of the screen, overcoming the defect that traditional global adjustment cannot respond to partition differences, and avoiding the problems of detail loss and uneven brightness.

[0018] (3) The present invention processes the partition adjustment requirement data through smoothing methods such as Gaussian filtering and curve fitting to obtain the partition brightness curve and smoothing real-time index, and combines the intermediate compensation coefficient to perform correlation fusion, suppress brightness fluctuation and inter-frame variation, reduce screen flicker and jitter, and improve display stability and visual comfort.

[0019] (4) Based on the compensation coefficient set and curve-driven basis, the present invention judges the business matching degree between compensation requirements and amplitude weight through the matching evaluation module, dynamically adjusts the partition brightness and compensates for inter-frame brightness variation, realizes the high synchronization between brightness adjustment and image content change, and optimizes the display accuracy, energy efficiency and overall user experience under dynamic content. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the LED backlight brightness intelligent dynamic adjustment method provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of the LED backlight brightness intelligent dynamic adjustment system provided in the second embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Reference Figure 1 The first embodiment of the present invention provides a method for intelligent dynamic adjustment of LED backlight brightness, including the following steps: S11, acquires the image frame sequence and real-time brightness of the display device; S12, based on the image frame sequence, extract the brightness distribution histogram and calculate the inter-frame brightness variation, extract image content change features, and generate image monitoring results; S13, Based on the image monitoring results, calculate the target brightness and zone adjustment parameters of the backlight zone, and combine the real-time brightness to perform deviation calculation and analyze the frequency amplitude weight matching degree to generate zone adjustment requirement data; S14, Smoothing is performed based on the frequency amplitude weight matching degree and the partition adjustment demand data to obtain the partition brightness curve and smoothing real-time index. S15, Based on the correlation and fusion of the partition brightness curve and the smooth real-time index, calculate the intermediate compensation coefficient for partition adjustment to obtain the compensation coefficient set and curve driving basis; S16. Based on the compensation coefficient set and the curve driving basis, combined with the partition brightness curve and the smoothing real-time index, determine the matching degree between the compensation demand and the weighted service, and obtain the matching evaluation result. S17, adjust the brightness level of the partition and compensate for inter-frame brightness variation based on the matching evaluation results, and determine the display output status and the range of changes in stable output content.

[0023] In step S11, the image frame sequence and real-time brightness of the display device are acquired.

[0024] Specifically, image frame sequence acquisition involves calling the display device's graphics processing interface to read continuous image frame data from the frame buffer at preset sampling intervals. Each image frame consists of a pixel matrix, containing the RGB color values ​​of each pixel. For example, 60 frames are acquired per second to ensure real-time data accuracy. Real-time brightness acquisition is achieved by querying the current value of the brightness register in the backlight driver circuit or by directly measuring the screen's brightness level using an ambient light sensor integrated into the display device. During acquisition, the image frame sequence is stored as a data stream in a time-series manner, while the real-time brightness is recorded as a scalar value. This step provides raw input data for subsequent image content change analysis and brightness adjustment, ensuring the system can dynamically adjust based on the actual display state, and is the foundation for intelligent brightness control.

[0025] In step S12, based on the image frame sequence, a brightness distribution histogram is extracted and inter-frame brightness variation is calculated to extract image content change features and generate image monitoring results.

[0026] In one specific implementation, the step of extracting a brightness distribution histogram and calculating inter-frame brightness variations based on the image frame sequence, extracting image content change features, and generating image monitoring results includes: Based on the image frame sequence, a weighted average algorithm is used to calculate the brightness value of each pixel to obtain the brightness component data of each frame image; Based on the brightness component data, the pixel number distribution is statistically analyzed using the histogram statistical method to obtain the brightness distribution histogram for each frame. Based on the brightness distribution histogram, the brightness similarity of the brightness distribution histograms of two adjacent frames is calculated using the histogram intersection algorithm. When the brightness similarity is lower than a preset difference threshold, it is determined that there is a significant change between the two corresponding frames, and inter-frame brightness variation data is obtained. Based on the inter-frame brightness variation data, a pre-trained convolutional neural network model is used to extract image content change features and generate image monitoring results that include the magnitude of content change and the degree of inter-frame brightness variation.

[0027] Specifically, firstly, each frame in the acquired image frame sequence is processed. The image frame data originates from the RGB pixel matrix obtained in step S11. A weighted average algorithm is used to calculate the brightness value of each pixel. This algorithm assigns specific weight coefficients to the red, green, and blue components of each pixel (e.g., red component weight is 0.299, green component weight is 0.587, and blue component weight is 0.114). The three weighted component values ​​are then summed to obtain the brightness value of the pixel. After traversing all pixels in a frame, the brightness component data of that frame, composed of the brightness values ​​of all pixels, is obtained.

[0028] Next, based on the obtained luminance component data of each frame, histogram statistics are used for analysis. The entire possible range of luminance values ​​(e.g., 0 to 255) is divided into several consecutive intervals, and the number of pixels in the frame whose luminance values ​​fall into each interval is counted. By statistically analyzing the pixel distribution across all intervals, a luminance distribution histogram for the frame is formed, which reflects the pixel distribution at different luminance levels in the image.

[0029] Then, based on the brightness distribution histograms of consecutive frames, the inter-frame brightness variation is calculated. A histogram intersection algorithm is applied, specifically calculating the sum of the minimum statistical values ​​of the brightness distribution histograms of two adjacent frames (e.g., frame k and frame (k-1)) across all brightness intervals, and then dividing by the sum of the statistical values ​​across all intervals of one of the histograms. The result is used as the brightness similarity between the two frames. The preset difference threshold is determined by statistically analyzing the brightness similarity distribution of adjacent frames during normal playback on a large number of typical video sequences (such as videos containing fast scene transitions and gradient effects), and selecting a quantile (e.g., the fifth percentile) that ensures significant changes can be identified. It is typically set to 0.85 and can be adjusted between 0.8 and 0.95 depending on the application scenario. When the calculated brightness similarity is lower than this preset difference threshold, a significant brightness change is determined between the two frames. This event is recorded, and inter-frame brightness variation data is generated, which identifies the frame pairs where significant changes have occurred.

[0030] Finally, based on the obtained inter-frame brightness variation data, a pre-trained convolutional neural network model is used to extract deeper image content change features. The training data for this model includes a large number of video clips labeled with the type and magnitude of content changes. These video clips were obtained from publicly available video datasets such as YouTube-8M and a self-built display content dynamics database. Each video clip is labeled with the type of content change (e.g., scene transition, object movement, or sudden brightness change) and the level of change magnitude. Frame sequences exhibiting significant brightness changes and their corresponding brightness distribution histogram differences are used as input to the model. This convolutional neural network model employs a multi-layer structure, including an input layer, three convolutional layers each followed by a max-pooling layer, two fully connected layers, and an output layer. The convolutional layers use small kernels to extract local spatial features, the pooling layers perform downsampling to reduce dimensionality, and the fully connected layers integrate global information and output a feature vector. During training, supervised learning and stochastic gradient descent algorithms were used for optimization. Mean squared error loss was employed as the loss function to minimize the difference between predicted and true annotations. For parameter selection, the learning rate was determined to be 0.001 through experimental grid search, the batch size was set to 32 based on hardware resources, and the number of training epochs was dynamically adjusted using an early stopping strategy based on performance on the validation set to avoid overfitting. The model automatically learns and extracts feature vectors representing the magnitude of content changes and the degree of inter-frame brightness variation through its convolutional and pooling layers. Based on the model's output, image monitoring results are generated. This result is a structured dataset that explicitly includes quantified values ​​of the magnitude of content changes and the degree of inter-frame brightness variation.

[0031] This step transforms the original image frame sequence into feature indicators that can accurately quantify the dynamic changes in image content, providing a crucial basis for subsequent zoned backlight adjustment. It enables real-time and accurate perception of changes in displayed content, overcoming the shortcomings of traditional methods with slow response times.

[0032] In step S13, based on the image monitoring results, the target brightness and zoning adjustment parameters of the backlight zone are calculated, and the deviation is calculated and the frequency amplitude weight matching degree is analyzed in combination with the real-time brightness to generate zoning adjustment requirement data.

[0033] In one specific implementation, the step of calculating the target brightness and zoning adjustment parameters of the backlight zone based on the image monitoring results, and combining the real-time brightness to perform deviation calculation and frequency amplitude weight matching degree analysis, and generating zoning adjustment requirement data includes: When the image monitoring result exceeds the preset inspection threshold, the number of variations of the inter-frame brightness variation data within a preset unit time is counted using the sliding window method to obtain the number of content changes. Based on the brightness component data corresponding to the number of content changes, the frequency domain feature analysis of the brightness change signal is performed using the Fast Fourier Transform algorithm to extract the main frequency component and obtain the content change frequency. Based on the number of content changes and the frequency of content changes, a weighted average algorithm is used to calculate the weight of the content change magnitude. Based on the content change frequency, the content change magnitude weight, and the brightness component data, a weighted average algorithm is used to perform a fusion calculation to obtain the target brightness value of the backlight zone. Based on the target brightness value, a PID algorithm is used to dynamically calculate the adjustment parameters and generate brightness adjustment parameters, resulting in partitioned adjustment parameters that include the weight of content change amplitude, brightness component data, brightness distribution histogram, target brightness and brightness adjustment parameters. The deviation between the target brightness and the real-time brightness is calculated to obtain the brightness deviation; Based on the brightness deviation and the content change magnitude weight, a weighted average algorithm is used to perform a fusion calculation to obtain the frequency amplitude weight matching degree. When the frequency amplitude weight matching degree is lower than the preset weight matching threshold, it is determined that there is an adjustment need, and partition adjustment need data containing partition adjustment need identifier and partition adjustment parameters is generated.

[0034] Specifically, the system compares the content change amplitude values ​​in the image monitoring results with a preset inspection threshold. The setting of this inspection threshold follows these steps: First, a historical display database covering typical application scenarios is constructed. These scenarios include static images, slow transitions, rapid scene switching, and high dynamic range content. Representative sample segments of stable and drastically changing images are identified through professional evaluation and user surveys. Next, the content change amplitude values ​​corresponding to all frames are extracted from these sample segments, forming a change amplitude dataset. Then, the K-means clustering algorithm is applied to this dataset for unsupervised learning, dividing the data points into two main categories: "stable" and "drastically changing." The algorithm automatically finds the optimal cluster centers through iterative calculation. Subsequently, the positions of these two cluster centers on the numerical axis are analyzed, and a suitable dividing point is selected on the line connecting them, based on the system's trade-off between sensitivity and stability. For example, a point closer to the "stable" cluster center is chosen to prioritize display stability, or the dividing point is set at the midpoint between the two cluster centers to balance response speed and the risk of false triggering. The value of this dividing point is then preliminarily determined as the inspection threshold. Finally, the initial threshold was validated and fine-tuned in a test environment consisting of an independent validation dataset. Its accuracy and recall in distinguishing between stable and drastically changing images were observed. The threshold value was adjusted iteratively until the predetermined discrimination performance target was achieved on the validation set, thus finalizing the adopted inspection threshold. When the content change exceeds this inspection threshold, the current image content is determined to have entered a state of significant dynamic change, requiring the initiation of fine-tuning analysis.

[0035] Subsequently, the system activates a sliding window with a time length of T (the unit time T is usually set to 1 second, but can be adjusted according to needs). Within this window, the system counts the number of frame pairs with brightness similarity below the difference threshold recorded by the inter-frame brightness variation data. This statistical value is taken as the content change count. This count reflects the frequency of significant scene changes per unit time.

[0036] Next, the luminance component data of each frame corresponding to the aforementioned sliding window time period are selected, and these luminance values ​​arranged in chronological order are considered as a discrete signal sequence. A Fast Fourier Transform (FFT) algorithm is applied to this sequence to transform the signal from the time domain to the frequency domain, obtaining its spectral distribution. In the spectrum, the frequency component with the largest amplitude, i.e., the dominant frequency component, is identified. The value of this dominant frequency is extracted as the content change frequency, which characterizes the speed of image brightness changes.

[0037] Then, the obtained content change count and content change frequency are normalized to the range of 0-1 using a max-min normalization method, and then fused using a weighted average algorithm. Appropriate weight coefficients are assigned to the content change count and content change frequency respectively; for example, the weight of the content change count is 0.4, and the weight of the content change frequency is 0.6. The two weighted values ​​are then added together to obtain the content change magnitude weight. This weight comprehensively reflects the intensity and rate of content change.

[0038] Next, the content change frequency, content change amplitude weights, and statistical characteristic values ​​(e.g., average brightness) of the current frame's luminance component data are normalized to the range of 0-1, and then fused again using a weighted average algorithm. This fusion assigns different weights to the three input variables; for example, the content change frequency weight is 0.3, the content change amplitude weight is 0.3, and the average brightness weight is 0.4. The calculated fusion result is the target brightness value for that backlight zone. This target brightness value is the ideal backlight brightness calculated based on the current image content characteristics.

[0039] Based on the calculated target brightness value, a proportional-integral-derivative (PID) algorithm is used to calculate the dynamic adjustment parameters. This algorithm generates brightness adjustment parameters for controlling the backlight brightness adjustment process by multiplying the difference between the target brightness value and a reference brightness value (proportional term), the integral of that difference (integral term), and the derivative of that difference (derivative term) by specific coefficients and then summing them. These parameters collectively constitute the zone adjustment parameters. This parameter set also includes the previously calculated weights of content change amplitudes, the brightness component data used, the brightness distribution histogram information, and the target brightness value. The PID parameters are experimentally tuned, with typical values ​​of Kp=0.5, Ki=0.1, and Kd=0.01, and can be adjusted according to the system response.

[0040] Simultaneously, the difference between the target brightness value and the real-time brightness value obtained from step S11 is calculated to obtain the brightness deviation. This deviation reflects the gap between the current actual backlight brightness and the ideal brightness.

[0041] Next, the calculated brightness deviation and content change amplitude weights are normalized to the range of 0-1, and then fused again using a weighted average algorithm. For example, the weight of brightness deviation is 0.7, and the weight of content change amplitude is 0.3, resulting in a frequency amplitude weight matching degree. The frequency amplitude weight matching degree is a real-time calculated scalar value used to instantly assess the degree of agreement between the current actual backlight state and the ideal state derived from image content change characteristics (including change frequency and amplitude). The lower the value, the greater the real-time deviation and the more urgent the adjustment need. The system presets a weight matching threshold, which is determined based on the following specific implementation steps: First, a test environment covering typical application scenarios such as web browsing, video playback, and game interaction is built, and testers with normal or corrected-to-normal vision are recruited to form an evaluation team. During the test, the system runs sequentially under different candidate thresholds and displays the same series of dynamically displayed content to the evaluation team. For each candidate threshold, two sets of data are recorded simultaneously: first, subjective ratings of visual comfort based on user feedback are collected through questionnaires, including aspects such as screen smoothness, flicker, and fatigue during prolonged viewing; second, real-time energy consumption data of the display device at this threshold is precisely measured using a power meter. After testing all candidate thresholds, the collected subjective comfort ratings and objective energy consumption data are normalized and assigned predefined importance weights. A weighted average is then used to generate a comprehensive performance score for each candidate threshold. Finally, the threshold that yields the highest comprehensive performance score is selected as the final weighted matching threshold used by the system. This value represents the optimal balance between visual comfort and system energy efficiency under given test conditions. When the calculated frequency amplitude weighted matching degree is lower than this preset weighted matching threshold, the system determines that there is a need to adjust the backlight zone.

[0042] The frequency-amplitude weighted matching degree is a scalar value calculated in real time. It is used to instantly assess the degree of agreement between the current actual backlight state and the ideal state derived based on the image content change characteristics (including change frequency and amplitude). The lower the value, the greater the real-time deviation and the more urgent the adjustment need. Finally, when an adjustment need is determined, the system generates zonal adjustment need data. This data includes a clear zonal adjustment need identifier (used to indicate the specific zonal that needs adjustment) and all the zonal adjustment parameters calculated above.

[0043] This step closely integrates the dynamic characteristics of the image content (number of changes, frequency, and amplitude) with the real-time status of the backlight zones. Through multi-level weighted fusion and PID control algorithms, abstract image changes are transformed into specific, executable zone brightness adjustment instructions and parameters, achieving a precise mapping from content perception to adjustment decisions. This lays a solid foundation for subsequent smoothing and drive execution, ensuring the accuracy and adaptability of zone brightness control.

[0044] In step S14, smoothing is performed based on the frequency amplitude weight matching degree and the partition adjustment requirement data to obtain the partition brightness curve and smoothing real-time index.

[0045] In one specific implementation, the smoothing process based on the partition adjustment demand data to obtain the partition brightness curve and real-time smoothing index includes: Based on the partition adjustment requirement data, a Gaussian filtering algorithm is used to suppress noise and smooth the brightness component data to generate a smooth brightness sequence. The statistical features of the brightness distribution histogram are extracted, and the statistical features of the brightness distribution histogram are fused into the smooth brightness sequence using a linear interpolation algorithm to obtain the fused adjustment sequence. Based on the fusion adjustment sequence, the least squares method is used to perform curve fitting to construct the partition brightness curve; Based on the partition brightness curve, a smoothness index is analyzed using a pre-trained convolutional neural network model to generate a smoothness real-time index.

[0046] Specifically, the system first smooths and denoises the luminance component data in the zonal adjustment demand data. A Gaussian filtering algorithm is used (the Gaussian filter kernel size is typically 5, and the standard deviation is 1.0). This algorithm calculates the weighted average of the neighboring data points for each data point in the luminance component data sequence. Data points closer to the center point are assigned higher weights, and these weights follow a Gaussian distribution (i.e., a normal distribution). Through this weighted averaging calculation, interference caused by random noise or instantaneous fluctuations in the luminance data is effectively suppressed, generating a new, smoother luminance sequence with more gradual changes.

[0047] Next, the system extracts key statistical features from the brightness distribution histogram contained in the zonal adjustment demand data, such as the peak position and distribution variance of the histogram. Then, using a linear interpolation algorithm, these statistical feature values ​​are fused into the generated smooth brightness sequence according to their corresponding time points. Specifically, at each corresponding time point in the smooth brightness sequence, a new adjustment value is calculated through a linear relationship based on the brightness value at that time and the histogram statistical feature value, thus forming a fused adjustment sequence that contains both smooth brightness information and reflects the brightness distribution characteristics.

[0048] Subsequently, based on the obtained fusion adjustment sequence, the system uses the least squares method for curve fitting (least squares fitting typically uses a quadratic polynomial) to construct the partitioned brightness curve. This method aims to find a continuous function curve that minimizes the sum of the squares of the distances between this curve and all data points in the fusion adjustment sequence along the vertical axis. By solving this optimization problem, a smooth and continuous partitioned brightness curve function that best represents the brightness variation trend can be obtained.

[0049] Finally, the constructed partitioned brightness curves are input into a pre-trained convolutional neural network model, which analyzes the smoothness of the curves. This model was trained on a large dataset of brightness curves labeled with different smoothness levels. It can automatically extract the morphological features of the curves and output a quantified smoothness index value, which is the real-time smoothness index. This index objectively reflects the smoothness of the partitioned brightness curves.

[0050] This step involves a series of processes, including Gaussian filtering to denoise the raw adjustment request data, fusing histogram statistical features, curve fitting, and intelligent smoothness evaluation. This transforms potentially noisy and abrupt brightness adjustment requirements into a smooth, continuous, and ideal brightness change trajectory that reflects the statistical characteristics of the image content, and provides a quantitative evaluation of its smoothness. This provides crucial information for generating stable, flicker-free backlight drive signals, avoiding visual discomfort caused by sudden brightness changes, and improving visual comfort and image stability.

[0051] In step S15, the intermediate compensation coefficients for partition adjustment are calculated by correlating and fusing the partition brightness curves and the smooth real-time indexes to obtain the compensation coefficient set and the curve driving basis.

[0052] In one specific implementation, the step of correlating and fusing the partition brightness curve and the smoothing real-time index to calculate the intermediate compensation coefficient for partition adjustment, thereby obtaining a set of compensation coefficients and a curve-driving basis, includes: Based on the partition brightness curve and the smoothing real-time index, combined with the content change magnitude weight and the brightness adjustment parameter, the content change magnitude weight and the brightness adjustment parameter are fused using a weighted average algorithm to obtain the adjustment fusion parameter; Based on the adjustment and fusion parameters, the proportional relationship between the brightness adjustment parameters and the content change magnitude weight is analyzed by linear regression algorithm, and the intermediate compensation coefficient corresponding to each partition is calculated. Based on the intermediate compensation coefficient and the partition brightness curve, a pre-trained convolutional neural network model is used to extract curve-driven features and generate a set of curve-driven criteria and compensation coefficients to guide backlight adjustment.

[0053] Specifically, the system first takes four data points as input: the current value of the partition brightness curve, the value of the smoothing real-time index, the weight of content change magnitude, and the brightness adjustment parameter. The partition brightness curve and the smoothing real-time index are dynamically updated time-series data. The partition brightness curve is re-generated by the smoothing index analysis module in each processing cycle based on the latest fusion adjustment sequence, while the smoothing real-time index is synchronously calculated by the convolutional neural network model in this module based on the newly generated partition brightness curve. These two features are updated at the end of each system processing cycle and used as input for the correlation fusion calculation in the next cycle, thus forming a continuously iterative, closed-loop data flow, ensuring that the brightness adjustment decision is always based on the latest system state. A weighted average algorithm is used to fuse these four data points. A preset weight coefficient is assigned to each of the partition brightness curve value, smoothing real-time index, content change magnitude weight, and brightness adjustment parameter. For example, these four weight coefficients can all be set to 0.25. Each data point is multiplied by its corresponding weight coefficient and then summed to obtain a comprehensive value, which is defined as the adjustment fusion parameter. This parameter comprehensively reflects the target brightness trend, curve smoothness, content change intensity, and instantaneous adjustment requirements.

[0054] Next, based on the obtained adjustment and fusion parameters, the system uses a linear regression algorithm to analyze the inherent proportional relationship between the brightness adjustment parameters and the content change magnitude weight. This algorithm retrieves brightness adjustment parameter data from a historical data buffer dedicated to storing recent states, representing multiple sampling moments within a recent consecutive time period, as the dependent variable set (data from the most recent 100 time points can be used). It then uses the content change magnitude weight data calculated by the adjustment demand analysis module at these same moments as the independent variable set. Subsequently, a univariate linear regression is performed on this series of data points using the least squares method. This method finds a line that best represents the overall trend of brightness adjustment parameters changing with the content change magnitude weight by minimizing the sum of the squares of the squares of the vertical distances (i.e., residuals) from all data points to the fitted line. Finally, the slope of this line is directly calculated through this fitting process, and this slope is determined as the ratio of the brightness adjustment parameter change with the content change magnitude weight. Based on this proportional relationship and the current adjustment and fusion parameters, an intermediate compensation coefficient is calculated for each backlight zone. This coefficient is a numerical value used to compensate the baseline brightness value in subsequent adjustments to more accurately match content dynamics.

[0055] Then, the system inputs the calculated set of intermediate compensation coefficients for all partitions, along with the partition brightness curves generated in step S14, into a pre-trained convolutional neural network model. This model employs a network structure containing an input layer, two one-dimensional convolutional layers, a max-pooling layer, a flattening layer, and two fully connected layers. The one-dimensional convolutional layers extract local temporal features from the partition brightness curves and compensation coefficient sequences, while the fully connected layers synthesize these features and perform high-level inference. The model's training dataset is constructed through the following automated process: collecting publicly available standard test video sequences covering standard and high dynamic ranges; for each frame, using three objective evaluation algorithms—structural similarity, detail retention, and local contrast maintenance—to generate an ideal backlight adjustment curve that simultaneously optimizes all the aforementioned metrics; and finally, selecting sample pairs with comprehensive scores higher than a preset standard to constitute the training dataset. The training process employs supervised learning, using mean squared error as the loss function, and iteratively updating the model parameters through backpropagation combined with the Adam optimizer. In terms of parameter selection, the initial learning rate was set to 0.01, and a learning rate decay strategy was adopted. The batch size was set to 64 based on computing resources, and the number of training epochs was automatically determined based on the performance on the validation set using early stopping. The model, through its convolutional and fully connected layers, extracts deep, curve-driven features from the input partition brightness curve shape and intermediate compensation coefficient set, effectively guiding backlight adjustment. The model's output is structured into two parts: first, the curve-driven basis, a set of feature vectors describing how the brightness curve should be driven to adapt to content changes; and second, the compensation coefficient set, the final set of intermediate compensation coefficients for each partition after model validation and optimization.

[0056] This step, through weighted fusion of multi-source information and linear regression analysis, correlates abstract content features, adjustment parameters, and specific brightness curves to calculate preliminary compensation coefficients. Then, a pre-trained intelligent model is used to perform deep feature extraction and optimization on the preliminary results, generating precise driving instructions that accurately reflect the dynamic needs of the content while meeting visual comfort requirements. This transforms backlight adjustment from a simple brightness mapping into predictive and adaptive intelligent control, providing a decision-making basis for ultimately achieving high-precision and high-stability display output.

[0057] In step S16, based on the compensation coefficient set and the curve driving basis, combined with the partition brightness curve and the smoothing real-time index, the matching degree between the compensation requirement and the weighted service is determined, and the matching evaluation result is obtained.

[0058] In one specific implementation, the step of determining the matching degree between compensation requirements and weighted services based on the compensation coefficient set and the curve-driven criteria, combined with the partition brightness curve and the smoothing real-time index, to obtain a matching evaluation result includes: Based on the aforementioned smoothing real-time index, a smoothing process is performed using a mean filtering algorithm to obtain a smoothed change sequence; Based on the smoothed change sequence, the mutation response features are extracted using the gradient descent algorithm to obtain the mutation response set; Based on the set of variant responses and the set of compensation coefficients, a weighted average algorithm is used to perform a fusion calculation to obtain the fusion weight parameters; Based on the fusion weight parameters and the curve-driven criteria, a convolutional neural network algorithm is used to analyze the business matching degree and compensation requirements to obtain the matching evaluation results.

[0059] Specifically, the smoothed real-time indicators are first further smoothed. A mean filtering algorithm is used to take the current smoothed real-time indicator value and the indicator values ​​of several adjacent time points, calculate the arithmetic mean of these values, and replace the original value at the current time with this average. This method is applied point-by-point to the entire smoothed real-time indicator sequence, generating a smoother sequence with smaller fluctuations and a gentler trend. This operation aims to eliminate potential short-term fluctuations in the indicators and highlight their long-term trends.

[0060] Next, based on the obtained smoothed change sequence, the gradient descent algorithm is applied to extract the mutation response features. This algorithm iteratively calculates the first-order difference (i.e., rate of change) between adjacent data points in the smoothed change sequence and adjusts a virtual probe point position along the direction that reduces the absolute value of this rate of change. This iterative process continues until the following two conditions are met simultaneously: first, a preset maximum number of iterations is reached to prevent infinite loops; second, in consecutive iterations, the difference in the absolute values ​​of the rates of change corresponding to the probe point positions is less than a set minimum positive threshold, indicating that the position has stabilized and converged to a local extremum. Finally, key points where the rate of change in the sequence undergoes a significant inflection are found. The sequence values ​​and their rates of change corresponding to these key points are combined to form a set of feature data, called the mutation response set. This set quantifies the response pattern of the smoothed change sequence to external disturbances (such as sudden changes in image content).

[0061] Subsequently, the variability response set and the compensation coefficient set are fused. A weighted average algorithm is used to assign weights to each feature data point in the variability response set and the corresponding compensation coefficient in the compensation coefficient set, respectively. The weighted average value for each partition is calculated; for example, the weights for the variability response features and the compensation coefficients are each set to 0.5, resulting in a comprehensive value called the fusion weight parameter. This parameter simultaneously incorporates the variability response characteristics extracted from historical smoothing trends and the pre-calculated brightness compensation requirements.

[0062] Finally, the fusion weight parameters and curve-driven criteria are input into a convolutional neural network algorithm for analysis. This algorithm is trained on a large number of data pairs (fusion weight parameters and curve-driven criteria) labeled "good match" or "needs adjustment." Through its multi-layer convolution and pooling operations, the model extracts deep features from the input fusion weight parameters and curve-driven criteria, ultimately outputting a quantitative evaluation result, i.e., the matching evaluation result. This result clearly characterizes the degree of matching between the current compensation requirement and the business requirement represented by the weight of content change magnitude.

[0063] This step, through sequence smoothing, gradient feature extraction, multi-source data fusion, and intelligent network analysis, deeply integrates and diagnoses the various parameters and data generated in the preceding steps, ultimately generating an authoritative assessment of whether the adjustment strategy dynamically and accurately matches the image content. This assessment result is the key criterion for deciding whether to perform the final brightness adjustment and how to adjust it, ensuring that the entire system's adjustment actions always remain highly consistent with the actual needs of the displayed content, thereby achieving precise, stable, and comfortable brightness control in complex dynamic scenarios.

[0064] In step S17, the brightness level of the partition is adjusted and the inter-frame brightness variation is compensated based on the matching evaluation result to determine the display output state and the range of changes in stable output content.

[0065] In one specific implementation, adjusting the brightness level of the partition and compensating for inter-frame brightness variations based on the matching evaluation result, and determining the display output state and the range of changes in stable output content, includes: Based on the matching evaluation results and the inter-frame brightness variation data, a backlight adjustment instruction set is generated through the Drools rule engine; According to the backlight adjustment instruction set, pulse width modulation is used to drive the backlight module to adjust the brightness level of the partition, and exponential smoothing algorithm is used to process and compensate for the brightness variation between frames, and finally display the output status and the range of changes in stable output content.

[0066] Specifically, the system inputs the matching evaluation results and inter-frame brightness variation data into the Drools rule engine. This rule engine contains a series of pre-defined business rules, derived from analyzing numerous successful adjustment cases in historical display optimization scenarios. For example, a rule might be defined as: when the matching evaluation result is below a certain threshold and the inter-frame brightness variation data indicates continuous and rapid changes, an instruction to increase the speed and magnitude of brightness adjustment is generated. The rule engine uses a pattern matching mechanism to compare the input data with conditions in the rule base, triggering all rules that meet the conditions, thereby logically generating a specific and executable set of backlight adjustment instructions. This instruction set explicitly specifies the target brightness level to be adjusted for each backlight zone and the compensation strategy for inter-frame variations.

[0067] Next, the system uses pulse width modulation (PWM) to drive the physical backlight module based on the generated backlight adjustment instruction set. For each zone requiring adjustment, the algorithm precisely controls its brightness by changing the duty cycle (the ratio of the high-level time to the total cycle) of the pulse signal input to the backlight LED. The higher the duty cycle, the greater the average luminous intensity of the LED in that zone within one cycle, and the higher the perceived brightness. By converting the target brightness value in the instruction set into a specific duty cycle parameter, the hardware circuit is directly driven, achieving precise control of the brightness level of each zone.

[0068] Meanwhile, to compensate for inter-frame brightness variations, the system utilizes an exponential smoothing algorithm to process the backlight brightness adjustment process. This algorithm weights and fuses the target brightness value calculated for the current frame based on the instruction set with the adjusted actual brightness value from the previous frame. For example, the weighting coefficient for the previous frame's brightness value is set to 0.7, and the weighting coefficient for the current frame's target brightness value is set to 0.3. Through this weighted calculation, a new, smoother final brightness value is obtained. This processing method ensures that inter-frame brightness changes are not abrupt but rather present a smooth transition, suppressing visual flickering or jitter that may be caused by rapid switching of image content.

[0069] Ultimately, the system determines the current display output state as the partition brightness state after pulse width modulation driving and exponential smoothing compensation, and records the range of image content changes in this stable state as the output stable content change range. This final output state is the result of the combined effect of brightness adjustment and variation compensation.

[0070] This step transforms all the intelligent decision-making results from the previous analysis, calculation, and evaluation into specific control commands through a rules engine. It then utilizes pulse width modulation (PWM) and exponential smoothing—algorithms adept at precise control and smooth transitions, respectively—to ultimately apply these commands to the physical hardware, achieving a closed loop from intelligent decision-making to physical execution. This ensures that backlight adjustment can both quickly respond to changes in image content and maintain smooth and stable output brightness, ultimately achieving the core goals of improving display accuracy, response speed, and visual comfort under dynamic content.

[0071] Reference Figure 2 The second embodiment of the present invention provides an intelligent dynamic adjustment system for LED backlight brightness, comprising: The data acquisition module is used to acquire image frame sequences and real-time brightness from the display device; The image monitoring module is used to extract the brightness distribution histogram and calculate the inter-frame brightness variation based on the image frame sequence, extract image content change features, and generate image monitoring results; The adjustment demand analysis module is used to calculate the target brightness and adjustment parameters of the backlight zone based on the image monitoring results, and to perform deviation calculation and frequency amplitude weight matching degree analysis in combination with the real-time brightness to generate zone adjustment demand data. The smoothing index analysis module is used to perform smoothing processing based on the frequency amplitude weight matching degree and the partition adjustment demand data to obtain the partition brightness curve and the real-time smoothing index. The curve driving module is used to correlate and fuse the partition brightness curve and the smooth real-time index to calculate the intermediate compensation coefficient for partition adjustment, and obtain the compensation coefficient set and curve driving basis. The matching evaluation module is used to determine the matching degree between compensation requirements and weighted services based on the compensation coefficient set and the curve driving basis, combined with the partition brightness curve and the smoothing real-time index, and obtain the matching evaluation result. The output module is used to adjust the brightness level of the partition and compensate for inter-frame brightness variations based on the matching evaluation results, and to determine the display output status and the range of changes in stable output content.

[0072] It should be noted that the LED backlight brightness intelligent dynamic adjustment system provided in this embodiment of the invention is used to execute all the process steps of the LED backlight brightness intelligent dynamic adjustment method in the above embodiment. The working principle and beneficial effects of the two are one-to-one, so they will not be described again.

[0073] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as an intelligent dynamic adjustment program for LED backlight brightness. When the processor executes the computer program, it implements the steps described in the various embodiments of the intelligent dynamic adjustment method for LED backlight brightness, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above system embodiments, such as the LED backlight brightness intelligent dynamic adjustment module.

[0074] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0075] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0076] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0077] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0078] If the modules / units integrated into the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or system capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0079] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0080] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for intelligent dynamic adjustment of LED backlight brightness, characterized in that, include: Acquire image frame sequences and real-time brightness from the display device; Based on the image frame sequence, a brightness distribution histogram is extracted and inter-frame brightness variation is calculated to extract image content change features and generate image monitoring results. Based on the image monitoring results, the target brightness and zoning adjustment parameters of the backlight zone are calculated, and the deviation is calculated and the frequency amplitude weight matching degree is analyzed in combination with the real-time brightness to generate zoning adjustment requirement data. Based on the frequency amplitude weight matching degree and the partition adjustment requirement data, smoothing is performed to obtain the partition brightness curve and smoothing real-time index. Based on the correlation and fusion of the partition brightness curve and the smooth real-time index, the intermediate compensation coefficient for partition adjustment is calculated to obtain the compensation coefficient set and the curve driving basis. Based on the compensation coefficient set and the curve driving basis, combined with the partition brightness curve and the smoothing real-time index, the matching degree between compensation demand and weighted services is determined, and the matching evaluation result is obtained. Based on the matching evaluation results, adjust the brightness level of the partition and compensate for inter-frame brightness variations to determine the display output status and the range of changes in stable output content.

2. The LED backlight brightness intelligent dynamic adjustment method according to claim 1, characterized in that, The step involves extracting a brightness distribution histogram and calculating inter-frame brightness variations based on the image frame sequence, extracting image content change features, and generating image monitoring results, including: Based on the image frame sequence, a weighted average algorithm is used to calculate the brightness value of each pixel to obtain the brightness component data of each frame image; Based on the brightness component data, the pixel number distribution is statistically analyzed using the histogram statistical method to obtain the brightness distribution histogram for each frame. Based on the brightness distribution histogram, the brightness similarity of the brightness distribution histograms of two adjacent frames is calculated using the histogram intersection algorithm. When the brightness similarity is lower than a preset difference threshold, it is determined that there is a significant change between the two corresponding frames, and inter-frame brightness variation data is obtained. Based on the inter-frame brightness variation data, a pre-trained convolutional neural network model is used to extract image content change features and generate image monitoring results that include the magnitude of content change and the degree of inter-frame brightness variation.

3. The intelligent dynamic adjustment method for LED backlight brightness according to claim 2, characterized in that, Based on the image monitoring results, the target brightness and zoning adjustment parameters of the backlight zones are calculated, and deviation calculation and frequency amplitude weight matching degree are performed in conjunction with the real-time brightness to generate zoning adjustment requirement data, including: When the image monitoring result exceeds the preset inspection threshold, the number of variations of the inter-frame brightness variation data within a preset unit time is counted using the sliding window method to obtain the number of content changes. Based on the brightness component data corresponding to the number of content changes, the frequency domain feature analysis of the brightness change signal is performed using the Fast Fourier Transform algorithm to extract the main frequency component and obtain the content change frequency. Based on the number of content changes and the frequency of content changes, a weighted average algorithm is used to calculate and obtain the weight of the content change magnitude. Based on the content change frequency, the content change magnitude weight, and the brightness component data, a weighted average algorithm is used to perform a fusion calculation to obtain the target brightness value of the backlight zone. Based on the target brightness value, a PID algorithm is used to dynamically calculate the adjustment parameters and generate brightness adjustment parameters, resulting in partitioned adjustment parameters that include the weight of content change amplitude, brightness component data, brightness distribution histogram, target brightness and brightness adjustment parameters. The deviation between the target brightness and the real-time brightness is calculated to obtain the brightness deviation; Based on the brightness deviation and the content change magnitude weight, a weighted average algorithm is used to perform a fusion calculation to obtain the frequency amplitude weight matching degree. When the frequency amplitude weight matching degree is lower than the preset weight matching threshold, it is determined that there is an adjustment need, and partition adjustment need data containing partition adjustment need identifier and partition adjustment parameters is generated.

4. The LED backlight brightness intelligent dynamic adjustment method according to claim 2, characterized in that, The step of smoothing the data based on the partition adjustment requirements to obtain the partition brightness curve and real-time smoothing index includes: Based on the partition adjustment requirement data, a Gaussian filtering algorithm is used to suppress noise and smooth the brightness component data to generate a smooth brightness sequence. The statistical features of the brightness distribution histogram are extracted, and the statistical features of the brightness distribution histogram are fused into the smooth brightness sequence using a linear interpolation algorithm to obtain the fused adjustment sequence. Based on the fusion adjustment sequence, the least squares method is used to perform curve fitting to construct the partition brightness curve; Based on the partition brightness curve, a smoothness index is analyzed using a pre-trained convolutional neural network model to generate a smoothness real-time index.

5. The intelligent dynamic adjustment method for LED backlight brightness according to claim 3, characterized in that, The step involves correlating and fusing the partition brightness curve and the smoothing real-time index to calculate intermediate compensation coefficients for partition adjustment, thereby obtaining a set of compensation coefficients and curve-driving basis, including: Based on the partition brightness curve and the smoothing real-time index, combined with the content change magnitude weight and the brightness adjustment parameter, the content change magnitude weight and the brightness adjustment parameter are fused using a weighted average algorithm to obtain the adjustment fusion parameter; Based on the adjustment and fusion parameters, the proportional relationship between the brightness adjustment parameters and the content change magnitude weight is analyzed by linear regression algorithm, and the intermediate compensation coefficient corresponding to each partition is calculated. Based on the intermediate compensation coefficient and the partition brightness curve, a pre-trained convolutional neural network model is used to extract curve-driven features and generate a set of curve-driven criteria and compensation coefficients to guide backlight adjustment.

6. The intelligent dynamic adjustment method for LED backlight brightness according to claim 1, characterized in that, The step of determining the matching degree between compensation requirements and weighted services based on the compensation coefficient set and the curve-driven criteria, combined with the partition brightness curve and the smoothing real-time index, to obtain a matching evaluation result includes: Based on the aforementioned smoothing real-time index, a smoothing process is performed using a mean filtering algorithm to obtain a smoothed change sequence; Based on the smoothed change sequence, the mutation response features are extracted using the gradient descent algorithm to obtain the mutation response set; Based on the set of variant responses and the set of compensation coefficients, a weighted average algorithm is used to perform a fusion calculation to obtain the fusion weight parameters; Based on the fusion weight parameters and the curve-driven criteria, a convolutional neural network algorithm is used to analyze the business matching degree and compensation requirements to obtain the matching evaluation results.

7. The intelligent dynamic adjustment method for LED backlight brightness according to claim 2, characterized in that, The step of adjusting the brightness level of the partition and compensating for inter-frame brightness variations based on the matching evaluation results, and determining the display output state and the range of changes in stable output content, includes: Based on the matching evaluation results and the inter-frame brightness variation data, a backlight adjustment instruction set is generated through the Drools rule engine; According to the backlight adjustment instruction set, pulse width modulation is used to drive the backlight module to adjust the brightness level of the partition, and exponential smoothing algorithm is used to process and compensate for the brightness variation between frames, and finally display the output status and the range of changes in stable output content.

8. An intelligent dynamic brightness adjustment system for LED backlights, characterized in that, include: The data acquisition module is used to acquire image frame sequences and real-time brightness from the display device; The image monitoring module is used to extract the brightness distribution histogram and calculate the inter-frame brightness variation based on the image frame sequence, extract image content change features, and generate image monitoring results; The adjustment demand analysis module is used to calculate the target brightness and adjustment parameters of the backlight zone based on the image monitoring results, and to perform deviation calculation and frequency amplitude weight matching degree analysis in combination with the real-time brightness to generate zone adjustment demand data. The smoothing index analysis module is used to perform smoothing processing based on the frequency amplitude weight matching degree and the partition adjustment demand data to obtain the partition brightness curve and the real-time smoothing index. The curve driving module is used to correlate and fuse the partition brightness curve and the smooth real-time index to calculate the intermediate compensation coefficient for partition adjustment, and obtain the compensation coefficient set and curve driving basis. The matching evaluation module is used to determine the matching degree between compensation requirements and weighted services based on the compensation coefficient set and the curve driving basis, combined with the partition brightness curve and the smoothing real-time index, and obtain the matching evaluation result. The output module is used to adjust the brightness level of the partition and compensate for inter-frame brightness variations based on the matching evaluation results, and to determine the display output status and the range of changes in stable output content.

Images