Adaptive adjustment method for ambient light brightness

Abstract

The invention provides an ambient light brightness adaptive adjustment method, which comprises the steps of collecting and preprocessing an ambient light brightness data sequence, determining a brightness analysis window, determining an ambient brightness label according to a target light brightness data sequence, generating a negative adjustment brightness strategy if the ambient brightness label is a high dynamic range label, and adjusting the brightness according to the negative adjustment brightness strategy. And if the label is a low dynamic range label, generating a forward adjustment brightness strategy, determining abnormal brightness information according to the current or updated brightness strategy in combination with the abnormal brightness information and a brightness mutation prediction model, and if there is an abnormal regulation and control, outputting a regulation and control instruction according to an abnormal regulation and control decision tree. And carrying out brightness locking on the display equipment or activating a safe brightness mode. According to the invention, accurate, dynamic and stable brightness adjustment of the display equipment can be realized, the user experience is improved, and the energy consumption is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic device display technology, and more particularly, to a method for adaptively adjusting ambient light brightness. Background Art

[0002] In today's digital age, electronic devices such as smartphones, tablets, and laptops play an indispensable role in people's daily lives. The displays of these devices need to adjust their brightness based on varying ambient light conditions to ensure a good viewing experience in all scenarios while avoiding eye discomfort caused by screens that are too bright or too dim. Currently, many electronic devices are equipped with light sensors and cameras to detect ambient light levels and adjust the display brightness accordingly. However, existing brightness adjustment methods often have limitations.

[0003] Existing methods for adjusting ambient light brightness primarily rely on a single data source collected by a photosensor, adjusting the display brightness through a simple threshold judgment. While this method can achieve brightness adjustment to a certain extent, it suffers from the following problems: First, single photosensor data cannot accurately reflect complex ambient light conditions, especially in high dynamic range (HDR) and low dynamic range (LDR) scenarios, resulting in poor adjustment results. Second, existing methods lack effective mechanisms for predicting and processing sudden changes in brightness, resulting in a lag in display brightness adjustment when ambient light changes rapidly, impacting the user experience. Furthermore, existing brightness adjustment strategies lack adaptive learning capabilities and are unable to dynamically adjust based on user habits and preferences.

[0004] In the process of implementing the embodiments of the present invention, there are at least the following problems or defects in the existing technology: it is impossible to accurately identify the changes in ambient light brightness in high dynamic range and low dynamic range scenarios, resulting in insufficient adjustment accuracy; there is a lack of effective prediction and processing mechanism for sudden brightness changes, which makes the display screen unable to respond in time when the ambient light brightness changes rapidly; it is impossible to adaptively learn and adjust according to the user's usage habits and preferences, resulting in insufficient personalization of the adjustment strategy. Summary of the Invention

[0005] The present invention provides an ambient light adaptive brightness adjustment method, which is applied to electronic devices equipped with light-sensitive sensors and cameras, comprising: collecting a current ambient light brightness data sequence; performing data preprocessing on the ambient light brightness data sequence to obtain a preprocessed brightness data sequence; determining a brightness analysis window according to the preprocessed brightness data sequence; determining an ambient light label according to a target brightness data sequence, wherein the target light brightness data sequence is preprocessed brightness data in the preprocessed brightness data sequence and located within the brightness analysis window, and the ambient light label represents the current ambient light intensity level; in response to determining that the ambient light label is a high dynamic range label, generating a display-specific brightness control strategy according to the current brightness strategy and the ambient light label corresponding to the electronic device. A negative adjustment brightness strategy for the device is used as the updated brightness strategy; in response to determining that the ambient brightness label is a low dynamic range label, a positive adjustment brightness strategy for the display device is generated according to the current brightness strategy and the ambient brightness label as the updated brightness strategy; according to the current brightness strategy or the updated brightness strategy, the display device is brightness controlled, including: determining abnormal brightness information according to the current brightness strategy or the updated brightness strategy, as well as the ambient brightness label and the brightness mutation prediction model; in response to determining that the abnormal brightness information represents that the current brightness strategy or the updated brightness strategy has a control abnormality, determining the control instruction corresponding to the abnormal brightness information in the abnormal control decision tree; according to the control instruction, the display device is brightness locked or a safe brightness mode is activated.

[0006] Furthermore, the ambient light brightness data sequence is preprocessed to obtain a preprocessed brightness data sequence, including: for each ambient light brightness data in the ambient light brightness data sequence, executing the following processing steps: performing sensor validity verification on the ambient light brightness data; in response to determining that the ambient light brightness data passes the validity verification, performing noise filtering on the ambient light brightness data to obtain filtered brightness data; performing data normalization on the filtered brightness data to obtain normalized brightness data; determining a sampling granularity corresponding to the ambient light brightness data sequence, wherein the sampling granularity represents the frequency of collecting the ambient light brightness data; in response to determining that the sampling granularity is not a preset sampling granularity, and the sampling granularity is smaller than the preset sampling granularity, downsampling the normalized brightness data sequence with the preset sampling granularity to obtain a preprocessed brightness data sequence.

[0007] Further, according to the preprocessed brightness data sequence, determining the brightness analysis window includes: initializing the window length corresponding to the brightness analysis window to obtain an initial brightness analysis window, wherein the window length of the initial brightness analysis window is a reference window length; performing the following window determination steps according to the reference window length and the preprocessed brightness data sequence: determining a target brightness data sequence, wherein the target brightness data in the target brightness data sequence is the preprocessed brightness data located within the initial brightness analysis window; calculating the brightness mean according to the target brightness data sequence; calculating the brightness mean according to the target brightness data sequence, the brightness mean, and the first target brightness; The first target light brightness data and the second target light brightness data are used to calculate the window characteristic value of the initial brightness analysis window, wherein the first target light brightness data is the maximum value of the target light brightness data sequence, and the second target light brightness data is the minimum value of the target light brightness data sequence; in response to determining that the light mean value is greater than or equal to the dynamic threshold, the initial brightness analysis window is determined as the brightness analysis window, wherein the dynamic threshold is dynamically configured according to the working scene of the display device; in response to determining that the light mean value is less than the dynamic threshold, the window length of the initial brightness analysis window is incremented to obtain the brightness analysis window with incremented length as the initial brightness analysis window, and the window determination step is performed again.

[0008] Furthermore, determining the ambient brightness label according to the target light brightness data sequence includes: for each target light brightness data in the target light brightness data sequence, calculating the ambient brightness value according to the following formula: ;in, is the calculated ambient brightness value, is the light sensor data, is the pixel position The weight coefficient of is the pixel grayscale value, is the predicted value of the convolutional neural network, is the weighting coefficient and .

[0009] Based on the calculated ambient brightness value sequence, a high dynamic range threshold and a low dynamic range threshold are determined; in response to the ambient brightness value continuously exceeding the high dynamic range threshold for a first time threshold, the ambient brightness label is determined to be a high dynamic range label; in response to the ambient brightness value continuously being lower than the low dynamic range threshold for a second time threshold, the ambient brightness label is determined to be a low dynamic range label.

[0010] Furthermore, the brightness mutation prediction model is constructed through the following steps: establishing a historical brightness control data set, which includes the ambient light brightness data sequence and the corresponding control strategy.

[0011] Extract brightness change gradient features: in, is the brightness change gradient at time t, is the brightness value at time t, is the brightness value at time t1, is the time interval.

[0012] Extract brightness fluctuation variance features: ;in, is the brightness fluctuation variance, is the sample size, is the ith brightness sample value, is the mean of brightness samples; the gradient features and variance features are input into the support vector machine classifier, and the brightness mutation prediction model is obtained through training.

[0013] Furthermore, the abnormal control decision tree includes: the first level node: judging whether the brightness change rate exceeds the safety threshold ,in They are respectively the maximum and minimum brightness values ​​allowed by the current environment.

[0014] Second level node: responds to the brightness change rate exceeding , to determine whether the time threshold is exceeded continuously .

[0015] Third-level node: In response to a duration exceeding , output the first control instruction to activate brightness lock.

[0016] Fourth-level node: Response duration does not exceed , output the second control instruction to activate the safe brightness mode.

[0017] Furthermore, the safe brightness mode includes: using an S-shaped brightness transition curve for brightness adjustment: ;in, is the adjusted brightness value at time t, To adjust the starting brightness value, is the target brightness value, is the transition rate factor, To adjust the starting time point.

[0018] Furthermore, it also includes a deep learning model update mechanism: recording the user's manual brightness adjustment operation data; when the deviation between manual adjustment and automatic adjustment continues to exceed the preset deviation threshold, triggering model retraining: ;in, is the loss function value, is the sample size, is the automatically adjusted brightness value of the i-th sample, is the manually adjusted brightness value of the i-th sample, is the regularization coefficient, is the convolutional neural network weight parameter vector; the convolutional neural network weight parameters are updated through the back propagation algorithm .

[0019] Furthermore, model retraining includes: freezing the basic feature layer using a transfer learning strategy; updating only the weight parameters of the fully connected layer; and using a momentum optimizer to accelerate convergence: ;in, is the weight parameter at time t+1, is the weight parameter at time t, is the weight parameter at time t1 is the learning rate, is the gradient of the loss function, is the momentum coefficient.

[0020] Furthermore, the brightness control execution includes: generating a PWM dimming control signal: ;in, is the PWM duty cycle, is the target brightness value, The maximum brightness value; transmit the control signal to the display driver chip through the I²C bus; monitor the actual output brightness value of the backlight in real time ; When detected When the abnormal fuse mechanism is activated, is the brightness tolerance threshold.

[0021] According to the above-mentioned embodiments of the present invention, there are at least the following beneficial effects: 1. By collecting the ambient light brightness data sequence and performing multi-step data preprocessing, including sensor validity verification, noise filtering, data normalization, and downsampling according to a preset sampling granularity, the accuracy and stability of the ambient light brightness data are effectively improved, and the problem of inaccurate brightness adjustment caused by sensor errors or noise interference is reduced, thereby improving the accuracy of brightness adjustment of the display device.

[0022] 2. By dynamically determining the brightness analysis window and calculating the ambient brightness value in combination with the target light brightness data sequence, and introducing the convolutional neural network prediction value as a weighting coefficient, a more comprehensive assessment of the current ambient light intensity can be achieved, avoiding misjudgments that may be caused by single sensor data, making the classification of ambient brightness labels more accurate, and thus optimizing the selection of brightness adjustment strategies.

[0023] 3. The brightness mutation prediction model is used to analyze the brightness change gradient characteristics and fluctuation variance characteristics, and combined with the abnormal control decision tree to make real-time control decisions. When brightness control abnormalities are detected, it can respond quickly and take measures such as brightness locking or activating safe brightness mode. This effectively prevents drastic fluctuations in the brightness of display devices caused by sudden changes in ambient light or system misjudgment, thereby improving user experience and device stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an illustrative and non-limiting manner, wherein: Figure 1 The figure is a flow chart of a method for adaptively adjusting ambient light brightness provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided solely to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. Rather, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0026] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software.

[0027] It should be noted that any number of elements in the drawings is for illustration only and not for limitation, and any naming is only for distinction and does not have any limiting meaning.

[0028] Reference below Figure 1 , Figure 1 Schematic diagram of the flow of the method for adaptively adjusting the ambient light brightness provided by one embodiment of the present invention. Figure 1As shown, a method for adaptively adjusting ambient light brightness includes: S1, collecting a current ambient light brightness data sequence; S2, performing data preprocessing on the ambient light brightness data sequence to obtain a preprocessed brightness data sequence; S3, determining a brightness analysis window based on the preprocessed brightness data sequence; S4, determining an ambient brightness label based on a target brightness data sequence, wherein the target brightness data sequence is preprocessed brightness data in the preprocessed brightness data sequence and located within the brightness analysis window, and the ambient brightness label represents the current ambient light intensity level; S5, in response to determining that the ambient brightness label is a high dynamic range label, generating a negative brightness adjustment strategy for a display device based on the current brightness strategy corresponding to the electronic device and the ambient brightness label, The updated brightness strategy is as follows: S6, in response to determining that the ambient brightness label is a low dynamic range label, generating a positive adjustment brightness strategy for the display device according to the current brightness strategy and the ambient brightness label as the updated brightness strategy; S7, performing brightness control on the display device according to the current brightness strategy or the updated brightness strategy, including: S8, determining abnormal brightness information according to the current brightness strategy or the updated brightness strategy, as well as the ambient brightness label and the brightness mutation prediction model; S9, in response to determining that the abnormal brightness information represents that the current brightness strategy or the updated brightness strategy has a control abnormality, determining the control instruction corresponding to the abnormal brightness information in the abnormal control decision tree; S10, performing brightness lock on the display device or activating a safe brightness mode according to the control instruction.

[0029] It should be noted that the present invention provides a method for adaptively adjusting ambient light brightness, applicable to electronic devices equipped with a photosensor and a camera. A photosensor is a sensor that can sense ambient light intensity, while a camera can capture image information. The combination of these two sensors enables more comprehensive perception of ambient light changes. The method first collects a current ambient light brightness data sequence—that is, ambient light brightness data continuously collected over a certain period of time. This data is then preprocessed to remove noise and normalize the data, thereby obtaining a more accurate brightness data sequence. Next, a brightness analysis window is determined based on the preprocessed data. This window is used to analyze brightness changes within a specific time period. Based on the data within the brightness analysis window, an ambient brightness label is determined. This label represents the current ambient light intensity level, namely, high dynamic range (HDR) or low dynamic range (LDR). Based on the ambient brightness label and the current brightness policy, a brightness adjustment policy for the display device is generated, and the display device is brightness controlled accordingly. The method also involves determining abnormal brightness information and applying an abnormal control decision tree to ensure the accuracy and safety of brightness adjustment.

[0030] Specifically, collecting the current ambient light brightness data sequence involves continuously collecting ambient light brightness data over a certain time interval using a photosensitive sensor to form a data sequence. Data preprocessing includes performing sensor validity verification on each ambient light brightness data point to ensure data reliability; removing noise from the data through noise filtering to obtain more accurate brightness data; and then performing data normalization to bring the data into a standard range for ease of subsequent processing. The brightness analysis window is a time window used to analyze brightness changes within a specific time period. Determining the brightness analysis window involves initializing the window length, calculating the brightness mean and eigenvalues ​​of the data within the window, and dynamically adjusting the window length based on these values ​​until specific conditions are met. The ambient light label is calculated based on the data within the brightness analysis window and represents the current ambient light intensity level. The brightness adjustment policy is generated based on the ambient light label and the current brightness policy and is used to adjust the brightness of the display device. Abnormal brightness information refers to information determined when the current brightness policy or an updated brightness policy has a control anomaly. The abnormal control decision tree is a decision tree model used to determine control instructions. The appropriate control instruction is selected based on the abnormal brightness information. Brightness lock means locking the brightness of the display device in specific circumstances to prevent over-adjustment; safe brightness mode is a brightness adjustment mode activated in specific circumstances to ensure that the brightness of the display device is within a safe range.

[0031] Preferably, the noise filtering process in data preprocessing can utilize a low-pass filter, which can effectively remove high-frequency noise while preserving the main features of the brightness data. Data normalization can normalize the brightness data to a range of 0 to 1, facilitating subsequent processing. The initial length of the brightness analysis window can be set based on the device's usage scenario and user needs. For example, a shorter window length can be set in indoor environments, while a longer window length can be set in outdoor environments. The ambient brightness label can be determined by calculating the brightness mean and characteristic values ​​within the brightness analysis window. For example, if the brightness mean exceeds a preset high dynamic range threshold, a high dynamic range label is determined; if the brightness mean falls below a preset low dynamic range threshold, a low dynamic range label is determined. The brightness adjustment policy can be generated based on the ambient brightness label and the current brightness policy. For example, if the ambient brightness label is a high dynamic range label, a negative brightness adjustment policy is generated to reduce the brightness of the display device; if the ambient brightness label is a low dynamic range label, a positive brightness adjustment policy is generated to increase the brightness of the display device. Abnormal brightness information can be determined using a brightness mutation prediction model. This model can be trained based on a historical brightness control dataset, extracting brightness change gradient features and brightness fluctuation variance features, and then inputting them into a support vector machine classifier for training. The construction of the abnormal control decision tree can be configured based on parameters such as the brightness change rate and duration. For example, if the brightness change rate exceeds a safety threshold and the duration exceeds a preset time threshold, a control instruction is output to activate brightness lock. If the brightness change rate exceeds the safety threshold but the duration does not exceed the preset time threshold, a control instruction is output to activate safe brightness mode.

[0032] In some embodiments, data preprocessing is performed on an ambient light brightness data sequence to obtain a preprocessed brightness data sequence, including: for each ambient light brightness data in the ambient light brightness data sequence, executing the following processing steps: performing sensor validity verification on the ambient light brightness data; in response to determining that the ambient light brightness data passes the validity verification, performing noise filtering on the ambient light brightness data to obtain filtered brightness data; performing data normalization on the filtered brightness data to obtain normalized brightness data; determining a sampling granularity corresponding to the ambient light brightness data sequence, wherein the sampling granularity represents the frequency of collecting the ambient light brightness data; in response to determining that the sampling granularity is not a preset sampling granularity, and the sampling granularity is smaller than the preset sampling granularity, downsampling the normalized brightness data sequence at the preset sampling granularity to obtain a preprocessed brightness data sequence.

[0033] It's important to note that data preprocessing of the ambient light brightness data sequence is intended to improve data quality and reliability, thereby providing an accurate data foundation for subsequent brightness analysis and adjustment. Data preprocessing includes performing sensor validation on each ambient light brightness data point to ensure data accuracy; performing noise filtering on the validated data to remove any potential interference; and performing data normalization on the filtered data to facilitate subsequent unified processing and analysis. Furthermore, based on a comparison of the sampling granularity with the preset sampling granularity, the normalized data sequence is downsampled to adjust the data acquisition frequency to meet the preset sampling standard. These steps collectively ensure the accuracy and consistency of the ambient light brightness data, providing reliable data support for subsequent brightness adjustment strategies.

[0034] Specifically, the ambient light brightness data sequence consists of ambient light brightness data continuously collected by a photosensor over a certain time interval. Sensor validity verification involves checking each piece of collected ambient light brightness data to determine its validity. This typically involves checking whether the data is within a reasonable range and whether the sensor is functioning properly. Noise filtering involves removing noise from the data using specific algorithms. Common methods include low-pass filtering and median filtering to improve data purity. Data normalization involves converting the data to a uniform range, typically between 0 and 1, to facilitate subsequent processing and comparison. Sampling granularity refers to the frequency at which ambient light brightness data is collected, i.e., the number of data points collected per unit time. The preset sampling granularity is a standard sampling frequency pre-set based on device performance and application scenarios. Downsampling involves processing the normalized data sequence according to the preset sampling granularity when the actual sampling granularity is less than the preset sampling granularity, reducing the data volume and improving processing efficiency. These steps ensure data quality and consistency, providing an accurate data foundation for subsequent brightness adjustments.

[0035] Preferably, sensor validity verification can be achieved by checking whether the output of the photosensor is within a preset reasonable range. For example, if the sensor output value exceeds its physical measurement range or deviates significantly from historical data, the data is considered invalid. Noise filtering can be performed using a low-pass filter, which can effectively remove high-frequency noise while preserving the key characteristics of the brightness data. Data normalization can be achieved by dividing each data point by its maximum value, thereby mapping all data points to a range of 0 to 1. When determining the sampling granularity, the preset sampling granularity can be set based on the device's usage scenario and performance requirements. For example, in an indoor environment, the preset sampling granularity can be set to 10 data points per second, while in an outdoor environment, due to the rapid changes in ambient light, the preset sampling granularity can be set to 20 data points per second. Downsampling can be achieved by selecting specific data points in the normalized data sequence. For example, if the actual sampling granularity is 20 data points per second and the preset sampling granularity is 10 data points per second, every other data point can be sampled to obtain a data sequence that meets the preset sampling granularity. These refined steps and parameter settings further improve the accuracy and efficiency of data preprocessing, providing high-quality data support for subsequent brightness adjustment.

[0036] In some embodiments, determining a brightness analysis window according to a preprocessed brightness data sequence includes: initializing a window length corresponding to the brightness analysis window to obtain an initial brightness analysis window, wherein the window length of the initial brightness analysis window is a reference window length; performing the following window determination steps according to the reference window length and the preprocessed brightness data sequence: determining a target brightness data sequence, wherein target brightness data in the target brightness data sequence is preprocessed brightness data located within the initial brightness analysis window; calculating a brightness mean according to the target brightness data sequence; and calculating a brightness mean according to the target brightness data sequence, the brightness mean, and the first target brightness analysis window. The brightness data and the second target brightness data are used to calculate the window characteristic value of the initial brightness analysis window, wherein the first target brightness data is the maximum value of the target brightness data sequence, and the second target brightness data is the minimum value of the target brightness data sequence; in response to determining that the brightness mean is greater than or equal to the dynamic threshold, the initial brightness analysis window is determined as the brightness analysis window, wherein the dynamic threshold is dynamically configured according to the working scene of the display device; in response to determining that the brightness mean is less than the dynamic threshold, the window length of the initial brightness analysis window is incremented to obtain the brightness analysis window with the incremented length as the initial brightness analysis window, and the window determination step is performed again.

[0037] It should be noted that the purpose of determining the luminance analysis window is to extract representative data segments from the preprocessed luminance data sequence for subsequent ambient luminance analysis. The determination of the luminance analysis window is a dynamic adjustment process, adjusting the window length based on the characteristics of the preprocessed luminance data sequence to ensure that the data within the window accurately reflects current ambient light variations. The initial luminance analysis window length is initialized based on a preset reference window length, which serves as the basis for initial analysis. By calculating the mean luminance value and window characteristic values ​​of the target luminance data sequence, it is possible to assess whether the data within the current window meets the analysis requirements. If the mean luminance value is greater than or equal to the dynamic threshold, the data within the current window is sufficiently representative and can be determined as a luminance analysis window. If it is less than the dynamic threshold, the window length is incrementally increased and reassessed until the conditions are met. The dynamic threshold is dynamically configured based on the operating scenario of the display device, meaning it can be adjusted to suit different usage environments and user needs to accommodate varying lighting conditions.

[0038] Specifically, the brightness analysis window is a time window used to analyze brightness changes within a specific time period. The initial brightness analysis window length refers to the initial size of this time window. It is a preset value typically determined based on device performance and application scenarios. The target brightness data sequence refers to the preprocessed brightness data within the initial brightness analysis window. This data is used to calculate the brightness mean and window characteristic value. The brightness mean is the average of all data points in the target brightness data sequence and reflects the overall brightness level within the window. The window characteristic value is calculated based on the maximum and minimum values ​​of the target brightness data sequence and is used to assess the range of data variation within the window. The dynamic threshold is a value dynamically configured based on the display device operating scenario. It can be adjusted according to different ambient lighting conditions to ensure that the brightness analysis window accurately reflects the current ambient light conditions. Incremental window length refers to gradually increasing the window length to include more data points when the brightness mean is less than the dynamic threshold until the brightness mean meets the required value. This process ensures that the data within the brightness analysis window accurately reflects ambient light changes, providing a reliable data foundation for subsequent brightness adjustments.

[0039] Preferably, the reference window length can be set based on the device's usage scenario and performance requirements. For example, in indoor environments, the reference window length can be set to a shorter time, such as 1 second. In outdoor environments, due to the rapid changes in ambient light, the reference window length can be set to an even shorter time, such as 0.5 seconds. The target brightness data sequence can be calculated using a simple averaging algorithm: summing all brightness data points within the window and dividing by the number of data points. The window characteristic value can be calculated as the difference between the maximum and minimum values: the window characteristic value equals the maximum brightness data point minus the minimum brightness data point. This value reflects the brightness range within the window. The dynamic threshold setting can be adjusted based on the device's display performance and user needs. For example, for high-resolution display devices, a higher dynamic threshold can be set to ensure accurate brightness adjustment in high dynamic range scenarios. When determining the brightness analysis window, a maximum window length limit can be set to prevent the window length from increasing indefinitely. For example, the maximum window length can be set to twice the reference window length. When the window length reaches the maximum limit, if the brightness mean value is still less than the dynamic threshold, an exception handling mechanism can be triggered, such as logging or user notification. These refined steps and parameter settings further improve the accuracy and adaptability of brightness analysis window determination, providing high-quality data support for subsequent brightness adjustment.

[0040] In some embodiments, determining the ambient brightness tag according to the target light brightness data sequence includes: for each target light brightness data in the target light brightness data sequence, calculating the ambient brightness value according to the following formula: ;in, is the calculated ambient brightness value, is the light sensor data, is the pixel position The weight coefficient of is the pixel grayscale value, is the predicted value of the convolutional neural network, is the weighting coefficient and .

[0041] Based on the calculated ambient brightness value sequence, a high dynamic range threshold and a low dynamic range threshold are determined; in response to the ambient brightness value continuously exceeding the high dynamic range threshold for a first time threshold, the ambient brightness label is determined to be a high dynamic range label; in response to the ambient brightness value continuously being lower than the low dynamic range threshold for a second time threshold, the ambient brightness label is determined to be a low dynamic range label.

[0042] It should be noted that determining the ambient brightness label is used to determine whether the current ambient light intensity falls within the high dynamic range (HDR) or low dynamic range (LDR) range based on the target brightness data sequence. This process involves performing a comprehensive calculation on each data point in the target brightness data sequence to obtain an ambient brightness value that represents the current ambient light intensity. This ambient brightness value is then used to determine the ambient brightness label, providing a basis for brightness adjustment of the display device. The ambient brightness value calculation combines photosensor data, pixel position weight coefficients, pixel grayscale values, and convolutional neural network predictions. These data are weighted and summed using weighting coefficients to obtain a more accurate ambient brightness value. The high dynamic range threshold and low dynamic range threshold are dynamically determined based on the ambient brightness value sequence and are used to determine the ambient light intensity level. If the ambient brightness value consistently exceeds the high dynamic range threshold or consistently falls below the low dynamic range threshold, the corresponding ambient brightness label is determined.

[0043] Specifically, the target light brightness data sequence is extracted from the preprocessed light brightness data sequence and is located within the brightness analysis window. The ambient brightness value is calculated by weighting each data point in the target light brightness data sequence. The photosensor data directly reflects the ambient light intensity, the pixel position weight coefficient indicates the importance of different pixel positions in the calculation, the pixel grayscale value reflects the brightness of each pixel in the image, and the convolutional neural network prediction value is based on the deep learning model's prediction of the ambient brightness. The weight coefficient is used to balance the contributions of these different data sources in the calculation, and their sum is 1 to ensure the rationality of the calculation result. The high dynamic range threshold and low dynamic range threshold are dynamically determined based on the ambient brightness value sequence to distinguish different light intensity levels. The first and second time thresholds are used to determine the duration that the ambient brightness value exceeds or falls below the threshold, ensuring the stability and reliability of the ambient brightness label. These parameters can be adjusted based on the specific usage scenario and device performance to adapt to different ambient lighting conditions.

[0044] Preferably, the calculation of the ambient brightness value can be achieved by the following steps: first, based on the data collected by the photosensor, the grayscale value of each pixel in the image captured by the camera, and the predicted value of the convolutional neural network, a weighted sum is performed in combination with the corresponding weight coefficients. For example, if the weight coefficient of the photosensor data is 0.4, the weight coefficient of the pixel grayscale value is 0.3, and the weight coefficient of the convolutional neural network predicted value is 0.3, then the ambient brightness value is the sum of these three data multiplied by the corresponding weight coefficients. The high dynamic range threshold and the low dynamic range threshold can be determined by analyzing the historical ambient brightness value data. For example, the 90% percentile in the historical data can be selected as the high dynamic range threshold, and the 10% percentile can be selected as the low dynamic range threshold. The first time threshold and the second time threshold can be set according to the actual application scenario. For example, in an indoor environment, it can be set to 3 seconds; in an outdoor environment, due to the rapid change of ambient light, it can be set to 1 second.

[0045] Furthermore, the convolutional neural network can include multiple convolutional and pooling layers to extract ambient light brightness features. This is then followed by classification or regression prediction via a fully connected layer to obtain a predicted ambient brightness value. These refined steps and parameter settings further improve the accuracy and adaptability of ambient brightness labeling, providing more reliable data support for brightness adjustment of display devices.

[0046] In some embodiments, the brightness mutation prediction model is constructed by the following steps: establishing a historical brightness control data set, including an ambient light brightness data sequence and a corresponding control strategy.

[0047] Extract brightness change gradient features: in, is the brightness change gradient at time t, is the brightness value at time t, is the brightness value at time t1, is the time interval.

[0048] Extract brightness fluctuation variance features: ;in, is the brightness fluctuation variance, is the sample size, is the ith brightness sample value, is the mean of brightness samples; the gradient features and variance features are input into the support vector machine classifier, and the brightness mutation prediction model is obtained through training.

[0049] The gradient features and variance features are input into the support vector machine classifier and trained to obtain the brightness mutation prediction model.

[0050] It's important to note that the brightness mutation prediction model is designed to predict sudden changes in ambient light brightness in advance, thereby providing a more accurate basis for display device brightness adjustment. This model analyzes ambient light data sequences and corresponding control strategies from a historical brightness control dataset to extract brightness gradient features and brightness fluctuation variance features. These features are then trained on a support vector machine classifier to produce a model capable of predicting sudden changes in brightness. The brightness gradient feature reflects the rate of change of brightness over time, while the brightness fluctuation variance feature reflects the stability of brightness changes. Using these two features, the model can identify sudden changes that could lead to abnormal display device brightness adjustment, allowing proactive action to improve the accuracy and stability of brightness adjustment.

[0051] Specifically, the historical brightness control data set is a set consisting of past ambient light brightness data sequences and their corresponding control strategies, and these data are used to train the brightness mutation prediction model. The brightness change gradient feature is obtained by calculating the brightness value difference between adjacent time points and dividing it by the time interval, which reflects the speed of brightness change over time. The brightness fluctuation variance feature is obtained by calculating the average value of the sum of the squares of the differences between the brightness sample value and the sample mean, which reflects the stability of the brightness change. The support vector machine classifier is a commonly used machine learning algorithm used for classification or regression prediction based on input feature data. In the present invention, it is used to predict brightness mutations based on the extracted brightness change gradient features and brightness fluctuation variance features. The parameter settings of these features and classifiers can be adjusted according to the specific usage scenarios and device performance to adapt to different ambient lighting conditions.

[0052] Preferably, the construction of the brightness mutation prediction model can be achieved by the following steps: first, a historical brightness control data set is established, which contains a sequence of ambient light brightness data collected over a period of time in the past and the corresponding control strategy. Then, the brightness change gradient feature is extracted. The specific method is to calculate the brightness value difference between adjacent time points and divide it by the time interval to obtain the rate of change of brightness over time. Next, the brightness fluctuation variance feature is extracted. The specific method is to calculate the average value of the sum of the squares of the differences between all brightness sample values ​​and the sample mean to obtain a stability index of the brightness change. Finally, the extracted brightness change gradient feature and brightness fluctuation variance feature are input into the support vector machine classifier for training to obtain a brightness mutation prediction model. During the training process, methods such as cross-validation can be used to optimize the parameters of the classifier, such as penalty parameters and kernel function parameters, to improve the prediction accuracy of the model. These refined steps and parameter settings further improve the accuracy and adaptability of the brightness mutation prediction model, providing more reliable support for brightness adjustment of display devices.

[0053] In some embodiments, the abnormal control decision tree includes: First level node: Determine whether the brightness change rate exceeds the safety threshold ,in They are respectively the maximum and minimum brightness values ​​allowed by the current environment.

[0054] Second level node: responds to the brightness change rate exceeding , to determine whether the time threshold is exceeded continuously .

[0055] Third-level node: In response to a duration exceeding , output the first control instruction to activate brightness lock.

[0056] Fourth-level node: Response duration does not exceed , output the second control instruction to activate the safe brightness mode.

[0057] It should be noted that the abnormal control decision tree is a decision model used to identify and handle abnormal brightness adjustment conditions on display devices. By setting a series of judgment conditions and corresponding control instructions, it can determine whether to lock the display device's brightness or activate safe brightness mode based on the current brightness change and ambient light data. The first-level node of this decision tree determines whether the brightness change rate exceeds a safety threshold. The brightness change rate here refers to the ratio of the current brightness to the maximum or minimum brightness value allowed by the environment. The safety threshold is a preset value used to determine whether the brightness change is within a safe range. If the brightness change rate exceeds the safety threshold, the decision tree proceeds to the second-level node to further determine whether the condition exceeding the safety threshold persists for a certain period of time. If the condition persists for more than the preset time threshold, the decision tree outputs the first control instruction, activating the brightness lock to prevent further abnormal brightness changes on the display device. If the condition persists for less than the preset time threshold, the decision tree outputs the second control instruction, activating safe brightness mode to ensure that the display device's brightness remains within a safe range.

[0058] Specifically, the abnormal control decision tree consists of multiple levels of nodes, each with specific judgment criteria and corresponding outputs. The first-level nodes determine whether the brightness change rate exceeds a safety threshold. The brightness change rate is calculated by comparing the current brightness with the maximum or minimum brightness allowed by the environment. The safety threshold is a pre-set ratio based on the characteristics of the display device and user requirements. If the brightness change rate exceeds the safety threshold, the second-level nodes determine whether this condition persists for a predetermined time threshold. The time threshold is a preset length of time used to determine the persistence of the abnormal brightness change. The first control instruction and the second control instruction are two different control instructions output by the decision tree based on the judgment results. The first control instruction activates brightness lock, which fixes the display device's brightness to prevent further abnormal changes. The second control instruction activates safe brightness mode, a backup brightness adjustment mode that provides a safe brightness range in the event of abnormal brightness changes. These parameters and instructions can be adjusted based on specific usage scenarios and device performance to ensure that the display device's brightness adjustment is both safe and effective.

[0059] Preferably, the construction of the abnormal control decision tree can take into account the following details: First, the safety threshold can be set based on the maximum and minimum brightness values ​​of the display device. For example, it can be set to 0.3 times the difference between the maximum and minimum brightness values. In this way, when the brightness change rate exceeds this ratio, it is considered that the brightness change may have an adverse impact on the display device or the user's visual experience. Second, the time threshold can be set based on the actual usage scenario. For example, in an indoor environment, it can be set to 3 seconds; in an outdoor environment, due to the rapid changes in ambient light, it can be set to 1 second. During the execution of the decision tree, when the brightness change rate exceeds the safety threshold, the system will start counting. If the brightness change rate continues to exceed the safety threshold within the preset time threshold, the brightness lock will be activated to prevent the display device's brightness from continuing to change abnormally. If the brightness change rate returns to normal within the time threshold, the safe brightness mode will be activated to ensure that the display device's brightness is within a safe range.

[0060] Furthermore, to further improve decision accuracy, more judgment conditions and corresponding control instructions can be added to the decision tree. For example, the activation conditions for brightness lock or safe brightness mode can be adjusted based on the changing trend of ambient light. These detailed steps and parameter settings can make the abnormal control decision tree more flexible and effective, better addressing various abnormal situations in display device brightness adjustment.

[0061] In some embodiments, the safe brightness mode includes: using an S-shaped brightness transition curve to adjust brightness: ;in, is the adjusted brightness value at time t, To adjust the starting brightness value, is the target brightness value, is the transition rate factor, To adjust the starting time point.

[0062] It should be noted that the safe brightness mode is a brightness adjustment mode activated under abnormal circumstances. It aims to ensure that the brightness of the display device is within a safe range and avoid discomfort to the user's vision or damage to the device due to sudden changes in brightness. This mode uses an S-shaped brightness transition curve for brightness adjustment. This curve can smoothly transition brightness values ​​and avoid sudden changes. The adjustment starting brightness value refers to the brightness when adjustment begins, the target brightness value refers to the brightness you want to achieve after adjustment, the transition rate factor controls the speed of brightness change, and the adjustment starting time point is the time when adjustment begins. Through this mode, the brightness of the display device can be adjusted in a more gentle and safe way when the brightness changes abnormally.

[0063] Specifically, the S-shaped brightness transition curve in the safe brightness mode is a nonlinear brightness adjustment method, which can ensure that the brightness changes are more natural and comfortable to the user's vision. The adjustment starting brightness value is the brightness value of the current display device, and the target brightness value is the ideal brightness value calculated based on the ambient light brightness and user preferences. The transition rate factor is a parameter used to control the speed at which the brightness changes from the starting value to the target value. This factor can be adjusted according to the performance of the device and the needs of the user. The adjustment starting time point is the time point when the brightness adjustment is started, usually when an abnormal change in brightness is detected. These parameters jointly determine the process and effect of brightness adjustment, ensuring that under abnormal circumstances, the brightness of the display device can smoothly transition to a safe range.

[0064] Preferably, the implementation of the safe brightness mode can be refined through the following steps: First, determine the starting brightness value for adjustment, which is usually the actual brightness of the current display device. Then, calculate the target brightness value based on the ambient light brightness and user preferences. This value can be obtained by combining the current ambient light brightness data with the brightness mutation prediction model. Next, select a suitable transition rate factor. This factor can be adjusted according to the response speed of the device and the user's sensitivity to brightness changes. For example, for devices with slower response speeds, a smaller transition rate factor can be set to ensure that the brightness changes do not happen too quickly. Finally, starting from the adjustment start time point, the brightness is adjusted according to the S-shaped brightness transition curve until the target brightness value is reached. During this process, the actual brightness of the display device can be monitored in real time to ensure the accuracy and safety of the adjustment process. This refined implementation method can ensure that the safe brightness mode adjusts the brightness of the display device in a smoother and safer manner when the brightness changes abnormally, thereby improving the user's visual experience and the stability of the device.

[0065] In some embodiments, a deep learning model update mechanism is also included: recording the user's manual brightness adjustment operation data; when the deviation between manual adjustment and automatic adjustment continues to exceed a preset deviation threshold, triggering model retraining: ;in, is the loss function value, is the sample size, is the automatically adjusted brightness value of the i-th sample, is the manually adjusted brightness value of the i-th sample, is the regularization coefficient, is the convolutional neural network weight parameter vector; the convolutional neural network weight parameters are updated through the back propagation algorithm .

[0066] It should be noted that the deep learning model update mechanism is designed to ensure that the adaptive ambient light brightness adjustment method can be dynamically adjusted based on actual user usage, thereby improving adjustment accuracy and user satisfaction. By recording the user's manual brightness adjustment operation data, when the deviation between manual adjustment and automatic adjustment continuously exceeds the preset deviation threshold, model retraining is triggered. This mechanism enables the system to learn user preferences and habits, thereby optimizing the brightness adjustment strategy. The loss function value is an indicator that measures the difference between the model's predicted value and the actual value. The regularization coefficient is used to prevent model overfitting. The convolutional neural network weight parameter vector is the core parameter of the model. These parameters are updated through the backpropagation algorithm to optimize the model's performance.

[0067] Specifically, the deep learning model update mechanism involves several key concepts. User manual brightness adjustment operation data refers to the data generated when users manually adjust the brightness of the display device according to their visual comfort. The preset deviation threshold is a set value used to determine whether the difference between manual adjustment and automatic adjustment is large enough to trigger model retraining. The loss function value is calculated by comparing the model's predicted brightness value with the brightness value actually manually adjusted by the user, and it reflects the accuracy of the model's prediction. The regularization coefficient is a parameter used to control the complexity of the model to prevent the model from overfitting to the training data. The convolutional neural network weight parameter vector is the set of parameters in the model used to process input data and generate prediction results. The backpropagation algorithm is an optimization algorithm used to adjust these parameters based on the loss function value. In this way, the model can learn a brightness adjustment strategy that is more in line with user habits.

[0068] Preferably, the deep learning model update mechanism can be implemented in detail through the following steps: First, continuously record user manual brightness adjustment operation data, including the brightness values ​​before and after the adjustment, as well as the adjustment timestamp. Then, calculate the difference between the manually adjusted brightness value and the automatically adjusted brightness value and compare it with a preset deviation threshold. If this difference consistently exceeds the preset deviation threshold, for example, if it exceeds the threshold for five consecutive adjustment operations, model retraining is triggered. During the retraining process, a loss function is constructed that takes into account the difference between the predicted brightness value and the actual manually adjusted brightness value for all samples, and a regularization term is added to control model complexity. The regularization coefficient can be adjusted based on the model training process; for example, if the model overfits, the regularization coefficient can be appropriately increased. Finally, the weight parameters of the convolutional neural network are updated using the backpropagation algorithm. This process can optimize the model performance through multiple iterations until the loss function value reaches a satisfactory level. In this way, the model can continuously learn and adapt to the user's brightness adjustment habits, thereby providing more personalized and accurate brightness adjustment services.

[0069] In some embodiments, model retraining includes: freezing the base feature layer using a transfer learning strategy; updating only the weight parameters of the fully connected layer; and accelerating convergence using a momentum optimizer. ;in, is the weight parameter at time t+1, is the weight parameter at time t, is the weight parameter at time t1 is the learning rate, is the gradient of the loss function, is the momentum coefficient.

[0070] It should be noted that the model retraining process adopts a transfer learning strategy, which is an efficient method in deep learning that can quickly adapt to new tasks or data using the existing model foundation. In the present invention, the transfer learning strategy freezes the basic feature layer and only updates the weight parameters of the fully connected layer, thereby quickly adjusting the model to adapt to new brightness adjustment requirements while maintaining the model's ability to extract basic features. The momentum optimizer is used to accelerate the convergence process of the model and improve training efficiency. The momentum optimizer introduces a momentum term in the gradient descent process to help the model converge to the optimal solution faster while avoiding falling into the local minimum.

[0071] Specifically, transfer learning is a method that leverages pre-trained models to continue training them on new tasks, saving training time and computing resources. Freezing the base feature layer means that during retraining, the underlying layers of the model, typically responsible for extracting common features, remain unchanged. This preserves the model's ability to extract common features like ambient light brightness. The fully connected layers are the higher-level layers of the model, responsible for mapping the extracted features to specific outputs, such as the brightness adjustment policy. Updating only the weight parameters of the fully connected layers means that only this layer is trained to adapt to new brightness adjustment requirements. The momentum optimizer is a modified gradient descent algorithm that accelerates convergence by introducing a momentum term. The learning rate controls the step size of each parameter update. The loss function gradient is the derivative of the loss function with respect to the weight parameters, indicating the direction of the parameter update. The momentum coefficient controls the influence of the momentum term. These parameters work together to enable the model to quickly adapt to new brightness adjustment requirements.

[0072] Preferably, the model retraining process can be further refined into the following steps: First, load a pre-trained model that has been trained on similar tasks and has good basic feature extraction capabilities. Then, freeze the basic feature layers of the model, which are responsible for extracting common features such as ambient light brightness, and keep their parameters unchanged. Next, only train the fully connected layers of the model, and update its weight parameters to adapt to the new brightness adjustment requirements. During the training process, use a momentum optimizer to accelerate convergence. The parameter settings of the momentum optimizer can include selecting an appropriate learning rate and momentum coefficient. For example, the learning rate can be set to 0.001 and the momentum coefficient can be set to 0.9. The settings of these parameters can be adjusted according to the training situation of the model to ensure that the model can converge quickly and stably. In this way, the model can adapt to new brightness adjustment requirements in a short period of time while maintaining the ability to extract common features such as ambient light brightness, thereby improving the accuracy and efficiency of brightness adjustment.

[0073] In some embodiments, brightness control execution includes: generating a PWM dimming control signal: ;in, is the PWM duty cycle, is the target brightness value, is the maximum brightness value.

[0074] The control signal is transmitted to the display driver chip via the I²C bus.

[0075] Real-time monitoring of the actual backlight output brightness value .

[0076] When detected When the abnormal fuse mechanism is activated, is the brightness tolerance threshold.

[0077] It should be noted that brightness control execution is a key step in the adaptive ambient light brightness adjustment method. It adjusts the display device's brightness by generating a PWM dimming control signal. PWM dimming is a technology that adjusts brightness by controlling the signal's duty cycle. The duty cycle is the ratio of the signal's high-level time to the total cycle time. The target brightness value is the ideal brightness value calculated based on the ambient light level and user preference, while the maximum brightness value is the maximum brightness the display device can achieve. The control signal is transmitted to the display driver chip via the I²C bus, a commonly used communication bus for transferring data between microcontrollers and peripheral devices. Real-time monitoring of the actual backlight output brightness value ensures the accuracy of brightness adjustment. When the deviation between the actual brightness and the target brightness exceeds the brightness tolerance threshold, the abnormal fuse mechanism is activated. This is a protection mechanism to prevent damage to the device caused by abnormal brightness adjustment.

[0078] Specifically, implementing brightness control involves several key concepts. A PWM dimming control signal adjusts brightness by controlling its duty cycle. The duty cycle is calculated as (target brightness / maximum brightness) × 100%, meaning the duty cycle is proportional to the target brightness. The target brightness is the ideal brightness value calculated based on ambient light levels and user preferences, while the maximum brightness is the maximum brightness the display device can achieve. The I²C bus is a simple and reliable communication bus for transmitting data between microcontrollers and peripheral devices. The actual backlight output brightness value refers to the brightness value actually achieved by the display device. Real-time monitoring of this value through a sensor ensures accurate brightness control. The brightness tolerance threshold is a preset value used to determine whether the deviation between the actual brightness and the target brightness is within an acceptable range. The abnormal fuse mechanism is a protection mechanism that activates when the deviation between the actual brightness and the target brightness exceeds the brightness tolerance threshold to prevent damage to the device caused by abnormal brightness control.

[0079] Preferably, the implementation of brightness control execution can be refined through the following steps: First, the duty cycle of the PWM dimming control signal is calculated based on the target brightness value and the maximum brightness value. For example, if the target brightness value is 50% of the maximum brightness value, the duty cycle is 50%. Then, the calculated PWM dimming control signal is transmitted to the display driver chip via the I²C bus to ensure accurate signal transmission. Next, the actual output brightness value of the backlight of the display device is monitored in real time, which can be achieved through a built-in brightness sensor. If the deviation between the monitored actual brightness and the target brightness exceeds a preset brightness tolerance threshold, for example, the deviation exceeds 5%, the abnormal fuse mechanism is activated. The abnormal fuse mechanism may include stopping the output of the PWM signal, or adjusting the brightness to a safe default value to prevent abnormal brightness adjustment from causing damage to the device. In this way, the brightness control execution can ensure that the brightness adjustment of the display device is both accurate and safe, improving the user's visual experience and the stability of the device.

[0080] The above-mentioned embodiments of the present invention have the following beneficial effects: 1. By collecting the ambient light brightness data sequence and performing multi-step data preprocessing, including sensor validity verification, noise filtering, data normalization, and downsampling according to a preset sampling granularity, the accuracy and stability of the ambient light brightness data are effectively improved, and the problem of inaccurate brightness adjustment caused by sensor errors or noise interference is reduced, thereby improving the accuracy of brightness adjustment of the display device.

[0081] 2. By dynamically determining the brightness analysis window and calculating the ambient brightness value in combination with the target light brightness data sequence, and introducing the convolutional neural network prediction value as a weighting coefficient, a more comprehensive assessment of the current ambient light intensity can be achieved, avoiding misjudgments that may be caused by single sensor data, making the classification of ambient brightness labels more accurate, and thus optimizing the selection of brightness adjustment strategies.

[0082] 3. The brightness mutation prediction model is used to analyze the brightness change gradient characteristics and fluctuation variance characteristics, and combined with the abnormal control decision tree to make real-time control decisions. When brightness control abnormalities are detected, it can respond quickly and take measures such as brightness locking or activating safe brightness mode. This effectively prevents drastic fluctuations in the brightness of display devices caused by sudden changes in ambient light or system misjudgment, thereby improving user experience and device stability.

[0083] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or a terminal device such as a computer, server, mobile phone, or tablet.

[0084] The above descriptions merely illustrate some preferred embodiments of the present invention and the underlying technical principles. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A method for adaptively adjusting ambient light brightness, characterized in that: include: Collect the current ambient light brightness data sequence; Performing data preprocessing on the ambient light brightness data sequence to obtain a preprocessed light brightness data sequence; Determine the brightness analysis window according to the pre-processed brightness data sequence; An ambient brightness label is determined based on a target brightness data sequence, wherein the target brightness data sequence is pre-processed brightness data in a pre-processed brightness data sequence and is located within a brightness analysis window, and the ambient brightness label represents the current ambient light intensity level; in response to determining that the ambient brightness label is a high dynamic range label, a negative adjustment brightness policy for the display device is generated as an updated brightness policy based on a current brightness policy and the ambient brightness label corresponding to the electronic device; in response to determining that the ambient brightness label is a low dynamic range label, a positive adjustment brightness policy for the display device is generated as an updated brightness policy based on the current brightness policy and the ambient brightness label; brightness control is performed on the display device based on the current brightness policy or the updated brightness policy, including: determining abnormal brightness information based on the current brightness policy or the updated brightness policy, the ambient brightness label, and a brightness mutation prediction model; in response to determining that the abnormal brightness information represents that the current brightness policy or the updated brightness policy has a control abnormality, determining a control instruction corresponding to the abnormal brightness information in an abnormal control decision tree; and according to the control instruction, brightness locking or activating a safe brightness mode for the display device.

2. The method for adaptively adjusting ambient light brightness according to claim 1, wherein: The method of performing data preprocessing on the ambient light brightness data sequence to obtain a preprocessed brightness data sequence includes: performing the following processing steps for each ambient light brightness data in the ambient light brightness data sequence: performing a sensor validity check on the ambient light brightness data; in response to determining that the ambient light brightness data passes the validity check, performing noise filtering on the ambient light brightness data to obtain filtered brightness data; performing data normalization on the filtered brightness data to obtain normalized brightness data; determining a sampling granularity corresponding to the ambient light brightness data sequence, wherein the sampling granularity represents a frequency of collecting the ambient light brightness data; and in response to determining that the sampling granularity is not a preset sampling granularity and is smaller than the preset sampling granularity, downsampling the normalized brightness data sequence at the preset sampling granularity to obtain the preprocessed brightness data sequence.

3. The method for adaptively adjusting ambient light brightness according to claim 2, wherein: The method of determining a brightness analysis window according to the preprocessed brightness data sequence includes: initializing a window length corresponding to the brightness analysis window to obtain an initial brightness analysis window, wherein the window length of the initial brightness analysis window is a reference window length; performing the following window determination steps according to the reference window length and the preprocessed brightness data sequence: determining a target brightness data sequence, wherein target brightness data in the target brightness data sequence is preprocessed brightness data located within the initial brightness analysis window; calculating a brightness mean according to the target brightness data sequence; and calculating a brightness mean according to the target brightness data sequence, the brightness mean, and the first target brightness. The first target light brightness data and the second target light brightness data are used to calculate the window characteristic value of the initial brightness analysis window, wherein the first target light brightness data is the maximum value of the target light brightness data sequence, and the second target light brightness data is the minimum value of the target light brightness data sequence; in response to determining that the light brightness mean is greater than or equal to the dynamic threshold, the initial brightness analysis window is determined as the brightness analysis window, wherein the dynamic threshold is dynamically configured according to the working scene of the display device; in response to determining that the light brightness mean is less than the dynamic threshold, the window length of the initial brightness analysis window is incremented to obtain the brightness analysis window with incremented length as the initial brightness analysis window, and the window determination step is performed again.

4. The method for adaptively adjusting ambient light brightness according to claim 3, wherein: Determining the ambient brightness label according to the target light brightness data sequence includes: for each target light brightness data in the target light brightness data sequence, calculating the ambient brightness value according to the following formula: ;in, is the calculated ambient brightness value, is the light sensor data, is the pixel position The weight coefficient of is the pixel grayscale value, is the predicted value of the convolutional neural network, is the weighting coefficient and ; Determine a high dynamic range threshold and a low dynamic range threshold based on the calculated ambient brightness value sequence; in response to the ambient brightness value continuously exceeding the high dynamic range threshold for a first time threshold, determine that the ambient brightness label is a high dynamic range label; in response to the ambient brightness value continuously being lower than the low dynamic range threshold for a second time threshold, determine that the ambient brightness label is a low dynamic range label.

5. The method for adaptively adjusting ambient light brightness according to claim 4, wherein: The brightness mutation prediction model is constructed by the following steps: establishing a historical brightness control data set, including an ambient light brightness data sequence and corresponding control strategies; extracting brightness change gradient features: in, is the brightness change gradient at time t, is the brightness value at time t, is the brightness value at time t1, is the time interval; extract the brightness fluctuation variance feature: ;in, is the brightness fluctuation variance, is the sample size, is the ith brightness sample value, is the mean of brightness samples; the gradient features and variance features are input into the support vector machine classifier, and the brightness mutation prediction model is obtained through training.

6. The method for adaptively adjusting ambient light brightness according to claim 1, wherein: The abnormal control decision tree includes: first level node: judging whether the brightness change rate exceeds the safety threshold ,in are the maximum and minimum brightness values ​​allowed by the current environment respectively; the second level node: responds to the brightness change rate exceeding , to determine whether the time threshold is exceeded continuously ; Third level node: In response to the duration exceeding , output the first control instruction to activate the brightness lock; the fourth level node: in response to the duration not exceeding , output the second control instruction to activate the safe brightness mode.

7. The method for adaptively adjusting ambient light brightness according to claim 6, characterized in that: The safe brightness mode includes: using an S-shaped brightness transition curve to adjust the brightness, as shown in the following formula: ;in, is the adjusted brightness value at time t, To adjust the starting brightness value, is the target brightness value, is the transition rate factor, To adjust the starting time point.

8. The method for adaptively adjusting ambient light brightness according to claim 1, wherein: The method also includes a deep learning model update mechanism that records user manual brightness adjustment operation data; when the deviation between manual adjustment and automatic adjustment continuously exceeds a preset deviation threshold, triggers model retraining: ;in, is the loss function value, is the sample size, is the automatically adjusted brightness value of the i-th sample, is the manually adjusted brightness value of the i-th sample, is the regularization coefficient, is the convolutional neural network weight parameter vector; the convolutional neural network weight parameters are updated through the back propagation algorithm .

9. The method for adaptively adjusting ambient light brightness according to claim 8, characterized in that: The model retraining includes: freezing the basic feature layer using a transfer learning strategy; updating only the weight parameters of the fully connected layer; and using a momentum optimizer to accelerate convergence: ;in, is the weight parameter at time t+1, is the weight parameter at time t, is the weight parameter at time t1 is the learning rate, is the gradient of the loss function, is the momentum coefficient.

10. The method for adaptively adjusting ambient light brightness according to claim 1, characterized in that: The brightness control execution includes: generating a PWM dimming control signal: ;in, is the PWM duty cycle, is the target brightness value, The maximum brightness value; transmit the control signal to the display driver chip through the I²C bus; monitor the actual output brightness value of the backlight in real time ; When detected When the abnormal fuse mechanism is activated, is the brightness tolerance threshold.

Images