A method for adjusting brightness of an LED display screen based on ambient light

Abstract

The present application relates to the technical field of display driving and brightness control of LED display screen, in particular to a brightness adjusting method of LED display screen based on ambient light. In the observation segment with unchanged image data, a current amplitude variation sequence is constructed, and a compensation current is programmed according to the relationship between current and light output, so that the cumulative light emission quantity is consistent with the target light emission quantity corresponding to the original driving current; the light sensing data is synchronously collected, the observation relationship composed of the ambient light variation term and the screen light emission contribution is established, the screen light coupling response is solved, and the sampling data of the repeated current state is used for checking; after the checking, the screen light emission contribution is deducted, the obtained ambient light is converted into target brightness and the constant current driving current is adjusted, and if the checking fails, the original driving current is maintained. The present application reduces the continuous dimming caused by the false judgment of the screen self light emission as the ambient light variation, and restricts the cumulative light quantity deviation generated in the identification process.

Description

Technical Field

[0001] This invention relates to the field of LED display driving and brightness control technology, specifically to a method for adjusting the brightness of an LED display based on ambient light. Background Technology

[0002] LED displays present text, graphics, and video content by controlling the luminous state of their pixels, and are commonly used in outdoor advertising and indoor information display. The illuminance around the display changes with natural and artificial lighting conditions, and the brightness adjustment system adjusts the overall luminous intensity of the screen accordingly. Existing systems typically place ambient light sensors around the screen to collect light signals, convert them into illuminance data, which is then read and filtered by the display controller. The target brightness is then determined based on a pre-established correspondence between illuminance and display brightness. This correspondence can be represented by a lookup table or a piecewise curve. The controller selects a brightness level based on the current illuminance and converts the target brightness into drive control commands, causing the screen to increase its luminous output when the ambient illuminance increases and decrease its luminous output when the ambient illuminance decreases.

[0003] In display systems employing constant current amplitude regulation, the display controller converts the target brightness into a current setpoint based on the brightness-drive current relationship calibrated by the screen, and then writes this setpoint into the current gain register of the constant current driver via a communication interface. The driver adjusts the conduction current of the output channel according to the written gain, and image data is processed through the existing display data path to form the grayscale and emission timing of each pixel. Under the condition of displaying the same image, changes in current amplitude correspond to changes in pixel emission intensity, thereby altering the overall brightness of the screen. Some systems also estimate the screen's contribution to the photosensor reading based on the display brightness and image grayscale, subtract this contribution using a pre-calibrated optical correspondence, and then use the processed reading for brightness calculation. This process is typically executed cyclically in the order of photosensor data reading, target brightness determination, and drive current update.

[0004] Light emitted from the screen itself, after being reflected by surrounding objects, may enter the ambient light sensor along with external ambient light, causing the light sensor reading to be affected by both external illumination and the screen's luminous state. When the displayed content or drive current changes, even if the external illumination remains constant, the sensor reading may change accordingly. If the display controller interprets this change as a change in ambient illuminance, it will continue to adjust the drive current, further altering the screen light contribution received by the sensor. When relying on fixed calibration relationships for subtraction, changes in the screen's luminous distribution or ambient reflection conditions may render the original correspondence inapplicable. Therefore, under conditions of continuous display operation and drive current participating in brightness adjustment, there is a mutual influence between the light sensor input and the screen output, making it difficult to distinguish between real environmental changes and the screen's own dimming response from the mixed light sensor readings. This can easily lead to continuous dimming caused by incorrect ambient light estimation. Summary of the Invention

[0005] The purpose of this invention is to provide a method for adjusting the brightness of an LED display screen based on ambient light, which can solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for adjusting the brightness of an LED display screen based on ambient light includes: within an observation segment where the image data remains unchanged, a compensation current is programmed into a current amplitude change sequence according to the relationship between current and light output, so that the cumulative light emission of the sequence is equal to the target light emission of the observation segment under the original driving current, the sequence is executed and light-sensing data is collected synchronously.

[0008] Establish an observation relationship whereby the light-sensing data equals the sum of the ambient light variation term and the screen light emission contribution obtained by converting the corresponding luminous quantity in the sequence through the screen light coupling response. Solve for the ambient light variation term and the screen light coupling response, and verify the observation relationship using light-sensing data of the repetitive current state.

[0009] If the verification passes, the screen light emission contribution is subtracted from the light sensing data according to the screen light coupling response, the resulting ambient light is mapped to the target brightness, and the LED constant current drive current is adjusted accordingly. If the verification fails, the original drive current is maintained.

[0010] Preferably, the current amplitude change sequence is constructed separately for each color channel, changing the current amplitude of one color channel at a time while keeping the original current amplitude of the other color channels.

[0011] Different current states are arranged in a time-mirrored manner, so that the equal-length sampling intervals corresponding to the same current state are symmetrical about the center of the observation segment. The compensation current is calculated in reverse according to the color channel, and the compensation current state is inserted into the time-mirrored arrangement.

[0012] Preferably, the solution includes: determining the luminous amount of each color channel within the sampling interval based on the actual effective current and image grayscale, and converting it into average light output according to the sampling interval duration;

[0013] The mean value of the light-sensing data that is symmetrical about the center of the observation segment is taken to obtain the central observation value corresponding to the current state;

[0014] The difference between two central observations with different current amplitudes in only one color channel is divided by the difference in the average light output of the corresponding color channel to obtain the optical coupling response coefficient of the color channel.

[0015] The screen luminous contribution is obtained by adding the product of the average light output of each color channel and the corresponding optical coupling response coefficient. The screen luminous contribution is then subtracted from the light-sensing data, and the ambient light variation term that changes linearly with the sampling time is obtained based on the remaining data.

[0016] Preferably, complete time-symmetric sampling pairs are retained from the photosensitive data corresponding to the repetitive current state as verification samples, and the remaining time-symmetric sampling pairs are used to complete the solution;

[0017] Based on the sensor measurement error and the light output calibration error, determine the joint feasible value that simultaneously meets the requirements of the optical coupling response coefficient of the solution sample and the ambient light change term, and calculate the predicted light sensitivity range corresponding to each verification sample.

[0018] The verification is considered successful only if there exists a common set of joint feasible values ​​that simultaneously satisfy the measurement error constraints of the corresponding predicted light sensitivity range for all verification samples; otherwise, the verification is considered unsuccessful.

[0019] Preferably, the actual current state interval is divided according to the effective time of the current gain register, and the photosensitive data that crosses the interval boundary is eliminated;

[0020] The difference between the actual effective current and the cumulative luminous emission amount and the target luminous emission amount is allocated to the compensation current state that has not yet been executed.

[0021] Preferably, for color channels with non-zero total light output, the pixel-by-pixel light output is divided by the total light output to form a normalized spatial distribution, which is then associated with and stored in relation to the light coupling response coefficient.

[0022] When the normalized spatial distribution of the subsequent image is consistent with the saved normalized spatial distribution, the corresponding optical coupling response coefficient is called; when they are inconsistent, the corresponding color channel is re-identified.

[0023] When the total light output is zero, the screen light emission contribution of the corresponding color channel is recorded as zero, and the corresponding color channel is re-identified after the light emission is restored.

[0024] Preferably, when the verification fails, the observation segment is divided into sub-segments using the sampling positions corresponding to the repetitive current states that are adjacent in time as candidate boundary positions;

[0025] The ambient light variation term is calculated for each sub-segment, and the optical coupling response coefficient is constrained to be shared by all sub-segments.

[0026] When there is a common response coefficient that satisfies the measurement error constraints of each sub-segment, the common response coefficient is retained, and the current amplitude change sequence is reconstructed using the current in effect as the original driving current. The ambient light change term is obtained based on the newly acquired time-symmetric sampling pair and a verification is performed.

[0027] If the common response coefficient is not present, the screen optical coupling response is recalculated.

[0028] Preferably, after adjusting the LED constant current drive current, the screen light emission contribution before and after the adjustment is calculated based on the actual effective current and the corresponding image grayscale, and the value before the adjustment is subtracted from the value after adjustment to generate a predicted value of light sensitivity change.

[0029] To verify the change in light perception, an equation is established between the measured change in light perception minus the predicted value and the change in ambient light at the corresponding sampling time. The measurement error constraint of the equation is then combined with the measurement error constraint of the verification.

[0030] If no joint feasible value exists, stop adding current adjustment based on measured light sensitivity changes, and re-execute the current amplitude change sequence and verification with the currently effective current.

[0031] Preferably, the current states in the current amplitude change sequence other than the compensation current state are used as the identification current states.

[0032] Before each current state identification is executed, the cumulative luminous amount after the color channel pre-simulation is subtracted from the target luminous amount to obtain the remaining target luminous amount.

[0033] Based on the combination of compensation current and duration allowed to be used during the remaining display period after the pre-performance, a set of remaining achievable light emission quantities is generated.

[0034] If the remaining target luminescence amount belongs to the set, the identification current state is executed; otherwise, the identification current state that has not yet been executed is cancelled.

[0035] After cancellation, the light-sensing data lacking time-symmetric samples are removed, the remaining target light emission is recalculated based on the actual cumulative light emission, the corresponding compensation current is calculated, and the compensation current is executed during the remaining display period.

[0036] Preferably, when the image grayscale changes before and after the current adjustment and the total light output of each color channel is not zero, the proportion of pixel light output to total light output is compared according to the color channel.

[0037] When the proportional distribution is consistent, the screen luminous contribution is calculated based on the image grayscale before and after adjustment;

[0038] When the proportional distribution is inconsistent, the light sensitivity change verification is paused. Under the adjusted image grayscale, the current amplitude change sequence is executed on the color channel with the changed proportional distribution. The optical coupling response coefficient is recalculated. The screen luminous contribution saved before adjustment is subtracted from the recalculated adjusted screen luminous contribution to update the predicted value. Then the light sensitivity change verification is executed again.

[0039] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0040] 1. By executing a current amplitude variation sequence within observation segments where image data remains unchanged, the light-sensing data obtains a basis for the screen light response corresponding to the actual current state. Sampling intervals with the same current state are organized symmetrically over time. Within segments where ambient light changes linearly with sampling time, the symmetrical sampling mean corresponds to the ambient light component at the same central moment. Combining the current variation of a single color channel, the corresponding optical coupling response can be obtained from the differences in observed values, and the screen luminescence contribution can be subtracted from the mixed readings. The retained repeated state samples are used to verify the same set of response coefficients and ambient light variation terms; observations that fail verification are not included in subsequent current adjustments. The measured light-sensing change after dimming is also verified together with the predicted contribution generated by the current change. This incorporates the reading changes caused by self-dimming into the subtraction and verification process, reducing the likelihood of it being repeatedly used to generate new dimming commands.

[0041] 2. A compensation current determined according to the light output relationship is incorporated into the current identification sequence, so that the cumulative light emission within the observation segment is constrained by the original target light emission. The executed light emission is calculated based on the actual effective time of the current gain register, allowing the light emission difference caused by the instruction activation deviation to be transferred to the unexecuted compensation state, and eliminating sampled data that crosses the current state boundary. Light emission pre-simulation is performed for subsequent identification states, comparing the remaining target light emission with the set of achievable light emission within the remaining display period. If the compensation conditions are not met, the unexecuted identification state can be cancelled, and the compensation current is then determined based on the actual output light emission. This process links identification sampling with display light emission calculation, constraining the cumulative light emission change caused by current probing and execution time deviation while performing screen light response identification, and avoiding the use of sampled data lacking corresponding current state basis for response solving. Attached Figure Description

[0042] Figure 1 Flowchart for ambient light decoupling and constant current brightness adjustment;

[0043] Figure 2 A graph showing the relationship between cumulative light intensity of color channels and target resolution;

[0044] Figure 3 An exploded view of the environment and screen items for light-sensing mixed readings;

[0045] Figure 4 Flowchart for the execution of residual light quantity pre-simulation and compensation;

[0046] Figure 5 Flowchart for dimming response verification and optical coupling update;

[0047] Figure 6 This is a graph showing the relationship between mixed illuminance, screen contribution, and environmental changes before and after dimming. Detailed Implementation

[0048] Please refer to Figure 1 In one embodiment, a method for adjusting the brightness of an LED display screen based on ambient light is provided, comprising: within an observation segment where image data remains unchanged, incorporating a compensation current into a current amplitude change sequence based on the relationship between current and light output, such that the cumulative luminous amount of the sequence is equal to the target luminous amount of the observation segment under the original driving current; executing the sequence and synchronously acquiring light-sensing data; establishing an observation relationship in which the light-sensing data equals the sum of the ambient light change term and the screen luminous contribution obtained by converting the corresponding luminous amount of the sequence through the screen light coupling response; solving for the ambient light change term and the screen light coupling response; and verifying the observation relationship using the light-sensing data under repeated current states; if the verification passes, subtracting the screen luminous contribution from the light-sensing data according to the screen light coupling response, mapping the obtained ambient light to the target brightness, and adjusting the LED constant current driving current accordingly; if the verification fails, maintaining the original driving current.

[0049] This embodiment is used for a display screen that emits light directly from LED pixels. The display controller receives the image to be displayed and outputs pixel grayscale data. The constant current driver determines the conduction current based on the effective value in the current gain register. The ambient light sensor is located at an existing detection position that can receive external lighting. Its readings are allowed to include the component formed by the reflection of the screen light through the surrounding surface. The controller processes these mixed readings through a program. The display screen cabinet, lamp board and sensor mounting structure follow the existing configuration. The improvements in the implementation focus on the organization of the current state, the correlation between the sampled data and the display state and the separation of light-sensing changes.

[0050] Specifically, the controller uses images that have entered the display buffer and remain unchanged during the observation period as the observed images, and the current that has been effective before entering the observation period is called the original driving current. The observation segment consists of a continuous display time. The same buffered image can be displayed repeatedly within the segment, but the light sensing data corresponding to different images are not allowed to be directly mixed into the same recognition. When an image update request arrives, the controller will retain the request until the existing frame boundary after the light quantity of the current observation segment is cleared, and then re-form the observation segment according to the new image. For platforms that do not allow delayed frame replacement, the observation segment is limited to the period of the original image retention. If the available time period is insufficient, the corresponding sequence will not be started.

[0051] Before entering observation, the controller reads the currently active current gain record and obtains the light output calibration record corresponding to the image to be displayed. The calibration record is established by optical measurements of the screen under different current states and image grayscale. The recorded content includes the average light output of pixels or drive groups and the measurement error boundary. During operation, the corresponding record is queried according to the image grayscale. Piecewise interpolation is used for the grayscale between adjacent records. The interpolation result carries the uncertainty formed by the error of adjacent calibration points and the interpolation deviation. The value of the calibration data comes from the screen measurement and is not replaced by the assumed ideal linear current relationship.

[0052] The current amplitude change sequence selects executable current states near the original driving current. Candidate states are directly derived from the gain states that the existing driver allows to write. The controller calculates the light output of the candidate states in the current image based on the calibration relationship. It selects states with different light outputs that can be compensated for in the remaining display time to form an identification combination. The sequence includes identification states used to obtain light response differences, as well as compensation states calculated inversely based on the cumulative light emission. The compensation states are still executed by the constant current driver, and their current value and duration jointly participate in the light emission calculation. The image grayscale maintains the content corresponding to the observed image.

[0053] The cumulative luminous intensity is calculated based on the cumulative amount of luminous flux during the actual display time, and has the dimension of lumen-second. The target luminous intensity of the same observation segment is calculated based on the light output when the original driving current continuously displays the observation image. The controller calculates each color channel separately and then summarizes them into a screen luminous intensity record. The non-luminous periods during the scanning display are also included in the time record, and their corresponding luminous flux is included according to the actual state. The scanning duty cycle change and current amplitude change are recorded separately. The repetitive current state has the same image and corresponding scanning luminous conditions.

[0054] In one specific implementation, the controller divides the observation segment into integral sub-intervals according to the change boundaries of the actual current state and the scanning state. Within each sub-interval, it queries the luminous flux corresponding to the current, multiplies it by the duration, and accumulates it. The compensation current is obtained by searching for candidate state combinations. During the search, executable compensation states and available time periods are enumerated, and combinations that can satisfy the target light quantity are retained. If multiple combinations can be executed, the already arranged time-symmetric position is used first. If no executable combination exists, the original driving current is retained and the identification arrangement that has not yet started is canceled. The light quantity constraint adopts the following relationship.

[0055] ;

[0056] Wherein, the color channel index c represents the emitted color being calculated, and the target emitted amount This represents the amount of color channel light that the original driving current should produce within the same observation segment, where N represents the total number of integration sub-intervals, and j represents one of the sub-intervals. This indicates the actual active color channel current within this sub-interval. This represents the calibrated luminous flux obtained by combining the current image and scanning status. The term represents the duration of a sub-interval, and the summation symbol indicates accumulation along the entire observation segment. The current is multiplied by time after entering the luminous flux function, so the calculation uses the product of the calibrated light output and the actual duration.

[0057] Please refer to Figure 2 , Figure 2 This corresponds to the aforementioned cumulative light quantity relationship of the color channels. The horizontal axis summarizes the same observation segment as start, first segment, middle segment, last segment, and end, with the actual time range being 0 to 3.2 seconds; the vertical axis represents the cumulative light quantity, in lumen-seconds. In the legend, the red, green, and blue cumulative values ​​track the execution ledgers of the three color channels, with the target line for each channel corresponding to a fixed target light quantity at the beginning of the segment. The increment of each segment of the curve is determined by the calibrated luminous flux under that state. Compared with actual duration The changes in the slope of the broken line reflect the switching of the light emission state, and the average value of the driving current cannot be directly taken as the average light quantity.

[0058] In this set of engineering operation records, all three channels accumulated from zero to 9.6 lumens per second and coincided with their respective targets at the end. At the 1.2-second mark, the red and blue channels both accumulated 3.6 lumens per second, and the green channel accumulated 4.0 lumens per second, demonstrating that the intermediate paths could differ while the endpoint could still be cleared. The target line maintained 9.6 lumens per second throughout the segment, used only for comparing endpoint differences and not indicating that this light intensity must be reached at all times. To identify locally overlapping trajectories, Figure 2 The inner lines have been slightly visually misaligned, but the original values ​​have not changed; the line segments only connect the boundaries of the saved ledgers.

[0059] To illustrate the light quantity calculation, we take a set of calibration data with available states of 2 lumens, 4 lumens, and 3 lumens, where 3 lumens corresponds to the original driving current. If the states with 2 lumens and 4 lumens occupy the same display period, then the cumulative light quantity of these two periods is consistent with the cumulative light quantity of the original state within the same duration. Here, 2 lumens and 4 lumens are optical calibration results and do not require equal increases or decreases in the driving current. If another set of candidate states has 1 lumen and 4 lumens, then the same target light quantity can be formed by arranging different cumulative durations. Whether this combination can be used depends on the actual available duration in the existing display timing sequence.

[0060] Actual devices have optical output measurement errors. The controller uses the light quantity equation as a combination constraint in the calibration model, and saves the error range involved in the calculation. The error range of the candidate state is used for subsequent observation and verification. The original state, the identified state and the compensated state use the same set of calibration data and units. The peak luminous flux needs to be converted into light quantity in combination with the actual emission period, and the average luminous flux needs to be converted into light quantity in combination with the corresponding average interval. The calculation range of all states covers the same observation segment. Therefore, different states can be compared on the same time basis, and the source of each light quantity can be traced.

[0061] The acquisition of light-sensing data uses the integration start and end information provided by the sensor or the synchronization marker corresponding to its integration period. The controller associates each reading with the display data actually covered by that integration period. If the sensor output is an integration count, it is converted into the average illuminance of the corresponding interval according to the calibration. If the average illuminance has already been output, the original dimensions are retained and the integration interval is recorded. Even if the readings in the communication buffer arrive late, they are assigned to the observation segment according to their actual integration interval. Readings whose integration assignment cannot be determined do not participate in the response solution. The data arrival time is only used for communication management and does not replace the optical sampling time.

[0062] The screen light coupling response represents the conversion relationship formed by the light sensor reading after the current screen light emission is reflected by the field and the spectral response of the sensor. The observation relationship incorporates the changes in external ambient light and the known light emission state of the screen into the calculation. The controller determines this conversion relationship based on the light output differences caused by multiple known current states, and then subtracts the screen contribution from the mixed reading. The subtracted ambient light enters the existing illuminance and target brightness mapping relationship. The target brightness is converted into the driving current through the brightness and current calibration curve of the screen. When the target exceeds the executable current state, the corresponding boundary is selected in the available state and the unmet brightness difference is recorded.

[0063] The observation verification and brightness command generation are stored separately. The current ambient light estimation updates the basis for the brightness command only when the input conditions for solving are met and the repeated state verification passes. If the verification fails, the original driving current before entering the segment is maintained. The already executed identification state is compensated according to the light quantity record and then exits. The controller attributes the light perception change caused by the compensation current to the screen light emission contribution. Subsequent observations will obtain valid data under the current display conditions again. The meaning of the original driving current always corresponds to the current that has taken effect at the beginning of each observation, rather than the target current that has not yet been executed.

[0064] In a preferred embodiment, the current amplitude change sequence is constructed according to color channels, and the current amplitude of one color channel is changed each time, while the current amplitude of the other color channels remains unchanged; different current states are arranged in time mirror, so that the equal-length sampling intervals corresponding to the same current state are symmetrical about the center of the observation segment, and the compensation current is calculated in reverse according to color channel, and the compensation current state is inserted into the time mirror arrangement.

[0065] In this embodiment, the color channels can correspond to red, green, and blue light emission channels. The monochrome display screen performs the same processing according to the actual existing color channels. The controller takes the state of all color channels in the original driving current as the common reference state, and pairs the states of changing only the red current, changing only the green current, and changing only the blue current with the common reference state respectively. When a color channel is changed, the image grayscale, current, and scanning coverage of the other channels remain consistent. The light output difference in the paired data is provided by the changed channel, and the response solution of each channel thus has the same comparison basis.

[0066] The time-mirrored arrangement uses the center moment of the observation segment as the center of symmetry. Each sampling interval before the center is arranged with corresponding sampling intervals with the same current state and integral duration after the center. The light-sensing data are paired according to the complete sampling interval, rather than according to the communication reception order. The compensation state can occupy the available time period near the center, or it can be inserted in pairs on both sides of the center. After insertion, the position of all sampling intervals relative to the center is recalculated. The symmetry center of the sequence that has already started is not arbitrarily changed due to the arrival of new data.

[0067] Before forming a mirror state, the controller reads the existing scan plan and compares the row gating signal and grayscale emission segment covered by the paired sampling intervals. Only when the scan coverage is the same, or when the average light output can be converted from the complete scan integration, are these two intervals classified as observation pairs of the same current state. If the sensor integration interval spans different scan stages, the emission amount of the corresponding pixel is accumulated segment by segment and then divided by the duration of the entire integration interval. The original image remains unchanged and does not directly replace the scan coverage consistency check.

[0068] For light-sensing data with limited precision, candidate current states need to provide distinguishable light output differences outside the calibration error range. The controller compares the light output ranges of candidate states. States with overlapping ranges and whose output differences cannot be confirmed do not participate in the differential solution. Different states participating in the solution must retain complete mirror sample pairs. At the same time, mirror sample pairs of repeated states are additionally retained for subsequent verification. Color channels with zero light output do not perform division operations, nor are new response coefficients derived based on zero light output. Their luminous contribution is directly recorded as zero.

[0069] Preferably, the solution includes: determining the luminous amount of each color channel within the sampling interval based on the actual effective current and image grayscale, and converting it into average light output according to the sampling interval duration; taking the mean of the light-sensing data symmetrical about the center of the observation segment to obtain the central observation value corresponding to the current state; dividing the difference between two central observation values ​​with different current amplitudes for only one color channel by the difference in the average light output of the corresponding color channel to obtain the light coupling response coefficient of the color channel; adding the product of the average light output of each color channel and the corresponding light coupling response coefficient to obtain the screen luminous contribution; subtracting the screen luminous contribution from the light-sensing data; and obtaining the ambient light variation term that changes linearly with the sampling time based on the remaining data.

[0070] The above average light output is based on the sensor sampling interval. After the controller obtains the cumulative light emission of each color channel within the interval, it uses the total duration of the interval as the divisor. The result represents the average luminous flux of that channel within the interval. The light output formed in the compensation state is also included in the corresponding sampling interval. If an integration interval covers multiple indistinguishable current states, the reading is removed in the state difference identification. The existing light emission records are still retained in the cumulative light emission calculation. Whether the reading can be used for the solution and whether the already output light emission needs to be compensated are determined separately.

[0071] Within a specific observation segment, the external ambient light is described by local linear changes over time. The controller takes the center of the segment as the time origin and divides the time difference between the actual sampling center time and the segment center by the total duration of the segment to obtain the normalized time coordinate. The time coordinate is only used to eliminate time scale differences in the calculation and does not change the actual integration interval. The time coordinate values ​​of the mirror interval are opposite. The average value of the linear ambient light in these two intervals is equal to the ambient light at the center of the segment. Therefore, the center observation value can be used to compare the screen luminescence contribution of different current states.

[0072] The optical coupling response of each color channel is shared within this segment, representing the contribution of the average light output of a unit color channel to the sensor illuminance reading under the same image and the same on-site optical path. The response coefficient simultaneously includes the propagation path and the sensor's response to the corresponding color light. If the spectral or spatial distribution changes caused by the candidate current change cannot be explained by the same set of response coefficients, it will be manifested as observation constraint incompatibility in the repeated state check. Based on this, the controller will exit the current state combination and reselect a candidate state that can be jointly explained, without taking the applicability of the response coefficient to all currents and all images as a premise.

[0073] When using average illuminance as light-sensing data, the specific observation relationship is as follows. The error term in the formula reflects both the sampling measurement deviation and the residuals not explained by the current local relationship. The calibration error is treated separately as an uncertainty of the light output input when solving the problem.

[0074] ;

[0075] Where m represents the index of the light-sensing sampling record, This represents the average illuminance of the record. This indicates the ambient illuminance at the center of the observed segment. This represents the rate of change of ambient illuminance relative to the normalized time coordinate. This represents the normalized time coordinate of the center of the corresponding sampling interval. This represents the set of actual emission color channels. Represents the optical coupling response coefficient of the color channel. This represents the average luminous flux of the color channel within the sampling interval. This represents the observation residual recorded in the unit of illuminance. The plus sign indicates the superposition of different light sources, and the product represents the illuminance contribution formed by the output of the color channel light through the optical path.

[0076] Please refer to Figure 3 , Figure 3 Used to interpret the effect of various factors in the observation relationship on the light sensor readings. The Chinese stage markers on the horizontal axis are arranged according to the actual sampling center sequence, corresponding to a time of approximately 0.1 to 3.5 seconds; the vertical axis represents the average illuminance, in lux. The solid black line represents the mixed readings. The gray dashed line is formed by the environmental baseline value. and time change items The obtained environmental estimate shows that the gray dashed lines represent the responses of each color channel. With the average luminous flux of the interval The resulting screen contribution. The sum of the two is... The small difference corresponds to the observation residual Therefore, the three curves cannot be compared independently based on their height, but their summation relationship should be checked at the same sampling location.

[0077] The initial effective recorded mixed illuminance was 35.46 lux, with an environmental estimate of approximately 27.697 lux and a screen contribution of approximately 7.763 lux. The highest mixed reading in the latter part of the record was 41.16 lux, corresponding to approximately 32.070 and 9.091 lux respectively. At the end of the record, the environmental estimate rose to approximately 33.007 lux, the screen contribution decreased to approximately 5.175 lux, and the mixed reading was 38.17 lux. This indicates that the variation in mixed illuminance cannot be directly attributed to either the external environment or the screen.

[0078] Continuing with the calculation example, suppose the ambient illuminance at the center of a segment is 30 lux, the average light output of the red channel in the common reference state is 2 lumens, the red response coefficient is 3 lux per lumen, and the total contribution of the other color channels is 10 lux. Then the center observation value of the common reference state is 46 lux. If the average light output of the red channel is changed to 4 lumens while keeping the other channels unchanged, the center observation value is 52 lux. Subtracting the two center observation values ​​gives 6 lux, and subtracting the corresponding light output gives 2 lumens. Dividing the former by the latter gives the red response coefficient. The data in this example is used to illustrate the calculation relationship.

[0079] Assuming the mirror readings of the common reference state are 44 lux and 48 lux, and the mirror readings of the red-changing state are 51 lux and 53 lux, the corresponding means are still 46 lux and 52 lux, respectively. In this example, the normalized time of the common reference sampling center is -0.4 and +0.4, the red-changing state is -0.2 and +0.2, the environmental change rate is 5 lux per normalized time unit, and the observation residual is 0. As long as these two sampling intervals are symmetrical about the same segment center and the external illuminance satisfies the linear relationship, the time change components contained in each will cancel each other out when taking the mean. Other color channels obtain their respective response coefficients according to the same single-channel difference, and then the screen contribution of all channels is subtracted from each original reading. The remaining data is used to calculate the central environmental illuminance and change rate.

[0080] In actual solution processing, the controller arranges the constant terms, normalized time coordinates, and average light output of each sampling record into an observation matrix. It checks the number of independent constraints by eliminating variables column by column. When there are 3 color response coefficients and ambient light constant and variable terms, at least 5 independent constraints are required. If the amount of data is sufficient but the observation matrix cannot independently distinguish these unknowns, no solution result that can be used for dimming will be generated. The controller will supplement the missing single-channel change state or end the current observation. The obtained color response coefficients retain non-negative values, and negative numerical solutions are entered into the verification process as input incompatibility cases.

[0081] In one embodiment that works in conjunction with current sequence execution, the actual current state interval is divided by the effective time of the current gain register, and photosensitive data that crosses the interval boundary is eliminated; the luminous emission is accumulated according to the actual effective current and duration, and the difference in relative luminous emission is allocated to the compensation current state that has not yet been executed.

[0082] After the current gain command is written, the controller temporarily stores it as pending activation. The current active current record is only updated after the latch activation flag given by the existing drive control chain arrives. If the driver updates multiple channels through a unified latch signal, the corresponding channel group is recorded according to the latch event. If the activation events of different channel groups occur separately, their respective time boundaries are recorded. The integral interval covered by the light sensor reading is compared with these boundaries. Only when the integral interval completely falls into the same interpretable current state interval can it have a clear attribution for state differential.

[0083] For example, if a light-sensing integration begins in current state A and state B has already taken effect before the integration ends, then the reading simultaneously includes the light emission contributions of both current states A and B. The controller excludes this reading from the calculation of the central observation value and the repeated state check, while retaining the actual duration of each of A and B. The output light quantity is calculated separately according to the calibrated luminous flux. If the sequence was originally expected to end earlier but A actually lasts longer, the light quantity difference is recalculated from the actual record. Subsequent compensation states use the recalculated difference instead of continuing to execute the old compensation result obtained based on the predetermined effective time.

[0084] For drive states where light output transition still exists after latching takes effect, the controller uses drive response calibration records to calculate the amount of light emitted during the transition period and marks the corresponding light sensing integral as not participating in steady state identification. The amount of light during the transition period still falls within the target light amount calculation range. The start and end of the steady state sample are determined based on the valid drive state marking and existing response records. During operation, only these existing capability information are read, and the operating parameters of the display scanning hardware are not changed. Control platforms that cannot provide actual effective attribution do not start identification combinations that depend on this attribution.

[0085] The actual execution records are saved in a time-sequential appending manner. Each record is associated with the observation image identifier, the effective current status and its corresponding light output record. The target light quantity record is frozen at the start of the observation. The light quantity that has been executed is only added according to the actual effective facts. The compensation light quantity that has not yet been executed can be rearranged according to the execution deviation, but will not write back the already executed records. Sampling rejection only changes the eligibility of the reading to participate in the solution, and does not delete the actual emission record of the corresponding time period. This makes the time source consistent with the time source used for observation data screening and light quantity compensation.

[0086] Please refer to Figure 4In a further embodiment, the current states other than the compensation current state in the current amplitude change sequence are used as the identification current states. Before each identification current state is executed, the cumulative luminous amount after the color channel pre-play is subtracted from the target luminous amount to obtain the remaining target luminous amount. Based on the combination of compensation current and duration allowed in the remaining display period after the pre-play, a set of remaining achievable luminous amounts is generated. If the remaining target luminous amount belongs to the set, the identification current state is executed; otherwise, the identification current state that has not yet been executed is cancelled. After cancellation, the light-sensing data lacking time-symmetric samples is removed, the remaining target luminous amount is recalculated based on the actual cumulative luminous amount executed, the corresponding compensation current is calculated, and the compensation current is executed in the remaining display period.

[0087] The rehearsal uses the current actual execution record as the starting point, adds the expected light quantity corresponding to the identified state to be executed to the already executed light quantity to obtain the candidate cumulative value, and then deducts the candidate cumulative value from the target light quantity of the same segment. The remaining time period is calculated according to the time position that can still be allocated in the existing display plan. The current state that has been submitted and cannot be withdrawn and its mirror corresponding state are deducted first. Only the unoccupied time position participates in the compensation combination search. Therefore, whether a certain identified state is allowed to be executed depends on whether there is a complete subsequent compensation path after it is executed.

[0088] In a feasible combined search, the controller builds a reachable light quantity table forward from the end of the remaining display time period. When there is no display time at the very end, only zero light quantity is retained. When an available time position is added forward, the compensation current that can be used at that position is enumerated, and the light quantity formed by the corresponding light output and duration is added to the reachable light quantity already saved at the back end to generate a new reachable record. Each record saves the correspondence between the current current state and subsequent records. After reaching the current time position, the table is queried with the remaining target light quantity. If the query is successful, the complete compensation sequence is restored along the saved correspondence.

[0089] When multiple color channels share the current update time, the reachability record is saved according to the vector composed of the light quantity of each color channel. After the individual reachability check of each channel is completed, it is also necessary to confirm that the same set of time positions and driving states can simultaneously achieve the remaining targets of all channels. If the combinations found for each color channel occupy incompatible time positions, they are not considered as executable results. The controller continues to enumerate other combinations. The feasible compensation path corresponding to the current remaining target light quantity is saved in advance before adding a new identification state. When the new state is rejected, the controller exits along the still executable compensation path.

[0090] The reachable light quantity is determined by comparing the calibrated data with the fixed-point records or bounded numerical records formed by the actual duration. When using bounded numerical records, candidate combinations must simultaneously satisfy the constraints of light quantity calculation error and subsequent observation error. It is not sufficient to judge whether the remaining target is located between the minimum and maximum values. For example, if two remaining time positions can each generate 2 units or 4 units of light quantity, then the combination can generate 4 units, 6 units, and 8 units of light quantity. Although 5 units of light quantity is located between the upper and lower bounds, it still does not belong to the reachable set of the combination. In this case, the identification arrangement that results in 5 units of remaining target is rejected.

[0091] Before each identification state begins, the controller also reserves the corresponding time position for mirror sampling pairs that have started but not yet completed. When the actual effective deviation causes the original mirror position to become unusable, the relevant samples are excluded as incomplete sampling pairs, and the identification arrangements that have not yet started can be canceled. The compensation state continues to be completed according to the current light quantity record. If the remaining valid sampling pairs are insufficient to support the observation solution, even if the accumulated light quantity can be cleared, no new brightness command will be generated for this observation. The completion of light quantity compensation and the completion of ambient light identification are recorded as different termination conditions.

[0092] For example, after a certain identification state is rehearsed, if the current color channel still needs 6 units of light, and there are compensation paths consisting of 2 units and 4 units of light for the remaining time period, then the identification state can be retained. If the execution record shows that the previous light quantity is different from the original rehearsal, the controller will regenerate the remaining target and query the reachability table before the next state begins. If it is still reachable, the compensation states that have not yet been executed will be rearranged. If it is not reachable, the new identification state that has not yet started will be rejected and the previously retained exit path will be restored. The actual effective boundary used for the rehearsal must fall within the time range that the drive control chain can confirm.

[0093] In this embodiment, the committed driver updates and their possible response transition periods are all included in the rehearsal. The controller does not continuously submit irrevocable trial instructions before confirming that the previous instruction has taken effect. The remaining light quantity calculation is updated with the effective record. The observation segment is only allowed to start when the available time period is sufficient to complete the selected compensation path. External image update requests are also coordinated through the existing display cache boundary. The committed observation segment is completed first and then the image is switched. From the execution constraints, it is ensured that the light quantity record is consistent with the image record participating in the identification.

[0094] The comparison object of the light quantity combination is always the cumulative amount of the observed segment. Satisfying the light quantity constraint does not mean that the light intensity is the same at each instant. There are still current changes arranged for identification during the display period. The time symmetry arrangement, the existing refresh sequence and the actual light output record jointly limit the location of these changes. The system will only perform active identification when the display control platform can support the corresponding sampling and compensation. It cannot rely on modifying the already displayed image data to make up for the light quantity when the remaining time period is insufficient, nor can it include the compensation amount that has not been completed in the executed record.

[0095] In another preferred embodiment, complete time-symmetric sampling pairs are retained from the light-sensing data corresponding to the repetitive current state as verification samples, and the remaining time-symmetric sampling pairs are used to complete the solution. Based on the sensor measurement error and the light output calibration error, the joint feasible value that simultaneously meets the optical coupling response coefficient and the ambient light variation term of the solution sample is determined, and the predicted light-sensing interval corresponding to each verification sample is calculated. The verification is deemed to have passed only when there is the same set of joint feasible values ​​that make all verification samples simultaneously meet the measurement error constraints of the corresponding predicted light-sensing interval; otherwise, the verification is deemed to have failed.

[0096] In this embodiment, the data is divided into solution samples and verification samples. The unit of division is a complete time-symmetric sampling pair. Two data points in the same pair will not be included in data sets with different purposes. When the controller executes the sampling arrangement, it records the purpose of the sample. After receiving the reading, it is classified according to the original record. The solution sample retains the current state that can distinguish all unknowns. The verification sample includes independent repeated observations of the current state that have appeared. If the solution sample no longer provides sufficient independent constraints due to loss or boundary removal, the corresponding state is supplemented and re-checked. It is not possible to directly add the specified verification sample into the solution set and still claim that independent verification has been completed.

[0097] The sensor measurement error is a composite of its counting conversion error, calibration error, and repeated measurement error under applicable working conditions. The light output calibration error is obtained by querying the current pixel grayscale and current status. The errors of multiple pixels are calculated as the average light output error by accumulating the actual light emission duration within the sampling interval. Current statuses lacking error records are not included in this identification. Photosensitive readings with saturation or invalid conversion flags are directly excluded. The error range comes from the measurement records and cannot be arbitrarily expanded in order to make inconsistent data pass the verification.

[0098] The controller treats the color response coefficient to be determined, along with the ambient light constant and variable terms, as the same set of unknowns. Both the solution sample constraints and the verification sample constraints are applied to this set of unknowns. The predicted light perception range is calculated jointly by the joint feasible range of the unknowns and the corresponding input error. It is not possible to select unrelated response coefficients for each verification sample. If one sample requires an increase in a certain color response while another sample requires a decrease, and there is no value that satisfies both simultaneously, then even if each sample can fall into its separately calculated range, the current observation is still judged as failing the verification.

[0099] The joint feasible values ​​are represented by the following set of observation constraints. When establishing the solution set, the solution sample index is used. When performing the verification, the retained sample index is added to the same set, and the intersection is checked to see if it still exists.

[0100] ;

[0101] in, H represents all parameter combinations that simultaneously satisfy the observation constraints, and H represents the response coefficients of each color channel. The vector formed This represents the set of sampled record indices currently participating in the constraints. This represents the illuminance error boundary of the m-th light sensor record. This represents the error bound of the average light output of the color channel corresponding to the record. The other symbols follow the previous definition. The absolute value symbol represents the magnitude of the deviation between the predicted illuminance and the measured illuminance. The less than or equal to symbol indicates that the deviation is covered by the measurement error and the input error after optical coupling conversion. The non-negative constraints correspond to the physical values ​​of the ambient illuminance and the optical transmission contribution, respectively.

[0102] In the specific solution process, since the response coefficients take non-negative values, the controller expands the absolute value constraints into two linear inequalities: positive deviation and negative deviation. The constraints generated for each record are added row by row to the same constraint matrix, and the parameter combination that satisfies all constraints is obtained through linear feasibility solution. The linear feasibility calculation can be implemented by adding constraints one by one and adjusting the current feasible point. When encountering mutually contradictory constraints, the record with no solution is returned. At the same time, the sampling index that causes the conflict is saved. The ambient illuminance at both ends of the observation segment is also added to the non-negative constraint to prevent the result of the intermediate value being positive while the endpoint is extrapolated to be negative from entering the brightness mapping.

[0103] When the joint feasible set is not a single point, the controller calculates the complete parameter combinations when the central ambient illuminance reaches the feasible lower and upper bounds respectively, and then averages these two complete combinations item by item to form the working solution for screen contribution deduction. The feasible set of the linear inequality constraint is a convex set, and the average value of the complete combination is still within the feasible set. Other color responses and environmental change rates take values ​​together with the complete combination. The midpoints of the intervals of each unknown quantity cannot be arbitrarily spliced. If the interval to be calculated is unbounded, the observation constraint is insufficient and distinguishable current states are supplemented.

[0104] For example, in a set of records used to illustrate the verification, the predicted screen contribution is 16 lux, the predicted ambient illuminance is 30 lux, the sensor illuminance error limit is 1 lux, and the total error limit of each color light output error after conversion by the response coefficient is 1 lux. Then, the total allowable deviation limit for the predicted reading of 46 lux is 2 lux. The observed value of 47 lux meets the constraints of this set of values, while the observed value of 50 lux does not. Whether it is finally approved depends on whether other jointly feasible values ​​can simultaneously explain all the retained samples. A single reading exceeding the error range of a certain working solution is not directly equivalent to the entire feasible set being empty.

[0105] In a preferred embodiment, if the verification fails, the observation segment is divided into sub-segments using the sampling positions corresponding to the repetitive current states that are adjacent in time as candidate boundary positions; the ambient light variation term is calculated for each sub-segment, and the sub-segments are constrained to share the optical coupling response coefficient; if there is a common response coefficient that satisfies the measurement error constraint of each sub-segment, the common response coefficient is retained, and the current amplitude variation sequence is reconstructed using the current effective current as the original driving current; the ambient light variation term is calculated based on the newly acquired time-symmetric sampling pair, and the verification is performed; if there is no common response coefficient, the screen optical coupling response is recalculated.

[0106] In this embodiment, the candidate boundaries are derived from the time records of repetitive current states. The controller prioritizes checking the adjacent occurrence positions of the same current state and attempts to segment the data sequentially between these positions. Each sub-segment has an independent ambient light constant and variation term. The color light coupling response is shared across all sub-segments. Sub-segments must retain independent observation and verification samples that can be used to examine their ambient light variation term. Segments that cannot provide these constraints are not considered valid segments. By limiting the identifiable conditions of each sub-segment, the system avoids fitting each reading individually to obtain segmentation results that lack verification significance.

[0107] If the external illumination changes abruptly in the original observation segment, while the optical path from the screen to the sensor remains unchanged, the single linear ambient light term of the entire segment may not be able to explain all the readings. After segmentation, it may be explained by different ambient light terms and the same set of color responses. At this time, the controller only retains the color response records with common feasible values. The original segment does not directly generate new dimming instructions. After the accumulated light amount is cleared, the mirror observation is re-executed under the display conditions after the segmentation, and the ambient light change term of the new segment is used to complete the verification and brightness calculation.

[0108] If there is no shared color response across sub-segments, the original color response is marked as unusable under the current conditions, and a new combination of current states is used for re-identification. The new combination still satisfies the constraints of image invariance and cumulative light amount. If multiple candidate boundaries can satisfy the shared response condition, the boundary with fewer sub-segments and each sub-segment being independently observed and complete is given priority. Segmented calculation is only used to determine which historical responses can be retained. Mixed segments that did not pass the verification are not directly regarded as valid segments after being split. Subsequent brightness updates are still determined by the newly acquired complete observation segments.

[0109] In one embodiment regarding the reuse of response records, for color channels with non-zero total light output, the pixel-by-pixel light output is divided by the total light output to form a normalized spatial distribution, which is then associated with and stored in relation to the optical coupling response coefficient. When the normalized spatial distribution of a subsequent image matches the stored normalized spatial distribution, the corresponding optical coupling response coefficient is invoked; otherwise, the corresponding color channel is re-identified. When the total light output is zero, the screen luminous contribution of the corresponding color channel is recorded as zero, and the corresponding color channel is re-identified after luminous recovery.

[0110] Specifically, the controller arranges the calibration light output of each color channel according to the fixed address of the pixel on the screen, divides each pixel value by the total amount of that color channel, and obtains a normalized vector stored in the same address order. The vector is associated with the image and effective current record when the response coefficient is obtained. When comparing subsequent images, the same calibration version and pixel address order are used. Therefore, the comparison object is the relative distribution of light emission in space. In the case where the image file identifier is the same but the grayscale data has been modified, the response will not be used based on the identifier alone. The same total light output is not directly considered to be the same spatial distribution.

[0111] Under the condition that the optical path remains unchanged, the total response of the color channel is formed by the contribution of different light emission positions weighted according to the relative light output. If the light output of all pixels changes in the same proportion and the normalized spatial distribution remains consistent, the saved response can be applied to the changed total light output. For example, if the light output of two light emission positions changes from 2 lumens and 4 lumens to 3 lumens and 6 lumens, the normalized distribution remains consistent. If it changes to 1 lumen and 5 lumens, although the total amount is still the same as before at 6 lumens, the normalized distribution has changed, and the corresponding color channel needs to be re-identified.

[0112] Normalization operations use the same numerical representation rules. Zero terms are retained for pixels with zero calibration light output. No division is performed when the total light output of the color channel is zero, and its screen light emission contribution is treated as zero. The sensor's own dark reading and fixed bias have been deducted in the basic illuminance conversion. The sensor's dark reading cannot be used as the screen light contribution generated by the black screen. After the light emission is restored, the mirror state of the color channel is reorganized using a new image and the response is obtained. The old response saved during the black period does not automatically replace this solution.

[0113] For large image caches, the search tags generated based on normalized data can be compared first to exclude obviously inconsistent records. If the tags are consistent, the actual normalized data is still compared. The search tags are only used for query purposes and cannot be used as the sole basis for completely consistent spatial distribution. If the calculated spatial distribution cannot be confirmed to be consistent, re-identification is performed. If some color channels are consistent while the other color channels are changed, they are handled separately. The response of the reserved channel is still checked with the repeated state data. If the reserved response fails due to changes in the field reflection path, it is expanded to solve all luminous color channels again.

[0114] Please refer to Figure 5 In one embodiment connected to dimming execution, after adjusting the LED constant current drive current, the screen light emission contribution before and after adjustment is calculated based on the actual effective current and the corresponding image grayscale. The adjusted value is subtracted from the original value to generate a light perception change prediction value. Light perception change verification is performed by establishing an equation between the measured light perception change after deducting the prediction value and the change in ambient light at the corresponding sampling time, and combining the measurement error constraint of the equation with the verification measurement error constraint. If no jointly feasible value exists, the current adjustment based on the measured light perception change is stopped, and the current amplitude change sequence and verification are re-executed with the current effective current.

[0115] When the controller sends a brightness update command, it saves the complete display state and observation working solution before the update. After the new current actually takes effect, it calculates the screen contribution based on the updated image grayscale and the actual current. The light sensor readings that are compared before and after the update must have valid integral attribution. Readings that cross the effective boundary are excluded according to the aforementioned rules. The comparison time interval is determined by the actual center time of two valid samples. The ambient light change term is calculated by the difference between these two time positions. The sending interval of the communication command cannot be used to replace the actual sampling interval.

[0116] The screen contribution before and after the update is calculated by the average light output of each color channel and its applicable response coefficient. When the spatial distribution remains unchanged and the original response still passes the verification, the same response record is used before and after. When the spatial distribution changes and re-identification is completed, response records matching their respective images are used before and after. After deducting the screen contribution change from the measured illuminance change, the remaining amount that needs to be explained by the change in external ambient light is obtained. The calculation relationship is as follows.

[0117] ;

[0118] in, This represents the estimate of ambient illuminance change after deducting the screen contribution change. and These represent the average illuminance of the effective samples before and after adjustment. and These represent the average luminous flux of the corresponding color channel in these two sampling intervals. and These represent the response coefficients and color sets that match the images before and after adjustment, as well as the optical conditions. Following the aforementioned definition, the values ​​within the parentheses are calculated by subtracting the previous value from the subsequent value, and the outer minus sign indicates that the change contributed by the screen itself is deducted from the measured change.

[0119] Please refer to Figure 6 , Figure 6Two valid samples were taken before and after the focus dimming command took effect. The horizontal axis shows only the values ​​before and after adjustment, while the vertical axis represents the change in illuminance relative to the pre-adjustment baseline, in lux. The three lines respectively show the change in mixed readings, the change in screen contribution, and the change in environment after subtracting the screen. (The corresponding relationships are...) This is not the adjusted absolute ambient illuminance, but rather a mixture of the readings before and after adjustment. , The difference, minus the light output of the front and rear color channels. , According to their respective responses , The difference is calculated based on the screen contribution. After adjustment, the mixed reading decreased by 1.970 lux compared to before adjustment, and the screen contribution decreased by approximately 2.588 lux. Subtracting these two, we get an increase of approximately 0.618 lux in environmental change. Meanwhile, the original environmental trend was projected to increase by approximately 0.625 lux over the same interval, resulting in a difference of approximately 0.007 lux. The decrease in the total visible light sensor reading does not necessarily indicate a darker external environment; changes in the screen itself are the primary cause. All three lines were initially set at zero before adjustment, serving only as a benchmark for comparing the changes. Figure 6 To avoid overlap, minor image misalignment does not change the saved value, and connecting the two ends does not represent continuous sampling in the middle.

[0120] To illustrate the calculation, the calculation condition of 3 lux per lumen for red response and 10 lux for the other colors is used. The average red light output is changed from 2 lumens to 6 lumens. The mixed reading before adjustment is 46 lux and the screen contribution is 16 lux. The mixed reading after adjustment is 58 lux and the screen contribution is 28 lux. The measured reading increases by 12 lux, and the predicted screen contribution also increases by 12 lux. After deducting these, the resulting environmental change is zero. When the original ambient light change item also gives zero change and the error constraint is compatible, the brightness will not increase again due to this 12 lux self-luminous change. If the mixed reading after adjustment is 63 lux and the screen contribution is still 28 lux, the remaining environmental change is 5 lux. The controller continues to check whether this remaining amount is consistent with the current ambient light change item.

[0121] When the remaining 5 lux cannot be explained by the current ambient light variation term and its error range, the controller will add the new comparison relationship to the original joint constraint. If the joint feasible set is empty, the subsequent additional current adjustment will be paused. The current that is currently in effect will be used as the new original driving current for re-observation. The ambient illuminance will be re-observed and then the target brightness will be mapped. The newly added remaining amount will not be directly superimposed on the driving current. The old samples participating in the new relationship will still retain the original measurement error source. When the same original sample is reused, the same error record will be used to avoid treating the same data as multiple independent pieces of evidence to repeatedly narrow the feasible range.

[0122] In a further embodiment, when the image grayscale changes before and after current adjustment and the total light output of each color channel is not zero, the proportion of pixel light output to total light output is compared according to color channel; when the proportion distribution is consistent, the screen light emission contribution is calculated based on the image grayscale before and after adjustment; when the proportion distribution is inconsistent, the light sensitivity change verification is paused, and the current amplitude change sequence is executed on the color channel with the changed proportion distribution under the adjusted image grayscale, the optical coupling response coefficient is recalculated, the screen light emission contribution saved before adjustment is subtracted from the recalculated adjusted screen light emission contribution, the predicted value is updated, and the light sensitivity change verification is performed again.

[0123] In this embodiment, when the proportional distribution changes, the screen contribution and its source record before adjustment are retained. The new identification is performed on the fixed image after adjustment. The current of the identified color channel is changed, while the other channels maintain their current active state. A repeat state is arranged to check the retained response. The new response only affects the light output of the adjusted image and cannot be used to recalculate the screen contribution before adjustment. The controller subtracts the contribution before adjustment obtained from the old record from the contribution after adjustment obtained from the new record and recalculates the predicted light perception change, so that the two different light emission distributions use the corresponding optical conversion basis respectively.

[0124] For example, in a calculation scenario where the optical path remains unchanged, two emission positions of a certain color contribute 2 lux and 1 lux to the sensor per lumen, respectively. When the original image outputs 2 lux and 4 lux at these two positions, respectively, the screen contribution is 8 lux. When the new image outputs 1 lumen and 5 lux, respectively, the screen contribution is 7 lux. The two images have the same total light output, but the sensor's received contribution is different. After re-identification, the calculation is performed using the contributions from before and after, and this 1 lux change in reading is attributed to the change in image emission distribution. Whether there are still changes in the external environment is determined by the remaining observation relationship.

[0125] When a frame change request occurs during the new image identification process, the controller will process the request after the promised observation segment ends and the light compensation is completed. Intermediate images that have not been effectively solved will not be passed on to the next generation of new response coefficients. If a color channel recovers its light emission from zero output, the recovery identification branch will be executed. The proportional comparison branch that requires the total light output before and after to be non-zero will not be entered. The results of each color channel will still be subject to the same set of ambient light variation terms. A feasible solution obtained by a single channel will not replace the joint verification of the entire observation image.

[0126] In this embodiment, the controller continuously maintains the observed images, effective current and corresponding light output records, uses the light sensing data of the effective integration interval for response separation, excludes readings that cross states or lack mirror samples from the solution, and retains the actual light emission in the compensation calculation. The ambient light estimation forms the target brightness after joint verification. After the target current actually takes effect, the light sensing change verification is performed. When the image distribution changes, the applicable response is updated. When the external lighting changes abruptly, the ambient light change item is re-determined. Each record is transferred according to the corresponding observation segment and settled at the end of the segment.

Claims

1. A method for adjusting the brightness of an LED display screen based on ambient light, characterized in that, include: Within the observation segment where the image data remains unchanged, a compensation current is programmed into the current amplitude change sequence based on the relationship between current and light output, so that the cumulative light emission of the sequence is equal to the target light emission of the observation segment under the original driving current. The sequence is then executed and photosensitive data is collected synchronously. Establish an observation relationship whereby the light-sensing data equals the sum of the ambient light variation term and the screen light emission contribution obtained by converting the corresponding luminous quantity in the sequence through the screen light coupling response. Solve for the ambient light variation term and the screen light coupling response, and verify the observation relationship using light-sensing data of the repetitive current state. If the verification passes, the screen light emission contribution is subtracted from the light sensing data according to the screen light coupling response, the resulting ambient light is mapped to the target brightness, and the LED constant current drive current is adjusted accordingly. If the verification fails, the original drive current is maintained.

2. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 1, characterized in that, The current amplitude change sequence is constructed separately for each color channel. Each time, the current amplitude of one color channel is changed, while the current amplitude of the other color channels remains unchanged. Different current states are arranged in a time-mirrored manner, so that the equal-length sampling intervals corresponding to the same current state are symmetrical about the center of the observation segment. The compensation current is calculated in reverse according to the color channel, and the compensation current state is inserted into the time-mirrored arrangement.

3. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 2, characterized in that, The solution includes: determining the amount of light emitted by each color channel within the sampling interval based on the actual effective current and image grayscale, and converting it into average light output according to the sampling interval duration; The mean value of the light-sensing data that is symmetrical about the center of the observation segment is taken to obtain the central observation value corresponding to the current state; The difference between two central observations with different current amplitudes in only one color channel is divided by the difference in the average light output of the corresponding color channel to obtain the optical coupling response coefficient of the color channel. The screen luminous contribution is obtained by adding the product of the average light output of each color channel and the corresponding optical coupling response coefficient. The screen luminous contribution is then subtracted from the light-sensing data, and the ambient light variation term that changes linearly with the sampling time is obtained based on the remaining data.

4. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 1, characterized in that, Complete time-symmetric sampling pairs are retained from the photosensitive data corresponding to the repetitive current state as verification samples, and the remaining time-symmetric sampling pairs are used to complete the solution. Based on the sensor measurement error and the light output calibration error, determine the joint feasible value that simultaneously meets the requirements of the optical coupling response coefficient of the solution sample and the ambient light change term, and calculate the predicted light sensitivity range corresponding to each verification sample. The verification is considered successful only if there exists a common set of joint feasible values ​​that simultaneously satisfy the measurement error constraints of the corresponding predicted light sensitivity range for all verification samples; otherwise, the verification is considered unsuccessful.

5. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 3, characterized in that, The actual current state interval is divided according to the effective time of the current gain register, and the photosensitive data that crosses the interval boundary is eliminated; The difference between the actual effective current and the cumulative luminous emission amount and the target luminous emission amount is allocated to the compensation current state that has not yet been executed.

6. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 3, characterized in that, For color channels with non-zero total light output, divide the pixel-by-pixel light output by the total light output to form a normalized spatial distribution, and store it in association with the light coupling response coefficient. When the normalized spatial distribution of the subsequent image is consistent with the saved normalized spatial distribution, the corresponding optical coupling response coefficient is called; when they are inconsistent, the corresponding color channel is re-identified. When the total light output is zero, the screen light emission contribution of the corresponding color channel is recorded as zero, and the corresponding color channel is re-identified after the light emission is restored.

7. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 4, characterized in that, If the verification fails, the observation segment is divided into sub-segments using the sampling positions corresponding to the repetitive current states that are adjacent in time as candidate boundary positions. The ambient light variation term is calculated for each sub-segment, and the optical coupling response coefficient is constrained to be shared by all sub-segments. When there is a common response coefficient that satisfies the measurement error constraints of each sub-segment, the common response coefficient is retained, and the current amplitude change sequence is reconstructed using the current in effect as the original driving current. The ambient light change term is obtained based on the newly acquired time-symmetric sampling pair and a verification is performed. If the common response coefficient is not present, the screen optical coupling response is recalculated.

8. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 5, characterized in that, After adjusting the LED constant current drive current, the screen light emission contribution before and after the adjustment is calculated based on the actual effective current and the corresponding image grayscale. The adjusted value is then subtracted from the original value to generate a predicted value of light sensitivity change. To verify the change in light perception, an equation is established between the measured change in light perception minus the predicted value and the change in ambient light at the corresponding sampling time. The measurement error constraint of the equation is then combined with the measurement error constraint of the verification. If no joint feasible value exists, stop adding current adjustment based on measured light sensitivity changes, and re-execute the current amplitude change sequence and verification with the currently effective current.

9. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 5, characterized in that, The current states other than the compensation current state in the current amplitude change sequence are used as the identification current states. Before each current state identification is executed, the cumulative luminous amount after the color channel pre-simulation is subtracted from the target luminous amount to obtain the remaining target luminous amount. Based on the combination of compensation current and duration allowed to be used during the remaining display period after the pre-performance, a set of remaining achievable light emission quantities is generated. If the remaining target luminescence amount belongs to the set, the identification current state is executed; otherwise, the identification current state that has not yet been executed is cancelled. After cancellation, the light-sensing data lacking time-symmetric samples are removed, the remaining target light emission is recalculated based on the actual cumulative light emission, the corresponding compensation current is calculated, and the compensation current is executed during the remaining display period.

10. The method for adjusting the brightness of an LED display screen based on ambient light according to claim 8, characterized in that, When the image grayscale changes before and after current adjustment and the total light output of each color channel is not zero, the proportion of pixel light output to total light output is compared according to color channel. When the proportional distribution is consistent, the screen luminous contribution is calculated based on the image grayscale before and after adjustment; When the proportional distribution is inconsistent, the light sensitivity change verification is paused. Under the adjusted image grayscale, the current amplitude change sequence is executed on the color channel with the changed proportional distribution. The optical coupling response coefficient is recalculated. The screen luminous contribution saved before adjustment is subtracted from the recalculated adjusted screen luminous contribution to update the predicted value. Then the light sensitivity change verification is executed again.

Images