Method for dynamically adjusting light source power of intelligent projector

Abstract

The invention relates to the technical field of intelligent dimming, in particular to an intelligent projector light source power dynamic adjustment method, which comprises the following steps of: acquiring an image through an image sensor, constructing a pyramid, dividing grids, aggregating HSV-V values by adopting a sliding window, outputting a brightness variation rate, and comparing a dynamic threshold value to generate an adjustment priority; and calculating a compensation coefficient in combination with an environment illumination interpolation to generate a correction weight parameter, extracting a main peak region span fusion output brightness structure stability, performing weighted fusion after normalization to obtain a driving current proportion, performing scaling to generate an adjustment parameter, and sending and executing the adjustment parameter. According to the method, the light source power is regulated and controlled through multi-dimensional data fusion, brightness variation is extracted through an image pyramid and a sliding window, a dynamic threshold value is used for screening an adjusting area, linear interpolation is used for fusing environment illumination, a histogram is used for detecting a main brightness interval, abnormal fluctuation is restrained, driving current is normalized and balanced, the response speed is increased, and the energy efficiency is optimized; and the picture stability and the visual comfort are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent dimming technology, and in particular to a method for dynamically adjusting the light source power of an intelligent projector. Background Art

[0002] The field of intelligent dimming technology includes a technical system that adjusts the output of light sources in real time to adapt to changes in the external environment. The core content of this technology is to use sensors to sense external brightness, image content characteristics, or user settings, and dynamically adjust the luminous intensity or color temperature parameters of the light source to achieve optimized visual comfort and energy efficiency control. This field mainly covers technical links such as light sensing acquisition, electro-optical adjustment control, signal transmission and feedback mechanisms, and is widely used in equipment such as intelligent lighting equipment, projection display systems, electronic display screens, and in-vehicle displays. From a systematic perspective, the development of intelligent dimming technology has tended towards multi-dimensional collaborative directions such as multi-parameter linkage control, environmental adaptive adjustment, and human-computer interactive perception adjustment.

[0003] Among them, the method for dynamic adjustment of light source power of intelligent projectors refers to a technical solution for automatically adjusting the power output of the light source in real time according to parameters such as ambient brightness, brightness distribution of the projected image content, and changes in projection distance during the operation of the projection equipment. The technical matters addressed by this topic include how to achieve automatic matching of light source power through dynamic power modulation under differentiated external lighting conditions, and how to adjust the overall light source output level based on local brightness distribution information in the projected image. The method adopted is: continuously collecting external lighting data through ambient light sensors, using image analysis to extract the brightness structure characteristics of the current image, and calculating the adapted power output value based on a preset dynamic power modulation model, and then using a power modulation drive circuit to adjust the output power of the light source in real time.

[0004] Existing technologies rely on a single ambient light sensor to collect data, making it difficult to accurately distinguish the coupling effect between ambient light and the projected image's own brightness. This makes light source power regulation susceptible to sudden changes in ambient light. For example, when room lighting suddenly turns on, the sensor mistakenly interprets an overall increase in ambient brightness, triggering unnecessary power reductions and causing the image to be partially too dark. Existing image brightness analysis often uses a global average calculation, ignoring the uneven distribution of bright and dark areas within the image. This results in loss of detail in highlight areas and wasted energy in dark areas. The power regulation model lacks a dynamic priority mechanism, making it unable to adjust the adjustment order in real time based on changes in image content. This can lead to adjustment lag or overcompensation when switching between complex scenes. The ambient light compensation process uses fixed weight coefficients, making it difficult to adapt to varying projection distances and the optical properties of reflective media. This can easily lead to color temperature deviations in scenarios involving long-distance projection or high-reflectivity screens. Furthermore, existing technologies rely on preset thresholds to control brightness stability and lack a quantitative assessment of the dynamic characteristics of image brightness distribution. This can cause periodic brightness jitter when playing fast-moving images, impacting the visual experience. Summary of the Invention

[0005] To address the problem that existing technologies rely on a single ambient light sensor to collect data, making it difficult to accurately distinguish the coupling effect between ambient light and the brightness of the projected image, resulting in light source power adjustment being easily disturbed by sudden changes in ambient light. For example, when indoor lighting is suddenly turned on, the sensor mistakenly judges that the ambient brightness has increased overall, triggering unnecessary power reduction, causing the image to be partially too dark. Existing image brightness analysis often uses a global average calculation, ignoring the uneven distribution of light and dark areas in the image, resulting in loss of detail in highlight areas or energy waste in dark areas. The power adjustment model lacks a dynamic priority mechanism and cannot adjust the adjustment order in real time according to changes in image content. Adjustment lag or overcompensation is prone to occur when switching between complex scenes. The ambient light compensation link uses a fixed weight coefficient, which is difficult to adapt to the different projection distances and the optical properties of the reflective medium. It is prone to color temperature deviation in long-distance projection or high-reflectivity screen scenes. In addition, the existing technology relies on preset thresholds to control brightness stability and lacks quantitative evaluation of the dynamic characteristics of the image brightness distribution. When playing fast-moving images, periodic brightness jitter occurs, affecting the visual experience. The embodiments of the present invention provide a method for dynamically adjusting the light source power of an intelligent projector. The technical solution is as follows: In one aspect, a method for dynamically adjusting the light source power of an intelligent projector is provided, the method comprising: S1: Acquire multiple frames of image data through the image sensor, construct an image pyramid by downsampling layer by layer and divide it into fixed grid areas, use the sliding window algorithm to aggregate the HSV-V values of pixels in the grid, and output the regional brightness variation rate data; S2: calling the brightness variation rate data of the region, performing binary comparison on the grid value and the dynamic threshold, sorting the results, and generating a power adjustment priority sequence; S3: calling the preceding grid data of the power adjustment priority sequence, obtaining the current illuminance value through the ambient light sensor, combining the original illuminance values of adjacent nodes and the adjustment parameters, calculating the compensation coefficient using the linear interpolation method, and generating the ambient light correction weight parameter; S4: calling the ambient light correction weight parameter, collecting multiple frames of brightness data through the image sensor, extracting the main peak area span using the histogram peak detection algorithm, combining the parameters for numerical fusion, and outputting the brightness structure stability coefficient; S5: Call the brightness structure stability coefficient, normalize it and input it into the weighted fusion function to calculate the driving current ratio value, perform proportional scaling based on the original driving current reference value, send it to the driving circuit, and obtain the grid light source driving current adjustment parameter.

[0006] As a further embodiment of the present invention, the mutation rate is normalized by Z-score to a dimensionless parameter, and the calculation formula is: ; in, is the original mutation rate value of the area to be evaluated, is the mean of samples in the corresponding area, is the standard deviation; The dynamic threshold is set according to the mean and standard deviation of the brightness variation rate of the current area, using the formula: ; in, represents the dynamic threshold, represents the mean brightness variation rate, represents the standard deviation, is the adjustment coefficient; The ambient illumination value is normalized into a dimensionless parameter by min-max standardization; The span of the main peak area is determined based on the maximum frequency grayscale value and its neighborhood distribution characteristics, and the specific range is adaptively set by the algorithm based on the statistical characteristics of the image; The driving current proportional value reflects the current adjustment value of the current system under various factors; The weighted fusion function needs to combine the temperature response weight factor and the load disturbance response factor; The regional brightness variation rate data includes the maximum variation rate, the minimum variation rate, and the average variation rate; the power adjustment priority sequence specifically refers to the brightness change level, the adjustment response level, and the grid index number; the ambient light correction weight parameters include the weight adjustment coefficient range, the linear interpolation change rate, and the ambient illumination sensitivity factor; the brightness structure stability coefficient specifically includes the main peak span range, the peak concentration, and the structural stability gradient; the grid light source drive current adjustment parameters include the current adjustment ratio value, the current output mode, and the current adjustment delay period.

[0007] As a further solution of the present invention, the specific steps of S1 include: S101: Acquire multiple frames of image data collected by an image sensor, call the RGB channel values in the multiple frames of image data, convert the RGB values into HSV values using a pixel-by-pixel calculation formula, establish an image pyramid hierarchical mapping relationship based on pixel coordinates, demarcate regions according to the grid size set by the image boundary, and obtain multi-scale grid region data; The image pyramid performs multi-scale detection, and the number of layers is determined by a trade-off between computational complexity. The number of layers in the image pyramid must meet the accuracy loss threshold of ΔL / L0 ≤ 5%, where L0 is the baseline layer accuracy and ΔL is the change in detection performance caused by increasing or decreasing the number of layers. S102: Based on the multi-scale grid area data, a sliding window algorithm is used to scan each grid area, locate the HSV-V value in the window, extract the pixel V value sequence, calculate its distribution amplitude within the spatial range, and obtain the window brightness spatial distribution amplitude data; The sliding window algorithm uses adaptive window size and step size settings, where the window size and step size are automatically adjusted according to the image pyramid level; S103: calling the window brightness spatial distribution amplitude data, performing offset analysis on the distribution amplitude at the same position according to the frame sequence, calculating the change rate in combination with the time interval, and obtaining regional brightness variation rate data.

[0008] As a further solution of the present invention, the regional brightness variation rate is calculated using the formula: ; in, represents the regional brightness variation rate, Representative Frame at time The brightness value is defined as the luminous flux per unit area, in lumens per square meter. Representative Frame at time The brightness value is defined as the luminous flux per unit area, in lumens per square meter. represents the inter-frame time interval, Represents the total number of frames analyzed.

[0009] As a further solution of the present invention, the specific steps of S2 include: S201: Calling the regional brightness variation rate data, extracting the brightness variation rate value corresponding to the grid position, marking the extracted value with the grid space identifier, constructing a data structure matching the grid position and the variation rate value, and obtaining the grid brightness variation rate distribution value; The spatial identification is implemented using Hilbert curve encoding; S202: Based on the grid brightness variation rate distribution value, the brightness variation rate value of the grid position is called, and a numerical judgment is performed according to the dynamic threshold set in the current time period. Those with values not lower than the dynamic threshold are marked as effective adjustment targets, and the rest are marked as secondary targets, thereby generating a grid adjustment effectiveness grading value; The update period of the dynamic threshold is set to be synchronized with the frame rate of the image acquisition device; S203: Extract the grid index marked as a valid adjustment target according to the grid adjustment effectiveness grading value, construct a pairing set of sequence numbers and brightness mutation rate values, sort them by mutation rate value using a maximum heap priority queue algorithm, and obtain a power adjustment priority sequence.

[0010] As a further solution of the present invention, the specific steps of S3 include: S301: calling the preceding grid data in the power adjustment priority sequence, extracting the corresponding time period adjustment parameters and basic grid illumination values according to the node arrangement order in the power adjustment sequence, integrating the node parameter items in order according to the node number, and generating a node illumination matching parameter set; The time period adjustment parameters include a color temperature compensation coefficient and a gamma correction value; S302: calling the node basic grid illumination value in the node illumination matching parameter set, obtaining the current illumination value of the ambient light sensor through the ambient light sensor, performing interval positioning, and calling the linear interpolation method based on the illumination values of adjacent nodes and corresponding adjustment parameters to calculate the deviation ratio and generate an illumination compensation coefficient value; The illumination compensation coefficient value compensates for the deviation between the real-time illumination value currently detected by the ambient light sensor and the preset basic illumination parameter; S303: combining the illumination compensation coefficient value with the node adjustment parameter in the node illumination matching parameter set, correcting the original node parameter data, reorganizing the correction results in node order and classifying and summarizing them according to region labels to generate ambient light correction weight parameters; The correction process needs to ensure that the sum of the weight parameters converges to the interval [0.8, 1.2].

[0011] As a further solution of the present invention, the specific steps of S4 include: S401: calling the ambient light correction weight parameter, collecting multiple frames of brightness data through the image sensor, constructing a brightness histogram of each frame of image, detecting the main peak band of grayscale frequency and extracting adjacent grayscale value segments to obtain the main peak grayscale span interval; The main peak band of the grayscale frequency detection adopts the sliding mean filtering method, and the window width is set according to the ratio of the total interval of the histogram; S402: Calling the main peak grayscale span interval, combining the ambient light correction weight parameter to set the weighting coefficient according to the grayscale ratio, applying the amplitude adjustment parameter in the histogram peak detection algorithm to perform grayscale amplitude conversion, and obtaining the brightness offset correction value interval; S403: Calling the brightness offset correction value interval, using the dynamic time warping algorithm to convert the normalized eigenvalues based on the grayscale value distribution on the time axis, screening the data group whose grayscale change amplitude is close to the overall sequence change mean, and generating the brightness structure stability coefficient.

[0012] As a further solution of the present invention, the normalized eigenvalue is calculated using the formula: ; in, represents the normalized eigenvalue of the time series, Represents the original grayscale sampling value at time t, Represents the arithmetic mean of the grayscale sequence of the time axis, Represents the total number of sampling points on the time axis, represents the time interval between the jth sampling point and the previous sampling point, Represents the number of valid sampling intervals, Represents the total span of the time axis, is the light intensity attenuation factor, and its value range is [0, 1]. Its value is positively correlated with the aging degree of the light source. is the nonlinear response coefficient, and its value range is [1, 3].

[0013] As a further solution of the present invention, the specific steps of S5 include: S501: calling the brightness structure stability coefficient, normalizing the parameter according to the difference between the minimum value and the maximum value in the periodic monitoring data, calling the weighted fusion function to integrate the normalized parameters with the set weights, and generating a fusion calculation intermediate value; S502: Based on the fusion calculation intermediate value, calling the set driving current proportion adjustment reference value to perform conversion processing on it, comparing the original driving current reference value with the conversion result, calculating the dynamic proportional factor according to the offset, and generating the driving current proportional value; The dynamic proportional factor calculation introduces a temperature compensation coefficient and a load impedance correction term; S503: Calling the driving current ratio value and the original driving current reference value, calculating the current output value according to the ratio relationship, converting the current output value into a control instruction and sending it to the grid light source driving circuit through the interface module to obtain the grid light source driving current adjustment parameter; The control instructions adopt PWM modulation mode, and the carrier frequency matches the response bandwidth of the light source.

[0014] As a further solution of the present invention, the dynamic scale factor is calculated using the formula: ; in, represents the dynamic scale factor, Represents the absolute difference between the original drive current reference value and the conversion result, Indicates the The temperature sampling value at the time of sampling, Indicates the Load change rate at sampling time, Represents the temperature response weight factor, the value is adjusted according to the thermal resistance coefficient of the radiator, and is negatively correlated with the thermal resistance coefficient of the radiator. represents the load disturbance response factor, Represents the total number of sampling times within the sampling period, Indicates the original driving current reference value, Indicates the current regulation reference scaling factor.

[0015] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: Precise adaptation of light source power is achieved through multi-dimensional data fusion and a dynamic priority control mechanism. During image processing, an image pyramid structure combined with a sliding window algorithm is used to rasterize and partition multiple frames. Regional brightness variation rates are calculated by aggregating HSV-V values, improving the precision of local brightness feature extraction and effectively capturing dynamically changing light and dark areas within the image. Binarization comparison and priority sequence generation based on dynamic thresholds automatically select key adjustment areas based on real-time data differences, avoiding energy waste and detail loss caused by global uniformity. When introducing ambient light correction weight parameters, the ambient illumination values are fused with priority sequence data through linear interpolation, enhancing the adaptability of ambient light compensation and reducing the impact of sudden external lighting changes on image stability. A histogram peak detection algorithm, combined with the calculation of the brightness structure stability coefficient, identifies the main brightness distribution range of the image and suppresses the negative impact of abnormal brightness fluctuations on power regulation. Through normalized weighted fusion and a scaling mechanism, dynamic balance of drive current is achieved, ensuring image uniformity while reducing light source power consumption. This solution achieves the combined effects of improved dimming response speed, optimized energy efficiency and enhanced visual comfort through layered processing, dynamic threshold screening, multi-parameter fusion compensation and stability quantitative control. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of the workflow of the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0018] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0019] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0020] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0021] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0022] See also Figure 1 The embodiment of the present invention provides a method for dynamically adjusting the light source power of an intelligent projector. The processing flow of the method may include the following steps: S1: Acquire multiple frames of image data through the image sensor, construct an image pyramid by downsampling layer by layer and divide it into fixed grid areas, use the sliding window algorithm to aggregate the HSV-V values of pixels in the grid, and output the regional brightness variation rate data; S2: Call the regional brightness variation rate data, perform binary comparison on the grid value and the dynamic threshold, sort the results and generate the power adjustment priority sequence; S3: Call the previous grid data of the power adjustment priority sequence, obtain the current illuminance value through the ambient light sensor, combine the original illuminance value of the adjacent nodes and the adjustment parameters, use the linear interpolation method to calculate the compensation coefficient, and generate the ambient light correction weight parameter; S4: Call the ambient light correction weight parameter, collect multiple frames of brightness data through the image sensor, use the histogram peak detection algorithm to extract the span of the main peak area, combine the parameters for numerical fusion, and output the brightness structure stability coefficient; S5: Call the brightness structure stability coefficient, normalize it, and input it into the weighted fusion function to calculate the driving current ratio value. Perform scaling based on the original driving current reference value and send it to the driving circuit to obtain the grid light source driving current adjustment parameter. The regional brightness variation rate data includes the maximum variation rate, the minimum variation rate, and the average variation rate. The power adjustment priority sequence specifically refers to the brightness change level, the adjustment response level, and the grid index number. The ambient light correction weight parameters include the weight adjustment coefficient range, the linear interpolation change rate, and the ambient illumination sensitivity factor. The brightness structure stability coefficient specifically includes the main peak span range, the peak concentration, and the structural stability gradient. The grid light source drive current adjustment parameters include the current adjustment ratio value, the current output mode, and the current adjustment delay period.

[0023] Specifically, the steps of S1 are: S101: Acquire multiple frames of image data collected by an image sensor, call the RGB channel values in the multiple frames of image data, convert the RGB values into HSV values using a pixel-by-pixel calculation formula, establish an image pyramid hierarchical mapping relationship based on pixel coordinates, demarcate regions according to the grid size set by the image boundary, and obtain multi-scale grid region data; Image pyramids are used for multi-scale detection. The number of layers is determined by a trade-off between computational complexity and accuracy. The number of layers in an image pyramid must meet the accuracy loss threshold of ΔL / L0 ≤ 5%, where L0 is the baseline layer accuracy and ΔL is the change in detection performance caused by increasing or decreasing the number of layers. Get the image sensor Select the third frame as the reference frame in the 5 consecutive frames of the high-resolution video stream and extract the pixel coordinates in the frame. The RGB value at ,Will 、 、 Substitute into the conversion formula: , , ; When building a four-layer pyramid structure, 1 / 2 rate downsampling is used layer by layer to maintain feature continuity and avoid phase shift caused by odd number of sampling. Resolution retention , second floor Downsample to , third floor Downsample to , fourth floor Downsample to , the grid size is set to layer Pixels, layer Pixels, layer Pixels, layer Pixels, when testing the effect of the number of pyramid layers on detection performance, a three-layer structure was measured (satisfy The requirement is derived from the computational accuracy-efficiency balance model proposed by the MIT Vision Lab in 2018. When the number of layers exceeds 4, the GPU memory usage increases exponentially). Five-layer structure (exceeding the threshold), a four-layer structure is selected, as shown in Table 1. The pyramid level parameters, when the grid area is divided, Layer Generation grids, Layer Generation grids, Layer Generation grids, Layer Generation grids.

[0024] Table 1: Image pyramid level parameter table: Hierarchy Resolution Grid size Number of grids <![CDATA[L0]]> 1280×720 64×64 20×11 <![CDATA[L1]]> 640×360 32×32 20×11 <![CDATA[L2]]> 320×180 16×16 20×11 <![CDATA[L3]]> 160×90 8×8 20×11 ; As shown in Table 1, the multi-level grid size is The ratio decreases step by step, and the number of grids remains stable. Layer Grid Area, collect all the pixels The average brightness is calculated as , the standard deviation is , when the brightness difference between adjacent grids exceeds the threshold Trigger area markers, such as grid With grid The average brightness difference is , it is determined to be an abnormal area.

[0025] S102: Based on the multi-scale grid region data, a sliding window algorithm is used to scan each grid region, locate the HSV-V value within the window, extract the pixel V value sequence, calculate its distribution amplitude within the spatial range, and obtain the window brightness spatial distribution amplitude data; The sliding window algorithm uses adaptive window size and step size settings, where the window size and step size are automatically adjusted according to the image pyramid level; exist Layer Grid Regional adoption The pixel window is scanned, and the window step is set to When the local contrast is too high, the multi-resolution joint scanning mode is automatically enabled. The high-level (low-resolution) quickly locates the suspicious area, and the bottom-level (high-resolution) performs sub-pixel verification. The actual measurement increases the scanning efficiency by 3.8 times, and the extraction window pixels Values form a sequence , when calculating the spatial distribution amplitude, first find the mean: ; Then calculate the sum of the squares of the differences between the pixel values and the mean: ; Finally calculate the standard deviation , when the window slides to When the position is detected The value suddenly increased to , exceeding the preset threshold , triggering abnormal recording, and automatically adjusting the window size to Pixels are rescanned and the step size is reduced to pixels, recalculated within the adjusted window .

[0026] S103: Calling the window brightness spatial distribution amplitude data, performing offset analysis on the distribution amplitude at the same position according to the frame sequence, calculating the change rate based on the time interval, and obtaining the regional brightness variation rate data; Calling Grid Region continuous Luminance data of the frame: ; Time interval per frame ,in Obtained through the following process: First extract the grid Pixel V value matrix , calculate the weighted average: ; Weight Used to enhance Nearby brightness sensitivity, such as a pixel hour , and then through the sensor conversion coefficient Converted to luminous flux ,when Time , calculate the mutation rate, using the formula: ; in, represents the regional brightness variation rate, Representative Frame at time The brightness value is defined as the luminous flux per unit area, in lumens per square meter. Representative Frame at time The brightness value is defined as the luminous flux per unit area, in lumens per square meter. represents the inter-frame time interval, Represents the total number of frames analyzed.

[0027] First find the absolute difference between frames: , , , ; Add up: ; Denominator: ; final: ; when The alarm is triggered when the threshold is set based on 200 sets of experimental data: normal fluctuation range and abnormal events Inter-existence Safety margin, the innovativeness of the formula is reflected in the Correction coefficient eliminates frame deviation (when Time correction ), and the absolute value summation method is used to avoid the offset of positive and negative fluctuations, and finally the mutation rate value of the region is output. and abnormal status markers.

[0028] Specifically, the steps of S2 are: S201: Calling regional brightness mutation rate data, extracting brightness mutation rate values corresponding to grid positions, marking the extracted values with corresponding grid space identifiers, constructing a data structure matching grid positions with mutation rate values, and obtaining grid brightness mutation rate distribution values; Spatial identification is implemented using Hilbert curve encoding; Call the brightness variation rate data set stored in the image processing system, which includes the position of each grid cell and its corresponding brightness variation rate, and read the grid from the storage unit. Brightness variation rate value , extract the position parameters of the grid in the image coordinate system , through the Hilbert curve encoding method, the position of each grid is encoded as a one-dimensional identifier for fast search and management. When converting the position coordinates into Hilbert code, first convert the two-dimensional coordinates Quantized to 8-bit binary number: , , take the first 8 digits , perform bit crossover operation through Hilbert curve coding algorithm to generate a 32-bit space identification code , establish data pairs Store in hash table, when processing adjacent grids of When generate , and finally build a key-value pair set containing 1024 entries .

[0029] S202: Based on the grid brightness variation rate distribution value, the brightness variation rate value of the grid position is called, and a numerical judgment is performed according to the dynamic threshold set in the current time period. Those with values not lower than the dynamic threshold are marked as effective adjustment targets, and the rest are marked as secondary targets, thereby generating a grid adjustment effectiveness grading value; The update period of the dynamic threshold is set to be synchronized with the frame rate of the image acquisition device; Get the dynamic threshold value for the current period ,in (corresponding to 09:30), calculated , traverse the hash table entries, when it is detected When executing judge: , marked as a valid regulation target, when processing hour, , marked as secondary targets, and finally generate a set of marks , of which effective targets accounted for 18.7%.

[0030] S203: According to the grid adjustment effectiveness grading value, extract the grid index marked as the effective adjustment target, construct a pairing set of sequence number and brightness mutation rate value, sort them by mutation rate value using the maximum heap priority queue algorithm, and obtain the power adjustment priority sequence.

[0031] From the previously generated grid adjustment effectiveness rating values, filter out the grid indexes marked as valid adjustment targets. The grids will be given priority for adjustment. The valid target index is extracted as A total of 192 items, which will be marked as effective adjustment targets and their corresponding brightness mutation rate values ​​paired to build a pairing set , use the maximum heap priority queue algorithm to sort the pairing set according to the brightness mutation rate value, and when initializing the maximum heap, put the first element Set as the root node and insert the secondary element Time comparison , place it in the left child node, and continue inserting When, because , execute node exchange: update the root node to 72.1, move the original root node downward, and After comparisons, the top element of the heap is , and finally output the priority sequence , among which the top 5% elements ( ) is assigned the highest regulation power level.

[0032] Specifically, the steps of S3 are: S301: Calling the preceding grid data in the power adjustment priority sequence, extracting the corresponding time period adjustment parameters and basic grid illumination values according to the node arrangement order in the power adjustment sequence, integrating the node parameter items in order according to the node number, and generating a node illumination matching parameter set; The time period adjustment parameters include color temperature compensation coefficient and gamma correction value; Extract the top 5% nodes (numbers 3, 1, 5, 8, 12, 15, 21, 25, 30, 35) from the priority sequence and read the nodes Time period parameters: color temperature compensation coefficient By querying the original color temperature data at 09:30 ,calculate , gamma correction value According to the ambient light contrast Obtained by looking up the table, basic illumination From calibration data, generate tuples ,node The parameters are ( hour Rounded to 1.1), the number, color temperature coefficient, gamma value, and basic illumination value parameters of each grid node are recorded in a complete manner, and arranged in the order of the nodes in the priority sequence. 10 sets of parameter sets are constructed: .

[0033] Table 2: Node parameter comparison table: node Color temperature coefficient Gamma value Basic illuminance (lx) H3 1.2 2.4 300 H1 1.1 2.2 300 H5 1.3 2.5 300 ; As shown in Table 2, the parameter set contains adjustment parameters of 10 nodes, and the gamma value is adjusted according to the contrast level. Mapping by table .

[0034] S302: Calling the node base grid illumination value in the node illumination matching parameter set, obtaining the current illumination value of the ambient light sensor through the ambient light sensor, positioning it within the interval, and using the linear interpolation method based on the illumination values of adjacent nodes and the corresponding adjustment parameters to calculate the deviation ratio and generate the illumination compensation coefficient value; The illumination compensation coefficient value compensates for the deviation between the real-time illumination value currently detected by the ambient light sensor and the preset basic illumination parameter; Ambient light sensor in (09:30) Detecting real-time illumination , locate the benchmark interval , select the node and of , calculate the deviation ratio: ; when When , interpolation calculation compensation coefficient: ; Correction The color temperature coefficient is: , the gamma value is corrected to , similarly The corrected parameters are , generate the compensation set: .

[0035] S303: combining the illumination compensation coefficient value with the node adjustment parameter in the node illumination matching parameter set, correcting the original node parameter data, reorganizing the correction results in node order and classifying and summarizing them according to the region label to generate the ambient light correction weight parameter; The correction process needs to ensure that the sum of weight parameters converges to the interval [0.8, 1.2].

[0036] Compute nodes Weight ,node Weight , summing up all weights to get , normalized ,For example Weight , verify the sum , generate classification weight set , where the maximum weight Corresponding node , minimum , all weight values are distributed in .

[0037] Specifically, the steps of S4 are: S401: Calling the ambient light correction weight parameter, collecting multiple frames of brightness data through the image sensor, constructing a brightness histogram for each frame of the image, detecting the main peak band of the grayscale frequency and extracting adjacent grayscale value segments to obtain the main peak grayscale span interval; The main peak band of grayscale frequency is detected using the sliding mean filtering method, and the window width is set according to the ratio of the total interval of the histogram; From the weight parameter set Extract Node Weight , collect 5 frames of images (frame numbers 101-105), the brightness histogram grayscale distribution of frame number 101 is , total pixels , set the sliding window width Round to 13 and calculate the grayscale Filter value at: take the interval Frequency: ; Sum , mean , the main peak band of the detection is (Average of 10 consecutive gray levels ), extract adjacent segments , span length , frame 102 main peak ,span , generating a set: ; Table 3: Main peak span interval data table: Frame number Starting grayscale End Grayscale Span length 101 110 130 21 102 118 128 22 103 112 132 21 ; As shown in Table 3, the span length fluctuation range is ,conform to stable range.

[0038] S402: Calling the main peak grayscale span interval, combining the ambient light correction weight parameter to set the weighting coefficient according to the grayscale ratio, applying the amplitude adjustment parameter in the histogram peak detection algorithm to perform grayscale amplitude conversion, and obtaining the brightness offset correction value interval; Span interval for frame 101 , calculate the grayscale ratio , combined with the node Weight , set the weighting coefficient , call the amplitude adjustment parameters (Ambient Light corresponding to the table value), calculate the lower limit of brightness offset , upper limit , generate the correction interval , calculated in frame 102 , the final set , the fluctuation range of the median value of the correction interval is , standard deviation , indicating that the brightness distribution is stable.

[0039] S403: Call the brightness offset correction value interval, use the dynamic time warping algorithm to convert the normalized eigenvalue based on the grayscale value distribution on the time axis, select the data group with the grayscale change amplitude close to the overall sequence change mean, and generate the brightness structure stability coefficient.

[0040] Take time series , calculate the arithmetic mean: ; Calculate the standard deviation term: ; Calculate the normalized eigenvalue using the formula: ; in, represents the normalized eigenvalue of the time series, Represents the original grayscale sampling value at time t, Represents the arithmetic mean of the grayscale sequence of the time axis, Represents the total number of sampling points on the time axis, represents the time interval between the jth sampling point and the previous sampling point, Represents the number of valid sampling intervals, Represents the total span of the time axis, is the light intensity attenuation factor, and its value range is [0, 1]. Its value is positively correlated with the aging degree of the light source. is the nonlinear response coefficient, and its value range is [1, 3].

[0041] Time interval seconds (30fps), total span Seconds, set (Light source used for 2000 hours), , calculate the time term: ; Substituting into the formula we get: ; when When it is determined to be a stable data group, the screening meets 3 sets of data to generate the stability coefficient .

[0042] Specifically, the steps of S5 are: S501: Calling the brightness structure stability coefficient, normalizing the parameter according to the difference between the minimum and maximum values in the periodic monitoring data, calling the weighted fusion function to integrate the normalized parameters with the set weights, and generating a fusion calculation intermediate value; Obtain periodic monitoring data , calculate the range , for the second data point Perform normalization: , set time weight (Monitoring cycle Minutes correspond to table values), spatial weight (node Weight and Weight The mean ), calculate the fusion intermediate value After weighted fusion, the obtained 0.4114 will be used as the final calculated intermediate value, which will be used for further processing and adjustment in subsequent steps.

[0043] S502: Based on the fusion calculation intermediate value, call the set driving current proportion adjustment reference value to perform conversion processing on it, compare the original driving current reference value with the conversion result, calculate the dynamic scale factor according to the offset, and generate the driving current proportion value; The dynamic proportional factor calculation introduces the temperature compensation coefficient and load impedance correction term; Compare the fusion calculation intermediate value with the driving current reference value, convert the fusion value into a current proportional value according to the set conversion rules, and set the driving current reference value. , temperature sampling value , load change rate , temperature response weight (Heat sink thermal resistance ), load disturbance factor , scaling factor , calculate the dynamic scale factor, using the formula: ; in, represents the dynamic scale factor, Represents the absolute difference between the original drive current reference value and the conversion result, Indicates the The temperature sampling value at the time of sampling, Indicates the Load change rate at sampling time, Represents the temperature response weight factor, the value is adjusted according to the thermal resistance coefficient of the radiator, and is negatively correlated with the thermal resistance coefficient of the radiator. represents the load disturbance response factor, Represents the total number of sampling times within the sampling period, Indicates the original driving current reference value, Indicates the current regulation reference scaling factor.

[0044] Calculate the denominator = By comparing the original driving current reference value with the converted current value, an offset is obtained to judge the difference between the current conversion result and the expected value. , calculate the scale factor ,when When the overload protection is triggered, the proportional value is generated Based on the calculated dynamic proportional factor, the driving current proportional value is obtained, which is used for subsequent current output calculations.

[0045] Table 4: Dynamic scaling factor calculation table: Sampling point Temperature (°C) Load change rate Temperature contribution Load Item Contribution 1 25 1.0 500 1.2 2 26 1.1 540.8 1.452 3 27 1.2 583.2 1.728 4 26.5 1.15 561.8 1.587 5 25.5 1.05 520.2 1.323 ; As shown in Table 4, the temperature term contributes 99.7% and is the main influencing factor.

[0046] S503: Calling the driving current ratio value and the original driving current reference value, calculating the current output value according to the ratio relationship, converting the current output value into a control instruction and sending it to the grid light source driving circuit through the interface module to obtain the grid light source driving current adjustment parameter; The control instructions adopt PWM modulation mode, and the carrier frequency matches the response bandwidth of the light source; The proportional value Converted to drive current , convert the current output value into a control instruction, which can be a voltage regulation signal, PWM (pulse width modulation) signal or other control format that can be recognized by the hardware. Calculate the 12-bit PWM value (maximum current 200mA), generates control instruction 0x020xB00x87 (address + frequency + duty cycle), and sends the generated control instruction to the driving circuit of the grid light source through the interface module. The driving circuit adjusts the current output according to the received instruction, thereby adjusting the brightness of the light source. The data is transmitted at a rate of 10Mbps through the SPI interface, and the carrier frequency is 0x020xB00x87. (Light source bandwidth ), the drive circuit returns the confirmation signal 0xBB, the measured output current (error 0.07%), generate adjustment parameters The adjustment parameters reflect the actual output of the driving circuit after adjusting the brightness of the light source, which is provided for monitoring and feedback adjustment by the subsequent adjustment system.

[0047] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for dynamically adjusting the light source power of an intelligent projector, characterized in that: The following steps are involved: S1: Acquire multiple frames of image data through the image sensor, construct an image pyramid by downsampling layer by layer and divide it into fixed grid areas, use the sliding window algorithm to aggregate the HSV-V values of pixels in the grid, and output the regional brightness variation rate data; S2: calling the brightness variation rate data of the region, performing binary comparison on the grid value and the dynamic threshold, sorting the results, and generating a power adjustment priority sequence; S3: calling the preceding grid data of the power adjustment priority sequence, obtaining the current illuminance value through the ambient light sensor, combining the original illuminance values of adjacent nodes and the adjustment parameters, calculating the compensation coefficient using the linear interpolation method, and generating the ambient light correction weight parameter; S4: calling the ambient light correction weight parameter, collecting multiple frames of brightness data through the image sensor, extracting the main peak area span using the histogram peak detection algorithm, combining the parameters for numerical fusion, and outputting the brightness structure stability coefficient; S5: Call the brightness structure stability coefficient, normalize it and input it into the weighted fusion function to calculate the driving current ratio value, perform proportional scaling based on the original driving current reference value, send it to the driving circuit, and obtain the grid light source driving current adjustment parameter.

2. The method for dynamically adjusting the light source power of an intelligent projector according to claim 1, characterized in that: The mutation rate is normalized by Z-score to become a dimensionless parameter; The dynamic threshold is set according to the mean and standard deviation of the brightness variation rate of the current area; The ambient illumination value is normalized into a dimensionless parameter by min-max standardization; The span of the main peak area is determined based on the frequency grayscale value and its neighborhood distribution characteristics, and the range is adaptively set by the algorithm based on the statistical characteristics of the image; The driving current proportional value reflects the current adjustment value of the current system under various factors; The weighted fusion function needs to combine the temperature response weight factor and the load disturbance response factor; The regional brightness variation rate data includes the maximum variation rate, the minimum variation rate, and the average variation rate; the power adjustment priority sequence specifically refers to the brightness change level, the adjustment response level, and the grid index number; the ambient light correction weight parameters include the weight adjustment coefficient range, the linear interpolation change rate, and the ambient illumination sensitivity factor; the brightness structure stability coefficient specifically includes the main peak span range, the peak concentration, and the structural stability gradient; the grid light source drive current adjustment parameters include the current adjustment ratio value, the current output mode, and the current adjustment delay period.

3. The method for dynamically adjusting the light source power of an intelligent projector according to claim 1, wherein: The specific steps of S1 include: S101: Acquire multiple frames of image data collected by an image sensor, call the RGB channel values in the multiple frames of image data, convert the RGB values into HSV values using a pixel-by-pixel calculation formula, establish an image pyramid hierarchical mapping relationship based on pixel coordinates, demarcate regions according to the grid size set by the image boundary, and obtain multi-scale grid region data; The image pyramid performs multi-scale detection, and the number of layers is determined by a trade-off between the computational complexity; S102: Based on the multi-scale grid area data, a sliding window algorithm is used to scan each grid area, locate the HSV-V value in the window, extract the pixel V value sequence, calculate its distribution amplitude within the spatial range, and obtain the window brightness spatial distribution amplitude data; The sliding window algorithm uses adaptive window size and step size settings, where the window size and step size are automatically adjusted according to the image pyramid level; S103: calling the window brightness spatial distribution amplitude data, performing offset analysis on the distribution amplitude at the same position according to the frame sequence, calculating the change rate in combination with the time interval, and obtaining regional brightness variation rate data.

4. The method for dynamically adjusting the light source power of an intelligent projector according to claim 3, characterized in that: The brightness variation rate of the region is calculated using the formula: ； in, represents the regional brightness variation rate, Representative Frame at time The brightness value is defined as the luminous flux per unit area, in lumens per square meter. Representative Frame at time The brightness value is defined as the luminous flux per unit area, in lumens per square meter. represents the inter-frame time interval, Represents the total number of frames analyzed.

5. The method for dynamically adjusting the light source power of an intelligent projector according to claim 3, wherein: The specific steps of S2 include: S201: Calling the regional brightness variation rate data, extracting the brightness variation rate value corresponding to the grid position, marking the extracted value with the grid space identifier, constructing a data structure matching the grid position and the variation rate value, and obtaining the grid brightness variation rate distribution value; The spatial identification is implemented using Hilbert curve encoding; S202: Based on the grid brightness variation rate distribution value, the brightness variation rate value of the grid position is called, and a numerical judgment is performed according to the dynamic threshold set in the current time period. Those with values not lower than the dynamic threshold are marked as effective adjustment targets, and the rest are marked as secondary targets, thereby generating a grid adjustment effectiveness grading value; The update period of the dynamic threshold is set to be synchronized with the frame rate of the image acquisition device; S203: Extract the grid index marked as a valid adjustment target according to the grid adjustment effectiveness grading value, construct a pairing set of sequence numbers and brightness mutation rate values, sort them by mutation rate value using a maximum heap priority queue algorithm, and obtain a power adjustment priority sequence.

6. The method for dynamically adjusting the light source power of an intelligent projector according to claim 5, characterized in that: The specific steps of S3 include: S301: calling the preceding grid data in the power adjustment priority sequence, extracting the corresponding time period adjustment parameters and basic grid illumination values according to the node arrangement order in the power adjustment sequence, integrating the node parameter items in order according to the node number, and generating a node illumination matching parameter set; The time period adjustment parameters include a color temperature compensation coefficient and a gamma correction value; S302: calling the node basic grid illumination value in the node illumination matching parameter set, obtaining the current illumination value of the ambient light sensor through the ambient light sensor, performing interval positioning, and calling the linear interpolation method based on the illumination values of adjacent nodes and corresponding adjustment parameters to calculate the deviation ratio and generate an illumination compensation coefficient value; The illumination compensation coefficient value compensates for the deviation between the real-time illumination value currently detected by the ambient light sensor and the preset basic illumination parameter; S303: combining the illumination compensation coefficient value with the node adjustment parameter in the node illumination matching parameter set, correcting the original node parameter data, reorganizing the correction results in node order and classifying and summarizing them according to region labels to generate ambient light correction weight parameters; The correction process needs to converge the sum of weight parameters to the interval [0.8, 1.2].

7. The method for dynamically adjusting the light source power of an intelligent projector according to claim 6, characterized in that: The specific steps of S4 include: S401: calling the ambient light correction weight parameter, collecting multiple frames of brightness data through the image sensor, constructing a brightness histogram of each frame of image, detecting the main peak band of grayscale frequency and extracting adjacent grayscale value segments to obtain the main peak grayscale span interval; The main peak band of the grayscale frequency detection adopts the sliding mean filtering method, and the window width is set according to the ratio of the total interval of the histogram; S402: Calling the main peak grayscale span interval, combining the ambient light correction weight parameter to set the weighting coefficient according to the grayscale ratio, applying the amplitude adjustment parameter in the histogram peak detection algorithm to perform grayscale amplitude conversion, and obtaining the brightness offset correction value interval; S403: Calling the brightness offset correction value interval, using the dynamic time warping algorithm to convert the normalized eigenvalues based on the grayscale value distribution on the time axis, screening the data group whose grayscale change amplitude is close to the overall sequence change mean, and generating the brightness structure stability coefficient.

8. The method for dynamically adjusting the light source power of an intelligent projector according to claim 7, wherein: The normalized eigenvalue is calculated using the formula: ； in, represents the normalized eigenvalue of the time series, Represents the original grayscale sampling value at time t, Represents the arithmetic mean of the grayscale sequence of the time axis, Represents the total number of sampling points on the time axis, represents the time interval between the jth sampling point and the previous sampling point, Represents the number of valid sampling intervals, Represents the total span of the time axis, is the light intensity attenuation factor, and its value range is [0, 1]. Its value is positively correlated with the aging degree of the light source. is the nonlinear response coefficient, and its value range is [1, 3].

9. The method for dynamically adjusting the light source power of an intelligent projector according to claim 7, wherein: The specific steps of S5 include: S501: calling the brightness structure stability coefficient, normalizing the parameter according to the difference between the minimum value and the maximum value in the periodic monitoring data, calling the weighted fusion function to integrate the normalized parameters with the set weights, and generating a fusion calculation intermediate value; S502: Based on the fusion calculation intermediate value, calling the set driving current proportion adjustment reference value to perform conversion processing on it, comparing the original driving current reference value with the conversion result, calculating the dynamic proportional factor according to the offset, and generating the driving current proportional value; The dynamic proportional factor calculation introduces a temperature compensation coefficient and a load impedance correction term; S503: Calling the driving current ratio value and the original driving current reference value, calculating the current output value according to the ratio relationship, converting the current output value into a control instruction and sending it to the grid light source driving circuit through the interface module to obtain the grid light source driving current adjustment parameter; The control instructions adopt PWM modulation mode, and the carrier frequency matches the response bandwidth of the light source.

10. The method for dynamically adjusting the light source power of an intelligent projector according to claim 9, characterized in that: The dynamic scaling factor is calculated using the formula: ； in, represents the dynamic scale factor, Represents the absolute difference between the original drive current reference value and the conversion result, Indicates the The temperature sampling value at the time of sampling, Indicates the Load change rate at sampling time, Represents the temperature response weight factor, the value is adjusted according to the thermal resistance coefficient of the radiator, and is negatively correlated with the thermal resistance coefficient of the radiator. represents the load disturbance response factor, Represents the total number of sampling times within the sampling period, Indicates the original driving current reference value, Indicates the current regulation reference scaling factor.

Images