Intelligent light-sensing adjustment method and system for variable message sign

Abstract

The present application relates to sensitive element and sensor manufacturing technical field, disclose a kind of road variable information board intelligent light feeling adjustment method and system, the method includes obtaining ambient light, the surface reflection of information board, current display brightness and illumination change data;Illumination interference vector is constructed, and the intensity value of light interference type and quantization are obtained;Fusion data determine glare intensity index, and threshold value is calculated glare influence coefficient;Determine brightness target value in combination with visual comfort benchmark, generate dynamic adjustment parameter and adjust brightness;Obtain feedback signal and calculate confidence score, and not reach threshold value then iterative optimization parameter, until information board display is clear.The present application can realize the accurate light feeling adjustment of road variable information board under complex illumination environment, satisfy the high-quality demand of information board display stability and environmental adaptability.

Description

Technical Field

[0001] This invention relates to the field of manufacturing sensitive elements and sensors, and in particular to a method and system for intelligent light-sensing adjustment of road variable message signs. Background Technology

[0002] Currently, in the field of sensitive component and sensor manufacturing, variable message signs serve as the core carrier for disseminating traffic information, and their display clarity directly affects driver safety and traffic scheduling efficiency. Intelligent light-sensing adjustment of these signs is a crucial element in addressing light interference and ensuring all-weather visibility. Achieving dynamic brightness adaptation by accurately capturing ambient light and glare characteristics has become a core technological direction for improving adjustment accuracy and environmental adaptability.

[0003] Existing methods for adjusting the light sensitivity of variable message signs in the industry mainly rely on a single light sensor or fixed threshold adjustment. For example, brightness is controlled by a single parameter of ambient light intensity, preset brightness levels are used for switching, or adjustments are made directly ignoring the source and intensity differences of glare. However, this approach is clearly inadequate in complex operating environments. Because data from a single sensor cannot fully reflect the multidimensional characteristics of light interference, it is easily affected by different types of glare, such as direct light and reflected light, leading to adjustment deviations. Fixed thresholds are difficult to adapt to dynamic changes in lighting, failing to respond accurately during peak glare times such as early morning and evening. Furthermore, the lack of real-time verification and iterative optimization of the adjustment effect, especially in scenarios with rapidly fluctuating light, can easily result in displays that are too bright or too dark, affecting the effectiveness of information transmission.

[0004] In summary, existing technologies are insufficient to achieve precise light-sensing adjustment of variable message signs under complex lighting conditions, and cannot meet the high-quality requirements of display stability and environmental adaptability. Summary of the Invention

[0005] This invention provides a method and system for intelligent light-sensing adjustment of road variable message signs, so as to achieve precise light-sensing adjustment of road variable message signs under complex lighting conditions and meet the high-quality requirements of display stability and environmental adaptability of the signs.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides an intelligent light-sensing adjustment method for road variable message signs, comprising: Acquire ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data; Based on the ambient light data and the surface reflection data, the light characteristics are analyzed and a light interference vector is constructed. By classifying and quantizing the position projection of the light interference vector, the light interference type and intensity value are obtained. By integrating the light interference type, the intensity value, the current display brightness data, and the light change data, the correlation pattern between the integrated data and the pre-acquired historical light data is analyzed to determine the glare intensity index. If the glare intensity index exceeds the preset glare judgment threshold, the spectral distribution data of the current environment is collected, effective spectral features are screened and the glare source is determined, and the glare influence coefficient is obtained by weighting the effective spectral features and the weight coefficients corresponding to the glare source. Based on the glare impact coefficient and the pre-acquired visual comfort benchmark, the brightness adjustment range is calculated, and the target brightness value is determined based on the adjustment range. Compare the target brightness value with the existing brightness of the information board. If the brightness difference exceeds a preset brightness difference threshold, generate dynamic brightness adjustment parameters and adjust the brightness of the information board according to the dynamic brightness adjustment parameters. Acquire the adjusted display brightness data and the real-time visibility feedback signal, and calculate the confidence score of the display brightness data and the feedback signal; If the confidence score reaches the preset effective judgment threshold, the dynamic brightness adjustment parameter remains unchanged; otherwise, the dynamic brightness adjustment parameter is iteratively optimized until the information board display is clear.

[0007] Secondly, the present invention provides an intelligent light-sensing adjustment system for road variable message signs, comprising: The data acquisition module is used to acquire ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data. The interference analysis module is used to analyze the light characteristics and construct the light interference vector based on the ambient light data and the surface reflection data. By classifying and quantizing the position projection of the light interference vector, the type and intensity value of the light interference are obtained. The index determination module is used to integrate the light interference type, the intensity value, the current display brightness data and the light change data, analyze the correlation pattern with the pre-acquired historical light data, and determine the glare intensity index of glare intensity. The coefficient calculation module is used to collect the spectral distribution data of the current environment, screen effective spectral features and determine the source of glare if the glare intensity index exceeds the preset glare judgment threshold, and obtain the glare influence coefficient through feature mapping and weighted calculation. The target determination module is used to calculate the brightness adjustment range based on the glare influence coefficient and a pre-acquired visual comfort benchmark, and determine the brightness target value based on the adjustment range. The parameter adjustment module is used to compare the target brightness value with the existing brightness of the information board. If the brightness difference exceeds the preset brightness difference threshold, dynamic brightness adjustment parameters are generated and sent to the control terminal of the information board. The feedback scoring module is used to acquire the adjusted display brightness data and the real-time visibility feedback signal, and to calculate the confidence score of the display brightness data and the feedback signal. The verification and optimization module is used to verify the effectiveness of the dynamic brightness adjustment parameters based on the confidence score. If the confidence score reaches the preset effective judgment threshold, the dynamic brightness adjustment parameters are kept unchanged. If not, the dynamic brightness adjustment parameters are iteratively optimized until the information board display is clear.

[0008] Compared with the prior art, the present invention has the following beneficial effects: (1) This invention collects ambient light data, information board surface reflection data, current display brightness data and illumination change data, uses the Otsu algorithm to process the light spot distribution and ambient light data, analyzes the characteristics such as light wavelength and incident angle, constructs a light interference vector with spatiotemporal dimensions, classifies and quantifies the light interference type and intensity value, breaks through the limitation of traditional single sensors that cannot fully reflect the multidimensional characteristics of illumination, explores the dynamic change law of light interference, eliminates the interference of different glare types and environmental noise, provides high-precision basic data support for glare assessment, effectively improves the capture rate of weak glare and complex illumination interference, and solves the brightness deviation problem caused by single parameter adjustment.

[0009] (2) This invention integrates light interference type, intensity value, current brightness and light change data, and determines the glare intensity index by combining historical light data correlation mode. After exceeding the threshold, spectral distribution data is collected, effective features are screened and the source of glare is determined. The glare influence coefficient is obtained by weighted calculation. This invention breaks through the limitation that the traditional fixed threshold cannot adapt to dynamic light, accurately captures the correlation features between glare intensity and source, provides multi-dimensional basis for brightness adjustment, significantly improves the pertinence of glare response under complex working conditions, and makes up for the defect of existing technology that is difficult to distinguish the source of glare.

[0010] (3) The present invention calculates the brightness adjustment range based on the glare influence coefficient and visual comfort benchmark, generates dynamic adjustment parameters by combining the brightness response characteristics of the information board, calculates the confidence score by adjusting the brightness data and visibility feedback signal, and iteratively optimizes the parameters until the display is clear. It breaks through the limitations of traditional lack of real-time verification and iterative optimization, provides an adaptive brightness adjustment basis for operation and maintenance, solves the problem of display being too bright or too dark under rapid light fluctuations, takes into account display stability and environmental adaptability, and meets the high-quality display requirements of information boards under complex lighting conditions. Attached Figure Description

[0011] Figure 1This is a schematic flowchart of an intelligent light-sensing adjustment method for road variable message signs provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of a road variable message sign intelligent light-sensing adjustment system provided in the second embodiment of the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0013] Reference Figure 1 The first embodiment of the present invention provides a method for intelligent light-sensing adjustment of road variable message signs, including the following steps: S101, acquire ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data; S102, Based on the ambient light data and the surface reflection data, analyze the light characteristics and construct a light interference vector. By classifying and quantizing the position projection of the light interference vector, obtain the light interference type and intensity value. S103, integrate the light interference type, the intensity value, the current display brightness data and the light change data, analyze the correlation pattern between the integrated data and the pre-acquired historical light data, and determine the glare intensity index; S104, if the glare intensity index exceeds the preset glare judgment threshold, collect the spectral distribution data of the current environment, screen the effective spectral features and determine the glare source, and calculate the glare influence coefficient by weighting the effective spectral features and the weight coefficients corresponding to the glare source. S105, Based on the glare impact coefficient and the pre-acquired visual comfort benchmark, calculate the brightness adjustment range, and determine the brightness target value based on the adjustment range; S106, compare the target brightness value with the existing brightness of the information board. If the brightness difference value exceeds the preset brightness difference threshold, generate dynamic brightness adjustment parameters and adjust the brightness of the information board according to the dynamic brightness adjustment parameters. S107, acquire the adjusted display brightness data and the real-time visibility feedback signal, and calculate the confidence score of the display brightness data and the feedback signal; S108, if the confidence score reaches the preset effective judgment threshold, the dynamic brightness adjustment parameter remains unchanged; if it does not reach the threshold, the dynamic brightness adjustment parameter is iteratively optimized until the information board display is clear.

[0014] In step S101, acquiring ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data includes: Collect ambient light radiation intensity and surface light spot distribution data of the information board, and process the ambient light radiation intensity and surface light spot distribution data using the Otsu algorithm to obtain ambient light data and surface reflection data of the information board. Read the current display brightness data of the information board; Continuous ambient light data is collected at preset fixed time intervals, and the ratio of the difference between adjacent data points of the ambient light data to the time interval is calculated to obtain the illumination change data of the surrounding environment.

[0015] It should be noted that the ambient light radiation intensity was collected using a high-precision photosensitive sensor with a measurement range of 0-20000 lx and a sampling frequency of 10 Hz. The surface light spot distribution data of the information board was captured by a high-definition industrial camera with a resolution of 1920×1080 pixels. The Otsu adaptive segmentation algorithm with maximum inter-class variance was used, combined with morphological opening and closing operations for optimization, to extract the area, shape, and brightness distribution information of the light spots. First, the original image captured by the camera was subjected to Gaussian filtering for noise reduction, with a filter kernel size of 3×3 and a standard deviation of 0.8, to eliminate interference noise such as environmental dust and lens reflection. Then, the optimal segmentation threshold was automatically calculated by the Otsu algorithm to divide the image into light spot foreground and background regions. Finally, morphological opening operations (5×5 erosion kernel) were used to remove small noise points, and closing operations (5×5 dilation kernel) were used to fill the internal holes of the light spots to ensure the integrity of the light spot region. For example, at noon, the sensor collects an ambient light radiation intensity of 18,000 lx. After the camera captures the image of the information board, the above segmentation process identifies a strongly reflective light spot with an area of ​​20 cm² and a regular outline in the upper left corner. At the same time, the circularity (0.85) and brightness distribution standard deviation (35 nits) of the light spot are obtained.

[0016] Next, the brightness data is read directly through the display driver interface of the information board, with the unit being nits and the measurement accuracy being ±1 nit. This data reflects the current actual luminous intensity of the information board, providing a benchmark for subsequent brightness adjustments. For example, if the current display brightness of the information board is read as 450 nits, it is in a medium brightness output state.

[0017] Finally, the preset fixed time interval is set according to the rate of change of light intensity, with a base interval of 1 second, which can be shortened to 0.5 seconds in scenarios with rapid changes in light. The unit of light intensity change data is lx / second, with positive values ​​indicating increased light intensity and negative values ​​indicating decreased light intensity. For example, if the continuously collected ambient light data are 15000 lx, 15800 lx, and 16500 lx, with a time interval of 1 second, the calculated light intensity change data are 800 lx / second and 700 lx / second, respectively, indicating that the light intensity is in a state of rapid increase.

[0018] In step S102, the step of analyzing light characteristics and constructing a light interference vector based on the ambient light data and the surface reflection data, and obtaining the light interference type and intensity value by classifying and quantizing the light interference vector through position projection, includes: The wavelength and incident angle characteristics of light rays are analyzed from the ambient light data, and the spot area and brightness distribution characteristics are extracted from the reflection data. By combining the light wavelength, the incident angle characteristics, the light spot area, the brightness distribution characteristics, as well as the acquisition time and the installation location information of the information board, a light interference vector containing spatiotemporal dimensions is constructed. Calculate the distance between the illumination interference vector and the boundary of the preset interference category, and determine the light interference type based on the illumination interference vector with the smallest distance; The illumination interference vector is projected into a high-dimensional space, and the distance between the projection position and the preset intensity reference point is calculated to obtain the intensity value.

[0019] It should be noted that the light wavelength was analyzed using a spectral analyzer, covering the visible light band of 380-780nm, and the incident angle was measured using an angle sensor with an accuracy of ±0.5°. The spot area was extracted from surface reflection data using image segmentation technology, employing the Otsu adaptive segmentation algorithm based on a brightness threshold, combined with morphological optimization processing. First, the image corresponding to the surface reflection data was processed, removing noise interference such as environmental dust and lens stray light using a 5×5 Gaussian filter kernel. Then, the Otsu algorithm was used to automatically analyze the image grayscale histogram to determine the optimal segmentation threshold between the spot and the background. Finally, morphological opening operations (3×3 erosion kernel) were used to remove residual small noise points after segmentation, and morphological closing operations (3×3 dilation kernel) were used to fill the holes inside the spot caused by noise, ensuring the integrity and contour accuracy of the spot area. Subsequently, the actual spot area was calculated by pixel coordinate conversion (combined with camera resolution and shooting distance calibration coefficients). Brightness distribution is characterized by calculating the standard deviation of the brightness of all pixels within the segmented spot area. The larger the standard deviation, the more uneven the brightness distribution within the spot, and vice versa.

[0020] It is worth noting that the calibration coefficients are obtained through a calibration board. The image is taken on a calibration board of known size that is coplanar with the information board, and the "actual length corresponding to each pixel" is calculated. Then, the pixel area obtained by dividing the light spot is converted into the actual area according to this ratio. If the shooting distance changes, the conversion coefficient is corrected proportionally using the distance measurement value.

[0021] Next, the acquisition time is converted to a numerical value (0-23) in 24-hour format, and the installation location information is quantized using latitude and longitude coordinates. This information is then concatenated dimensionally with the light wavelength, incident angle, spot area, and brightness distribution characteristics. The acquisition time is first converted to a numerical value of 0–23 and normalized; the installation location is converted from latitude and longitude to numerical values ​​and normalized according to the equipment's coverage area. These two numerical values ​​are then directly concatenated with the wavelength, incident angle, spot area, and brightness distribution in a fixed order to form a feature vector. This results in a 6-dimensional illumination interference vector. This vector takes into account both spatiotemporal and light characteristics, providing a comprehensive basis for subsequent classification. For example, if the acquisition time is 18:00 (normalized to 0.78), the installation location is coded as 0.42, the light wavelength is 620nm, the incident angle is 15°, the spot area is 3.6cm², and the brightness distribution standard deviation is 22, then the 6-dimensional illumination interference vector can be represented as [0.78, 0.42, 620, 15, 3.6, 22] for subsequent classification.

[0022] Subsequently, the preset interference categories include direct sunlight, reflected light, and illuminator interference. The category boundaries are based on over 100,000 light sample data points from different time periods (morning, noon, and evening), weather conditions (sunny, rainy, and foggy), and scenes (urban main roads / suburban roads) over the past year. A Support Vector Machine (SVM) algorithm is used to train the classification model. The penalty coefficient C=1.0 and the kernel function gamma=0.1 are optimized through grid search. The sample feature vectors are used as inputs, and the manually labeled interference types are used as outputs. After training, the optimal classification hyperplanes for the three types of interference are obtained, i.e., the category boundaries. At the same time, because the dimensions of the light interference vectors are different (e.g., wavelength nm, angle °, area cm²), the data of each dimension needs to be min-max normalized first, mapped to the [0,1] interval to eliminate the difference in dimensions. Then, the Euclidean distance algorithm is used to calculate the distance between the vector and the three category boundaries. The one with the smallest distance is the light interference type to be determined. For example, after normalization, a certain light interference vector has a distance of 2.3 from the boundary of the direct sunlight class, and distances of 4.1 and 3.8 from the boundaries of the other two classes, respectively. The light interference type is determined to be direct sunlight.

[0023] Finally, the high-dimensional spatial projection employs kernel principal component analysis (KPCA) with the RBF kernel function (kernel parameter gamma=0.2). The top 10 principal components (cumulative variance contribution rate ≥95%) are selected through eigenvalue decomposition to map the 6-dimensional vector to a 10-dimensional high-dimensional space. During mapping, the normalized 6-dimensional vector is first centered, then the similarity matrix between samples is calculated using the kernel function, transforming it into a covariance matrix in the high-dimensional space. After eigenvalue decomposition, the top 10 principal components are selected to complete the mapping. The preset intensity reference point is the vector projection position in an interference-free scene (uniform ambient light, no reflected light spots). The intensity value is mapped to a numerical range of 0-100 at a distance of 20 times the distance. This linear mapping rule is set based on the projection distance distribution of interference-free to interference-strongly interfering samples over the past year (interference-free distance 0, interference-strongly interfering distance 5), aiming to quantify the strength of interference with the intensity value; a higher value indicates stronger interference. For example, if the distance between the projected vector and the reference point is 3.5, the calculated intensity value is 70, indicating strong light interference.

[0024] In step S103, the process of fusing the light interference type, the intensity value, the current display brightness data, and the illumination change data, analyzing the correlation pattern between the fused data and the pre-acquired historical illumination data, and determining the glare intensity index includes: The light interference type, the intensity value, the current display brightness data, and the illumination change data are normalized to form a feature matrix of real-time illumination. Calculate the similarity between the feature matrix and the pre-acquired historical illumination data, and use the mean of the similarity as the correlation strength. If the correlation strength is within a preset glare sensitivity range, calculate the initial glare intensity index based on the correlation strength. The initial glare intensity index is verified based on the light spot area and the brightness distribution characteristics, and the glare intensity index is obtained after correction.

[0025] It should be noted that after the light interference type is coded by category (direct light 1, reflected light 2, light interference 3), the coded value range is [1,3]. Min-max normalization is used to map it to the [0,1] interval. Intensity values, current display brightness data, and illumination change data are all normalized using the min-max method and mapped to the same interval. The four types of normalized data are then concatenated column-wise to form an N×4 dimensional real-time illumination feature matrix (N is the number of data samples).

[0026] It is worth noting that the historical illumination data comes from the national database of 100 main roads (≥100,000 samples) published by traffic management departments over the past year, covering six weather conditions including sunny, rainy, and foggy days, as well as peak hours such as morning and evening rush hours. The statistics are validated using a 95% confidence interval Bootstrap sampling method, and the interval boundaries can be manually fine-tuned according to road type (e.g., tightening the interval to 0.75-1.0 for highways). Each historical sample and the real-time feature uses the same four features and the same normalization method, stacked row-wise to form a K×4 dimensional historical illumination feature matrix (K is the number of historical samples), ensuring that the similarity with the real-time N×4 dimensional real-time illumination feature matrix can be directly calculated in terms of dimension and scale, covering common scenarios such as urban main roads, suburban roads, and tunnel entrances and exits. Association patterns are extracted using cosine similarity, which measures the degree of association by calculating the cosine of the angle between the real-time illumination feature matrix and the historical sample vector; the closer the value is to 1, the more similar the features are. The glare sensitivity range is set based on the similarity distribution of historical glare scenes over the past year. Statistics show that 90% of historical glare scenes have a similarity between 0.7 and 1.0, therefore this range is set as the basic range. For complex lighting scenarios such as heavy rain or hazy days with large light fluctuations and more dispersed similarity distributions, the range needs to be expanded to 0.65 to 1.0 to avoid missing glare detections. The correlation strength is the average similarity between the real-time lighting feature matrix and all historical glare samples. The initial glare intensity index is obtained by multiplying the correlation strength by 100. This calculation method transforms the abstract similarity into an intuitive glare intensity index. For example, if the average similarity between the real-time lighting feature matrix and 500 historical glare samples is 0.85, which falls within the basic sensitivity range, the calculated initial glare intensity index is 85.

[0027] Finally, the spot area is quantified as the ratio of the actual area to the information board display area, and the brightness feature is quantified as the ratio of the spot brightness to the ambient light brightness. Both are normalized to the interval [0, 0.2] as verification coefficients. The correction formula is: G = G0 × (1 + max(0, (S−0.1) + (L−0.1))), where G represents the final glare intensity index, G0 represents the initial glare intensity index, S represents the spot area coefficient, L represents the brightness feature coefficient, and 0.1 is the baseline coefficient, determined based on the statistical average of the spot proportion and brightness ratio in glare-free scenes over the past year. When the spot area is small and the brightness is low, it indicates that the spot contributes weakly to glare. In this case, the correction term is truncated to 0 at the lower limit, and the final index remains the initial index G0 to avoid incorrectly lowering the glare intensity caused by other strong interference sources due to weak spot characteristics. When the spot area is large and the brightness is high, the correction term is positive and G0 is adjusted upward, thereby reflecting the additional contribution of the spot to glare and achieving positive enhancement verification. For example, if the initial glare intensity index G0 is 85, and the spot area coefficient S is 0.05 and the brightness characteristic coefficient L is 0.08, then (S−0.1)+(L−0.1) is negative, and the correction term is truncated to 0 at the lower limit, finally obtaining G=85, indicating that the initial estimate is not lowered when the spot contribution is weak; if S is 0.15 and L is 0.12, then the correction term is 0.07, finally obtaining G=85×1.07=90.95, reflecting the positive aggravation effect of large spot and high brightness on glare.

[0028] In step S104, if the glare intensity index exceeds a preset glare determination threshold, then the spectral distribution data of the current environment is collected, effective spectral features are screened and the glare source is determined. The glare influence coefficient is obtained by weighting the effective spectral features and the weighting coefficients corresponding to the glare source, including: If the glare intensity index exceeds the preset glare determination threshold, then collect the spectral distribution data of the current environment; The spectral distribution data is filtered and effective spectral features are selected. The effective spectral features are compared with a preset spectral feature library to determine the source of glare. The effective spectral features are mapped to a preset weight matrix, and the weight coefficients corresponding to the glare sources are combined to obtain the glare influence coefficient by weighted summation.

[0029] It should be noted that the specific rules for setting the glare judgment threshold are based on the statistical data of glare occurrence under different lighting scenarios over the past year. The glare intensity index of all valid glare scenarios over the past year is used as the sample. Valid glare scenarios refer to samples in which the system triggers glare judgment during the corresponding time period and is confirmed to have glare impact by maintenance records / video playback. At the same time, the spectrometer and brightness data are continuous and complete (e.g., continuous sampling ≥2s), the sensor is in normal condition and there are no abnormalities such as maintenance obstruction, lens stains / raindrop obstruction, or exposure saturation. Samples that do not meet the above conditions are not included in the statistics.

[0030] The base threshold is set at the 80th percentile of the sample, corresponding to 70. This threshold covers 90% of actual glare scenarios with a low false positive rate. For high-light scenarios such as urban main roads, where light intensity is high and glare is more significant, the 90th percentile is set to 75 to increase the stringency of the judgment. For low-light scenarios such as suburban roads, where light fluctuations are small, the 75th percentile is set to 65 to avoid missing minor glare. The threshold can be manually fine-tuned according to the light characteristics of different areas to adapt to diverse road environments. Environmental spectral distribution data is collected using a miniature spectrometer, covering the 380-780nm visible light band, with a sampling frequency of 5Hz to ensure the capture of subtle spectral changes. For example, if the glare intensity index is 82, exceeding the base threshold of 70, the spectrometer is activated to collect the current environmental spectral distribution data and obtain light intensity distribution curves at different wavelengths.

[0031] Next, when filtering the spectral distribution data to select effective spectral features with high signal-to-noise ratio, an adaptive bandpass filtering technique based on Gaussian windows is employed. The adaptive characteristic is achieved through three steps: First, the spectral curve is smoothed and denoised (e.g., using Savitzky-Golay filtering or moving average), and then local maxima are searched on the smoothed curve: points are compared point by point using a sliding window, and points whose intensity is greater than their left and right neighbors and exceeds the noise threshold are selected as candidate peaks; these candidate peaks are then filtered based on peak prominence and minimum peak distance, ultimately selecting the peak with the highest intensity as the main peak, and obtaining its corresponding peak wavelength and peak intensity; Second, the half-width at half-maximum (WHM) of the peak is calculated with the peak wavelength as the center, and the filtering bandwidth is set to 1.5 times the WHM, achieving automatic bandwidth adjustment based on peak characteristics; Third, a Gaussian window function is dynamically generated based on the bandwidth, and convolution operations are performed on the spectral data to retain the effective signal in the band where the peak is located and significantly attenuate noise signals outside the bandwidth. The entire process requires no manual intervention and can adapt to the spectral morphology of different light sources. Effective spectral features are obtained by extracting three key parameters from the filtered spectrum: peak wavelength, peak intensity, and full width at half maximum (FWHM). For example, if the original spectral data contains multiple stray peaks, after adaptive bandpass filtering, effective spectral features with a center wavelength of 550 nm, a peak intensity of 800 nm, and a FWHM of 30 nm are retained, improving the signal-to-noise ratio to over 30 dB.

[0032] It should be noted that the preset spectral feature library contains spectral parameter samples of common light sources such as sunlight, vehicle headlights, and building reflections. The library samples were obtained through actual measurements using a NIST standard spectrometer (model SR-3501), with no fewer than 100 data sets collected for each light source, and stored after min-max normalization. The library supports dynamic updates: when a new light source type (such as new energy vehicle headlights) is added, it is automatically classified using a clustering algorithm (such as K-means), with an update cycle of once per quarter. Each sample consists of three key parameters and has undergone min-max processing. Before comparison, data preprocessing is required for the effective spectral features and samples in the library. First, the peak wavelength, peak intensity, and half-width at half-maximum (FWHM) are normalized according to their respective value ranges using min-max normalization. The comparison uses a cosine similarity algorithm, which measures feature similarity by calculating the cosine of the angle between two vectors; the closer the value is to 1, the better the feature match. If the highest similarity score exceeds 80% after comparison, the light source type of the corresponding sample is the source of glare; if the highest similarity score does not reach 80%, it is determined to be an unknown light source, and the weighting coefficient corresponding to the reflected light is used for subsequent calculations. For example, the effective spectral feature vector is... =[550,800,30], the sunlight sample vector is =[545,820,28], the building reflected light sample vector is =[530,700,35]; First, calculate the dot product of the two. =550×545+800×820+30×28=299750+656000+840=956590, and the modulus of both. = , Substitute into the formula The similarity with the sunlight sample is approximately 956590÷(971.6×984.3)≈0.99; similarly, the similarity with the building reflected light sample is approximately 0.88. The sunlight sample has the highest similarity, exceeding 80%, indicating that the glare source is direct sunlight.

[0033] Peak intensity mapping values ​​are obtained by normalizing the original peak intensity to the 0-1 range, with the normalization range set based on the peak intensity distribution of spectral data over the past year; peak wavelength mapping values ​​are divided into intervals based on the influence of wavelength on human visual sensitivity, with higher mapping values ​​for the sensitive band around 550nm.

[0034] It should be noted that the peak wavelength mapping value uses the publicly available CIE1931 luminous efficiency function V(λ) as the human eye sensitivity benchmark: the system presets V(λ) lookup data for 380–780 nm and normalizes it to 0–1; when the peak wavelength is input, the corresponding sensitivity is obtained through linear interpolation as the mapping value. Since V(λ) reaches its maximum value around 555 nm, the mapping value is higher around 550 nm, while the mapping value automatically decreases near the two ends of the visible light spectrum (such as <450 nm or >650 nm).

[0035] The full width at half maximum (FWHM) value was then obtained by normalizing the original FWHM to the 0-1 range, reflecting the concentration of the spectrum. The weight matrix was set based on visual comfort experimental data, following the CIE 218:2016 visual comfort guidelines. 120 subjects (aged 20-60 years, corrected visual acuity to 1.0) were recruited and tested in a simulated laboratory (illuminance range 0-100,000 lx). Weights were determined through multiple regression analysis (R²≥0.9), and the weight recommendations from the International Commission on Illumination (CIE) were referenced (e.g., peak intensity weight 0.5 corresponds to CIE S 026 / E:2022): peak intensity directly determines the glare intensity and has the greatest impact on display effect, so its weight was set to 0.5; peak wavelength affects the human eye's sensitivity to glare, so its weight was set to 0.3; FWHM reflects the concentration of spectral energy and affects the sustained range of glare, so its weight was set to 0.2, and the sum of all weights was 1. The weighting coefficients for glare sources are set based on the degree of glare hazard from different light sources. Direct light has a high intensity and great hazard, corresponding to 1.0, while reflected light has a relatively weak intensity, corresponding to 0.8.

[0036] Then, a weighted summation is performed. For example, the effective spectral characteristic mapping values ​​are peak intensity 0.8, peak wavelength 0.7, and half-width at half-maximum 0.6, with weights of 0.5, 0.3, and 0.2, respectively. The glare source is direct light, corresponding to a coefficient of 1.0. The calculation process is 0.8×0.5+0.7×0.3+0.6×0.2=0.73. Multiplying by 1.0 gives the glare influence coefficient of 0.73, which accurately quantifies the degree of glare's impact on the information board display.

[0037] In step S105, the step of calculating the brightness adjustment range based on the glare influence coefficient and a pre-acquired visual comfort benchmark, and determining the brightness target value based on the adjustment range, includes: The glare impact coefficient is mapped to a pre-acquired visual comfort benchmark to obtain the basic brightness adjustment ratio; The basic brightness adjustment ratio is corrected based on the illumination change data, and the brightness adjustment range is calculated. Based on the adjustment range and the current display brightness data, combined with the preset maximum and minimum brightness limits of the information board, the target brightness value is determined.

[0038] It should be noted that the recruitment of participants spanned youth, middle-aged, and elderly individuals. In a laboratory environment, scenarios with varying glare intensities were simulated. Participants adjusted the brightness of the information board to a subjectively comfortable visual state in each scenario, and the appropriate brightness data corresponding to each glare intensity was recorded. Statistical analysis was performed on the experimental data of all participants. After removing outliers, a cubic polynomial was used to fit the relationship between glare intensity and appropriate brightness, resulting in a curve that serves as the visual comfort benchmark.

[0039] It should be noted that the above-mentioned outlier removal refers to removing outliers from suitable brightness data under the same glare intensity using the interquartile range (IQR) method, eliminating extreme samples that deviate significantly from the main distribution, in order to avoid occasional misoperation or subjective bias affecting the fitting curve.

[0040] The mapping between the glare impact coefficient and the visual comfort benchmark is achieved through piecewise linear interpolation. The glare impact coefficient is first divided into three intervals: 0-0.3, 0.3-0.6, and 0.6-1.0. Each interval corresponds to a different brightness adjustment ratio range in the visual comfort benchmark: 0-0.3 corresponds to 1.0-1.5 times, 0.3-0.6 corresponds to 1.5-2.0 times, and 0.6-1.0 corresponds to 2.0-2.5 times. Based on the interval of the glare impact coefficient, the specific basic brightness adjustment ratio is calculated through linear interpolation; the larger the coefficient, the higher the adjustment ratio. For example, a glare impact coefficient of 0.73 falls within the 0.6-1.0 interval. Interpolation yields a basic brightness adjustment ratio of 1.8, indicating that an 80% increase in brightness is needed to match visual comfort.

[0041] Next, the illumination change data is categorized by positive and negative values ​​to distinguish between increasing and decreasing illumination trends. Since there's no need to excessively increase the information board brightness when ambient light increases, but a more significant increase is needed when illumination decreases, the correction formula is as follows: ; in, This represents the adjusted proportions after the revision. This represents the base brightness adjustment ratio. This represents the normalized value of the illumination variation data. The illumination variation data is mapped to the [-0.2, 0.2] interval using the min-max normalization method to avoid over-correction affecting the adjustment accuracy. The formula for the brightness adjustment range is: ; in, This represents the brightness adjustment range. This represents the current display brightness data. For example, the base brightness adjustment ratio. =1.8, normalized value of illumination variation data =-0.1, substituting into the formula, we get =1.8×(1+0.1)=1.98; Current display brightness =450 nits, the brightness adjustment range is calculated. =450×1.98-450=441 nits, precisely matching the needs of dynamic changes in lighting.

[0042] Finally, when calculating the target brightness value based on the adjustment range and the current display brightness data, the target brightness value is the sum of the current display brightness and the adjustment range, and must be limited to the range between the maximum brightness of 1000 nits and the minimum brightness of 50 nits on the information board. If the calculated target brightness value exceeds 1000 nits, then 1000 nits is taken as the final target value to avoid excessive current causing overheating and damage to the equipment; if it is less than 50 nits, then 50 nits is taken as the final target value to ensure that the information board has basic visibility. For example, if the current display brightness is 450 nits and the adjustment range is 600 nits, the calculated target brightness value is 1050 nits, which exceeds the maximum brightness limit, so the final target value is determined to be 1000 nits; if the adjustment range is -420 nits, the calculated target value is 30 nits, which is less than the minimum brightness limit, so the final target value is determined to be 50 nits.

[0043] In step S106, comparing the target brightness value with the existing brightness of the information board, if the brightness difference exceeds a preset brightness difference threshold, generates dynamic brightness adjustment parameters, and adjusts the brightness of the information board according to the dynamic brightness adjustment parameters, including: Calculate the difference between the target brightness value and the existing brightness of the information board to obtain the brightness difference value; If the brightness difference value exceeds the preset brightness difference threshold, the drive current gain is determined by combining the pre-acquired brightness response characteristics of the information board. If the drive current gain is within a preset safe range, dynamic brightness adjustment parameters are generated; The dynamic brightness adjustment parameters are sent to the control terminal of the information board to drive the information board to perform brightness adjustment.

[0044] It should be noted that the target brightness value is the ideal brightness calculated using the glare impact coefficient and visual comfort benchmark. The current brightness is read in real time through the information board display driver interface, and the difference between the two directly reflects the absolute magnitude of the required brightness adjustment. For example, if the target brightness value is 891 nits and the current brightness of the information board is 450 nits, the calculated brightness difference is 441 nits, clearly indicating a need for a significant increase in brightness.

[0045] It's worth noting that the brightness difference threshold is set based on the brightness adjustment sensitivity of the information board. This sensitivity was calibrated experimentally by collecting data on the brightness change rate of different information board models under varying currents. Statistics showed that for most models, under typical viewing distances and ambient light conditions, the minimum perceptible threshold for brightness change by the human eye was approximately 50 nits (the smallest brightness difference that a subject can subjectively and stably distinguish). This was used as the base threshold. High-sensitivity information boards have a fast brightness response rate, allowing for perceptible adjustments with minimal changes; the threshold for these is lowered to 30 nits. Low-sensitivity information boards have a slower response, requiring larger adjustments to show any effect; the threshold for these is raised to 70 nits. The threshold can be manually fine-tuned according to the specific information board model to adapt to different hardware performance.

[0046] The brightness response characteristics of the information board exhibit a nonlinear relationship between brightness and driving current. By collecting multiple sets of driving current and actual brightness data, a cubic polynomial model is fitted using the least squares method. This model characterizes the nonlinear response of LED devices. The coefficients a, b, c, and d are obtained from calibration data points (I_k, L_k) (k=1…M) by minimizing the squared error between the predicted and measured brightness. These coefficients are then stored according to the specific model of the information board for subsequent back-calculation. This model accurately characterizes the nonlinear characteristics of LED devices. The drive current adjustment is derived by inputting the brightness difference value into the nonlinear model, eliminating the need for linear gain calculations and directly matching the actual hardware response.

[0047] For example, with a brightness difference of 441 nits, the fitted cubic polynomial model is used to inversely deduce that the drive current needs to be adjusted from the current 45 mA to 60 mA, and the current adjustment amount is determined to be 15 mA. Next, it is verified whether the target current after the drive current adjustment is within the preset safe range. If it is, dynamic brightness adjustment parameters are generated. The safe range is set according to the rated parameters of the information board hardware, ranging from 20 mA to 80 mA. This range is determined by the rated operating current provided by the hardware manufacturer to ensure that the current is not too high, causing the device to overheat and age, nor too low, preventing the display panel from emitting light normally.

[0048] The generation process of dynamic brightness adjustment parameters is as follows: First, the drive current value is determined based on the target current derived from the nonlinear model. Then, the adjustment rate is set according to the hardware capacity of the information board. The adjustment rate is based on the maximum current change rate in the hardware manual to avoid sudden changes in brightness or device damage caused by sudden current changes. It can be set to 50 nits / second, and the corresponding current change rate is found to be 1 mA / second in the equipment manual. Finally, the parameters are formatted according to the information board control protocol, including key information such as target current, adjustment rate, and adjustment duration. For example, if the target current of 60 mA is within the safe range, the generated dynamic brightness adjustment parameters are a drive current of 60 mA, an adjustment rate of 50 nits / second, and an adjustment duration of 8.82 seconds, ensuring that the parameters meet the requirements of the control protocol.

[0049] Finally, parameters are transmitted via a wired communication interface RS485 or a wireless communication module LoRa. The control unit receives and analyzes these parameters, then gradually adjusts the drive current according to the set adjustment rate, smoothly increasing the brightness of the backlight module or display panel to avoid sudden brightness changes that could cause visual discomfort to the driver. For example, after receiving the parameters, the control unit gradually increases the current from the current 45 mA to 60 mA at a rate of 1 mA / s, corresponding to a brightness increase from 450 nits to 891 nits at a rate of 50 nits / s. The entire adjustment process is smooth and without fluctuations.

[0050] In step S107, acquiring the adjusted display brightness data and the real-time visibility feedback signal, and calculating the confidence score of the display brightness data and the feedback signal, includes: Acquire adjusted display brightness data and real-time visibility feedback signals; The display brightness data and the feedback signal are mapped into a multi-dimensional feature vector to construct a current state sample; Calculate the similarity between the current state sample and the preset ideal display state sample, and convert the similarity into a confidence score.

[0051] It should be noted that the adjusted display brightness data is collected by the built-in brightness sensor on the information board, with an accuracy of ±1 nit, reflecting the actual luminous intensity after adjustment in real time. Real-time visibility feedback signals are obtained by capturing the information board display from the driver's perspective using a high-definition camera and analyzing it using the Laplacian variance algorithm. This algorithm first converts the color image to grayscale to reduce the impact of color interference on sharpness judgment, and then calculates the variance of the Laplacian operator on the grayscale image. The larger the variance value, the sharper the image edges, the richer the details, and the better the visibility. The quantization rule is to normalize the Laplacian variance value to a scoring range of 0-100 using min-max normalization, with the minimum variance corresponding to 0 points (completely invisible) and the maximum variance corresponding to 100 points (completely clear). This mapping range is determined by calibrating image sharpness data under different lighting scenarios over the past year. For example, after processing, the Laplacian variance of the display image acquired after brightness adjustment is 800, and after normalization, a visibility score of 92 points is obtained, providing basic data for subsequent feature mapping.

[0052] It is worth noting that the multidimensional feature vector includes four dimensions: display brightness, visibility score, ambient light intensity, and light change rate. The data for each dimension are normalized to the [0,1] interval using min-max normalization. The current state sample is a four-dimensional feature vector that comprehensively represents the adjusted display effect and environmental adaptation. For example, the display brightness of 890 nits is normalized to 0.89, the visibility score of 92 is normalized to 0.92, the ambient light intensity of 18000 lx is normalized to 0.9, and the light change rate of 700 lx / second is normalized to 0.7. The constructed current state sample is [0.89, 0.92, 0.9, 0.7].

[0053] Finally, the construction process for the ideal display state sample involved recruiting subjects of different ages (youth, middle-aged, and elderly) covering different visual levels. In a laboratory environment, scenarios with varying ambient light intensities and glare types were simulated. Subjects adjusted the brightness of the information board and their viewing angle to achieve the most comfortable subjective visual experience and clearest information retrieval. Display brightness, visibility score, ambient light intensity, and rate of light change were recorded for each comfortable state. After removing extreme outliers, the statistical mean of each dimension was calculated, and then the mean was normalized to the [0,1] interval using min-max normalization. The final ideal display state sample was determined to be [0.9, 0.95, 0.85, 0.6]. Similarity was calculated using the cosine similarity algorithm, which measures the angle between two vectors to reflect the degree of feature similarity, with a value range of [0,1]. The confidence score was converted to a score of 0-100 based on a factor of 100 of the similarity score; a higher score indicates a better adjustment effect. For example, the cosine similarity between the current state sample and the ideal sample is 0.93, and the confidence score after conversion is 93, indicating that the brightness adjustment effect is good.

[0054] In step S108, if the confidence score reaches a preset effective judgment threshold, the dynamic brightness adjustment parameter remains unchanged; if it does not reach the threshold, the dynamic brightness adjustment parameter is iteratively optimized until the information board display is clear, including: If the confidence score reaches the preset effective judgment threshold, the dynamic brightness adjustment parameters remain unchanged; If the confidence score does not reach the preset effective judgment threshold, the deviation feature component that caused the failure to meet the standard is extracted and the offset of brightness compensation is calculated. The offset is added to the dynamic brightness adjustment parameters to generate the fine-tuned brightness adjustment parameters; The fine-tuned brightness adjustment parameters are sent to the control terminal of the information board to drive the information board to perform brightness fine-tuning, and obtain the fine-tuned display brightness data and real-time visibility feedback signal. The confidence score is recalculated based on the display brightness data and the feedback signal; If the confidence score reaches the preset effective judgment threshold, the fine-tuned brightness adjustment parameters are maintained; otherwise, ambient light data is re-collected and the dynamic brightness adjustment parameters are iteratively optimized.

[0055] It should be noted that the specific rules for setting the effective threshold are based on visual comfort experimental data from different scenarios over the past year. The experiments covered common scenarios such as urban main roads, highways, and suburban roads, recruiting participants of different ages to record display effect scores under comfortable conditions, resulting in a large sample size. The basic threshold is set at the 80th percentile of the sample, corresponding to 85 points. At this quantile, 90% of comfortable scenarios score higher than this value, balancing stringency and practicality. High-precision scenarios such as highways have higher requirements for display clarity, so the 85th percentile is set at 90 points. Suburban road scenarios have less light interference, so the 75th percentile is set at 80 points. The threshold can be manually fine-tuned based on the lighting characteristics of different areas and maintenance feedback to adapt to diverse road environments. If the score reaches or exceeds the threshold, the parameter is considered effective; otherwise, optimization is required. For example, a confidence score of 92 points exceeds the basic threshold of 85 points, validating the effectiveness of the current dynamic brightness adjustment parameters.

[0056] If the confidence score reaches the valid judgment threshold, the dynamic brightness adjustment parameters remain unchanged, the system keeps the existing drive current, adjustment rate and other parameters unchanged, and continuously monitors the ambient light and display effect to ensure that the information board displays stably under the current lighting conditions.

[0057] If the confidence score does not reach the effective decision threshold, the random forest algorithm is used to extract the biased feature components that lead to low effectiveness and calculate the brightness compensation offset. The random forest algorithm constructs 100 independent CART decision trees. Each tree draws samples from the training set based on bootstrap sampling, randomly selects features for training according to the square root rule, and does not prune to preserve model diversity. Finally, the prediction results of all trees are integrated through voting, achieving both high accuracy and resistance to overfitting. This algorithm quantifies the influence weight of each feature dimension on the confidence score, which is used here to identify the core dimensions of difference between the current state sample and the ideal sample.

[0058] The specific process is as follows: First, input the four-dimensional data (display brightness, visibility score, ambient light intensity, and rate of change of illumination) of the current state sample and the ideal sample into a trained random forest model. The training set contains over 50,000 samples from different lighting scenarios over the past year. Each sample contains four-dimensional normalized features and a binary label indicating whether the confidence level meets the standard. Key parameters are set as follows: 100 decision trees, 5 minimum leaf node samples, and unlimited maximum depth. The Gini coefficient is used as the loss function. The model outputs the feature importance weights for each dimension: display brightness 0.4, visibility score 0.3, ambient light intensity 0.2, and rate of change of illumination 0.1, with the sum of all weights being 1. Then, calculate the deviation value for each dimension (current value minus ideal value), multiply the deviation value by the corresponding weight, and sum them by weight to obtain the comprehensive deviation coefficient. Finally, calculate the brightness compensation offset based on the confidence difference, with the formula adjusted as follows: ΔBc=(Tv-Sc)×Kc×(1+W); Where ΔBc represents the brightness compensation offset, Tv represents the effective judgment threshold, Sc represents the current confidence score, Kc represents the experimentally calibrated basic compensation coefficient of 5 nits / minute, and W represents the comprehensive deviation coefficient.

[0059] It is worth noting that Kc's experimental calibration method involved recruiting 30 subjects of different ages and recording the amount of luminance compensation required to achieve the confidence score and the score difference under 10 typical glare intensities. The relationship between the two was fitted by linear regression, and the slope was finally determined to be 5 nits / min.

[0060] This formula considers both the confidence gap and the influence of differences in features across various dimensions, avoiding the limitations of a single linear correction. For example, with an effective judgment threshold Tv = 85 points, a current confidence score Sc = 78 points, and a comprehensive deviation coefficient W = 0.15, substituting these values ​​into the formula yields ΔBc = (85-78) × 5 × (1 + 0.15) = 7 × 5 × 1.15 = 40.25 nits, accurately compensating for insufficient brightness caused by multi-dimensional deviations.

[0061] Next, the offset is superimposed on the original dynamic brightness adjustment parameters to generate fine-tuned brightness adjustment parameters. First, the target brightness is updated based on the brightness compensation offset. The original target brightness is then increased by adding the offset; if the overall deviation coefficient is positive, the brightness is increased, and if it is negative, the brightness is decreased, ensuring the target brightness matches actual needs. Then, the updated target brightness is input into the previously calibrated cubic polynomial nonlinear model. This model accurately represents the nonlinear relationship between brightness and drive current, allowing for the derivation of the corresponding drive current value. During fine-tuning, the adjustment rate is set to 20 nits / second, corresponding to a current change rate of 0.5 mA / second, avoiding sudden brightness changes that could affect the visual experience, while ensuring the parameters remain within the safe operating range of the information board: drive current 20-80 mA and brightness 50-1000 nits. For example, if the original target brightness is 891 nits, after adding a 40.25 nit offset, it is updated to 931.25 nits. Inputting this into the nonlinear model, the drive current needs to be adjusted to 65 mA, generating fine-tuned brightness adjustment parameters of a drive current of 65 mA and an adjustment rate of 20 nits / second.

[0062] Then, the confidence score is recalculated. If it still does not meet the standard, ambient light data is collected again at the same frequency as the initial collection. The glare intensity, adjustment range, etc. are recalculated according to steps S101 to S107. The process is iterated until the confidence score reaches the effective judgment threshold.

[0063] In summary, this invention discloses an intelligent light-sensing adjustment method for road variable message signs, which can achieve precise light-sensing adjustment of road variable message signs under complex lighting conditions, meeting the high-quality requirements of display stability and environmental adaptability of the signs.

[0064] Reference Figure 2 The second embodiment of the present invention provides an intelligent light-sensing adjustment system for road variable message signs, comprising: The data acquisition module is used to acquire ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data. The interference analysis module is used to analyze the light characteristics and construct the light interference vector based on the ambient light data and the surface reflection data. By classifying and quantizing the position projection of the light interference vector, the type and intensity value of the light interference are obtained. The index determination module is used to fuse the light interference type, the intensity value, the current display brightness data and the light change data, analyze the correlation pattern between the fused data and the pre-acquired historical light data, and determine the glare intensity index. The coefficient calculation module is used to collect the spectral distribution data of the current environment, filter effective spectral features and determine the source of glare if the glare intensity index exceeds the preset glare judgment threshold, and calculate the glare influence coefficient by weighting the effective spectral features and the weight coefficients corresponding to the glare source. The target determination module is used to calculate the brightness adjustment range based on the glare influence coefficient and a pre-acquired visual comfort benchmark, and determine the brightness target value based on the adjustment range. The parameter adjustment module is used to compare the target brightness value with the existing brightness of the information board. If the brightness difference exceeds the preset brightness difference threshold, dynamic brightness adjustment parameters are generated, and the brightness of the information board is adjusted according to the dynamic brightness adjustment parameters. The feedback scoring module is used to acquire the adjusted display brightness data and the real-time visibility feedback signal, and to calculate the confidence score of the display brightness data and the feedback signal. The verification and optimization module is used to keep the dynamic brightness adjustment parameters unchanged if the confidence score reaches the preset effective judgment threshold, and to iteratively optimize the dynamic brightness adjustment parameters if the confidence score does not reach the threshold, until the information board is clearly displayed.

[0065] It should be noted that the intelligent light-sensing adjustment system for road variable message signs provided in this embodiment of the invention is used to execute all the process steps of the intelligent light-sensing adjustment method for road variable message signs in the above embodiment. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0066] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a parameter adjustment program. When the processor executes the computer program, it implements the steps described in the various embodiments of intelligent light-sensing adjustment for variable message signs, for example... Figure 1 The step S101 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above system embodiments, such as the parameter adjustment module.

[0067] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0068] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0069] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0070] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0071] Wherein, if the modules / units integrated in the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or system capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0072] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0073] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for intelligent light-sensing adjustment of road variable message signs, characterized in that, include: Acquire ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data; Based on the ambient light data and the surface reflection data, the light characteristics are analyzed and a light interference vector is constructed. By classifying and quantizing the position projection of the light interference vector, the light interference type and intensity value are obtained. By integrating the light interference type, the intensity value, the current display brightness data, and the light change data, the correlation pattern between the integrated data and the pre-acquired historical light data is analyzed to determine the glare intensity index. If the glare intensity index exceeds the preset glare judgment threshold, the spectral distribution data of the current environment is collected, effective spectral features are screened and the glare source is determined, and the glare influence coefficient is obtained by weighting the effective spectral features and the weight coefficients corresponding to the glare source. Based on the glare impact coefficient and the pre-acquired visual comfort benchmark, the brightness adjustment range is calculated, and the target brightness value is determined based on the adjustment range. Compare the target brightness value with the existing brightness of the information board. If the brightness difference exceeds a preset brightness difference threshold, generate dynamic brightness adjustment parameters and adjust the brightness of the information board according to the dynamic brightness adjustment parameters. Acquire the adjusted display brightness data and the real-time visibility feedback signal, and calculate the confidence score of the display brightness data and the feedback signal; If the confidence score reaches the preset effective judgment threshold, the dynamic brightness adjustment parameter remains unchanged; otherwise, the dynamic brightness adjustment parameter is iteratively optimized until the information board display is clear.

2. The intelligent light-sensing adjustment method for road variable message signs according to claim 1, characterized in that, The acquisition of ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data includes: Collect ambient light radiation intensity and surface light spot distribution data of the information board, and process the ambient light radiation intensity and surface light spot distribution data using the Otsu algorithm to obtain ambient light data and surface reflection data of the information board. Read the current display brightness data of the information board; Continuous ambient light data is collected at preset fixed time intervals, and the ratio of the difference between adjacent data points of the ambient light data to the time interval is calculated to obtain the illumination change data of the surrounding environment.

3. The intelligent light-sensing adjustment method for road variable message signs according to claim 1, characterized in that, The process involves analyzing light characteristics and constructing a light interference vector based on the ambient light data and surface reflection data. By classifying and quantizing the light interference vector through position projection, the type and intensity value of the light interference are obtained, including: The wavelength and incident angle characteristics of light rays are analyzed from the ambient light data, and the area of ​​light spot and brightness distribution characteristics are extracted from the surface reflection data. By combining the light wavelength, the incident angle characteristics, the light spot area, the brightness distribution characteristics, as well as the acquisition time and the installation location information of the information board, a light interference vector containing spatiotemporal dimensions is constructed. Calculate the distance between the illumination interference vector and the boundary of the preset interference category, and determine the light interference type based on the illumination interference vector with the smallest distance; The illumination interference vector is projected into a high-dimensional space, and the distance between the projection position and the preset intensity reference point is calculated to obtain the intensity value.

4. The intelligent light-sensing adjustment method for road variable message signs according to claim 3, characterized in that, The process involves fusing the light interference type, the intensity value, the current display brightness data, and the illumination change data, analyzing the correlation pattern between the fused data and pre-acquired historical illumination data, and determining the glare intensity index, including: The light interference type, the intensity value, the current display brightness data, and the illumination change data are normalized to form a feature matrix of real-time illumination. Calculate the similarity between the feature matrix and the pre-acquired historical illumination data, and use the mean of the similarity as the correlation strength. If the correlation strength is within a preset glare sensitivity range, calculate the initial glare intensity index based on the correlation strength. The initial glare intensity index is verified based on the light spot area and the brightness distribution characteristics, and the glare intensity index is obtained after correction.

5. The intelligent light-sensing adjustment method for road variable message signs according to claim 1, characterized in that, If the glare intensity index exceeds a preset glare determination threshold, then the spectral distribution data of the current environment is collected, effective spectral features are screened and the glare source is determined. The glare influence coefficient is calculated by weighting the effective spectral features and the weighting coefficients corresponding to the glare source, including: If the glare intensity index exceeds the preset glare determination threshold, then collect the spectral distribution data of the current environment; The spectral distribution data is filtered and effective spectral features are selected. The effective spectral features are compared with a preset spectral feature library to determine the source of glare. The effective spectral features are mapped to a preset weight matrix, and the weight coefficients corresponding to the glare sources are combined to obtain the glare influence coefficient by weighted summation.

6. The intelligent light-sensing adjustment method for road variable message signs according to claim 1, characterized in that, The step of calculating the brightness adjustment range based on the glare impact coefficient and a pre-acquired visual comfort benchmark, and determining the target brightness value based on the adjustment range, includes: The glare impact coefficient is mapped to a pre-acquired visual comfort benchmark to obtain the basic brightness adjustment ratio; The basic brightness adjustment ratio is corrected based on the illumination change data, and the brightness adjustment range is calculated. Based on the adjustment range and the current display brightness data, combined with the preset maximum and minimum brightness limits of the information board, the target brightness value is determined.

7. The intelligent light-sensing adjustment method for road variable message signs according to claim 1, characterized in that, The process of comparing the target brightness value with the existing brightness of the information board, and if the brightness difference exceeds a preset brightness difference threshold, generates dynamic brightness adjustment parameters, and adjusts the brightness of the information board according to the dynamic brightness adjustment parameters, including: Calculate the difference between the target brightness value and the existing brightness of the information board to obtain the brightness difference value; If the brightness difference value exceeds the preset brightness difference threshold, the drive current gain is determined by combining the pre-acquired brightness response characteristics of the information board. If the drive current gain is within a preset safe range, dynamic brightness adjustment parameters are generated; The dynamic brightness adjustment parameters are sent to the control terminal of the information board to drive the information board to perform brightness adjustment.

8. The intelligent light-sensing adjustment method for road variable message signs according to claim 1, characterized in that, The process of acquiring the adjusted display brightness data and the real-time visibility feedback signal, and calculating the confidence score of the display brightness data and the feedback signal, includes: Acquire adjusted display brightness data and real-time visibility feedback signals; The display brightness data and the feedback signal are mapped into a multi-dimensional feature vector to construct a current state sample; Calculate the similarity between the current state sample and the preset ideal display state sample, and convert the similarity into a confidence score.

9. The intelligent light-sensing adjustment method for road variable message signs according to claim 1, characterized in that, If the confidence score reaches a preset effective judgment threshold, the dynamic brightness adjustment parameters remain unchanged; if not, the dynamic brightness adjustment parameters are iteratively optimized until the information board display is clear, including: If the confidence score reaches the preset effective judgment threshold, the dynamic brightness adjustment parameters remain unchanged; If the confidence score does not reach the preset effective judgment threshold, the deviation feature component that caused the failure to meet the standard is extracted and the offset of brightness compensation is calculated. The offset is added to the dynamic brightness adjustment parameters to generate the fine-tuned brightness adjustment parameters; The fine-tuned brightness adjustment parameters are sent to the control terminal of the information board to drive the information board to perform brightness fine-tuning, and obtain the fine-tuned display brightness data and real-time visibility feedback signal. The confidence score is recalculated based on the display brightness data and the feedback signal; If the confidence score reaches the preset effective judgment threshold, the fine-tuned brightness adjustment parameters are maintained; otherwise, ambient light data is re-collected and the dynamic brightness adjustment parameters are iteratively optimized.

10. A smart light-sensing adjustment system for road variable message signs, characterized in that, include: The data acquisition module is used to acquire ambient light data, surface reflection data of the information board, current display brightness data of the information board, and ambient light change data. The interference analysis module is used to analyze the light characteristics and construct the light interference vector based on the ambient light data and the surface reflection data. By classifying and quantizing the position projection of the light interference vector, the type and intensity value of the light interference are obtained. The index determination module is used to fuse the light interference type, the intensity value, the current display brightness data and the light change data, analyze the correlation pattern between the fused data and the pre-acquired historical light data, and determine the glare intensity index. The coefficient calculation module is used to collect the spectral distribution data of the current environment, filter effective spectral features and determine the source of glare if the glare intensity index exceeds the preset glare judgment threshold, and calculate the glare influence coefficient by weighting the effective spectral features and the weight coefficients corresponding to the glare source. The target determination module is used to calculate the brightness adjustment range based on the glare influence coefficient and a pre-acquired visual comfort benchmark, and determine the brightness target value based on the adjustment range. The parameter adjustment module is used to compare the target brightness value with the existing brightness of the information board. If the brightness difference exceeds the preset brightness difference threshold, dynamic brightness adjustment parameters are generated, and the brightness of the information board is adjusted according to the dynamic brightness adjustment parameters. The feedback scoring module is used to acquire the adjusted display brightness data and the real-time visibility feedback signal, and to calculate the confidence score of the display brightness data and the feedback signal. The verification and optimization module is used to keep the dynamic brightness adjustment parameters unchanged if the confidence score reaches the preset effective judgment threshold, and to iteratively optimize the dynamic brightness adjustment parameters if the confidence score does not reach the threshold, until the information board is clearly displayed.

Images