Vehicle-mounted DMS camera light supplement lamp brightness self-adaptive adjusting system

Abstract

The invention belongs to the technical field of vehicle-mounted DMS control, and discloses a vehicle-mounted DMS camera light supplement lamp brightness self-adaptive adjustment system, which comprises a data acquisition module used for acquiring ambient light data and a figure image; the data processing module is used for removing invalid data from the ambient light data to obtain valid light data, and screening the figure image to obtain an optimal image; the influence factor module is used for obtaining a light temperature coefficient based on the constructed light intensity influence coefficient model; the light adjusting module is used for carrying out light calculation based on the effective light data and the light temperature coefficient to obtain ambient light intensity, carrying out graph analysis based on the optimal image to obtain optimal ambient light intensity, and carrying out light adjustment based on the ambient light intensity and the optimal ambient light intensity; therefore, the monitoring precision and the driving safety of the DMS are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle-mounted DMS control, and more specifically, to a vehicle-mounted DMS camera fill light brightness adaptive adjustment system. Background Art

[0002] The patent application with the publication number CN112261309A discloses a shooting method and system based on an IMS camera and a DMS camera, wherein the IMS camera is a rolling shutter, and the DMS camera includes an infrared fill light, wherein the infrared fill light is used to fill light for the shooting area of ​​the DMS camera, and the shooting area of ​​the IMS camera and the shooting area of ​​the DMS camera include a part of overlap, and the shooting method includes: setting the IMS camera and the DMS camera to the same shooting frame rate; exposing the IMS camera and the DMS camera in a cycle of the shooting frame rate according to a preset rule, wherein the preset rule is that the exposure of the DMS camera is synchronized with the fill light of the infrared fill light, and the exposure time of the DMS camera and the exposure time of the blank invalid line of the IMS camera include an overlapping time; the images shot by the IMS camera and the DMS camera are more fidelity, and the equipment has a long service life and low maintenance cost.

[0003] Under the existing technology, the fill light equipped with a DMS camera generally uses a fixed voltage for POC power supply, and the power that can be transmitted is fixed. However, when the ambient light changes drastically, the fill light will be underpowered and the camera cannot obtain the best image quality. Moreover, when the output voltage is fixed, the higher the input voltage, the lower the conversion efficiency, the more heat is converted into heat, and the greater the heat generated when the camera is working. The longer the semiconductor device works in a higher temperature environment, the shorter its life will be.

[0004] In view of this, the present invention proposes a vehicle-mounted DMS camera fill light brightness adaptive adjustment system to solve the above-mentioned problem. Summary of the invention

[0005] In order to overcome the above defects of the prior art and to achieve the above purpose, the present invention provides the following technical solutions: a vehicle-mounted DMS camera fill light brightness adaptive adjustment system, comprising: a data acquisition module: collecting ambient light data and character images; Data processing module: remove invalid data from ambient light data to obtain valid light data, and screen the character images to obtain the most suitable image; Impact factor module: obtain the light temperature coefficient based on the constructed light intensity impact coefficient model; Lighting adjustment module: performs light calculation based on effective light data and light temperature coefficient to obtain ambient light intensity, performs graph analysis based on the most suitable image to obtain the optimal ambient light intensity, and adjusts the lighting based on the ambient light intensity and the optimal ambient light intensity.

[0006] Furthermore, the ambient light data includes: light intensity, spectral distribution, ambient temperature, light wavelength, illuminated area and reflectivity.

[0007] Furthermore, the method of removing invalid data includes: The ambient light data is normalized by the Min-Max normalization algorithm to obtain normalized light data, and the normalized light data is vectorized to obtain the ambient light vector. Calculate the data points Other data points and data points The similarity is calculated as follows: ;in, Representative The ambient light vector is The similarity of the ambient light vectors, represents the bandwidth parameter, represents the base of natural logarithms, Representative Ambient light vectors, Representative Ambient light vectors, and represents the index of the ambient light vector, and ; Constructing an undirected similarity graph based on similarity ,in, represents the ambient light vector, Represents a set of edges; preset similarity threshold , and define the edge weight, the weight calculation formula is: ;in, Representative The ambient light vector is The weight of the edge connecting the ambient light vectors; For the ambient light vector in the undirected similarity graph , all with The indices of other ambient light vectors that have directly connected edges Neighborhood index set , the ambient light vector corresponding to the index in the neighborhood index set and the ambient light vector Construct a local point set, and count the total number of edges that are actually connected to each ambient light vector in the local point set based on the undirected similarity graph , based on the total number of Calculating the ambient light vector The clustering coefficient is calculated as follows: ;in Representative The clustering coefficient of the ambient light vectors, Represents the total amount of data in the neighborhood index set; The basic outlier degree of each data point in the ambient light vector is calculated based on the degree of the ambient light vector in the undirected similarity graph. The calculation formula of the basic outlier degree is: ;in, Representative The basic outlier of the ambient light vector, No. The degree of the ambient light vector, Represents the maximum degree of the ambient light vector in the undirected similarity graph; The local outlier degree of each data point in the ambient light vector is calculated based on the neighborhood index set and the clustering coefficient. The calculation formula of the local outlier degree is: ;in, Representative The local outlier of the ambient light vector, Representative The neighborhood index set of ambient light vectors, Representative A similarity list of ambient light vectors; The comprehensive outlier degree is calculated based on the basic outlier degree and the local outlier degree. The formula for calculating the comprehensive outlier degree is: ;in, Representative The comprehensive outlier degree of the ambient light vectors; the abnormal threshold is set based on the comprehensive outlier degree ,in, represents the abnormal threshold, represents the mean of the comprehensive outlier degree, Represents the standard deviation of the comprehensive outlier degree; the ambient light data corresponding to the ambient light vector whose comprehensive outlier degree is greater than or equal to the abnormal threshold is marked as abnormal data, and all ambient light data not marked as abnormal data is regarded as valid light data.

[0008] Furthermore, the step of performing graph screening includes: Step 1: preset K groups of random person images, each group of person images is represented as a chromosome, and the genes in the chromosome correspond to the pixels in the person image; Step 2: For each chromosome, the character image is gray-processed by the quantization method to obtain a gray-scale image. The fitness of the chromosome is evaluated based on the gray-scale image. The calculation formula for fitness evaluation is: ;in, represents the fitness of the chromosome, represents the number of pixels of the chromosome, Represents the chromosome The gray value of a pixel, Represents the mean gray value of each pixel in the chromosome, Represents the index of the pixel in the chromosome; the parent chromosome is selected by elite selection method according to the fitness value; Step 3: Compare the genes in the parent chromosome one by one according to the grayscale value, use the parent chromosome as a template, exchange the genes with high grayscale values ​​in the mother chromosome with the genes in the parent chromosome to generate the daughter chromosome; Step 4: Using a random mutation algorithm, randomly select U genes in the offspring chromosome to replace them with genes at corresponding positions in the pre-stored chromosome; Step 5, evaluate the fitness of the offspring chromosome. When the fitness value of the offspring chromosome is greater than the fitness value of the pre-stored chromosome, replace the pre-stored chromosome with the offspring chromosome as the new pre-stored chromosome. When the maximum number of iterations has not been reached and the fitness value of the pre-stored chromosome is less than the preset fitness threshold, return to step 2 to continue iterating. When the maximum number of iterations is reached or the fitness value of the pre-stored chromosome is greater than the preset fitness threshold, stop the iteration and output the pre-stored chromosome as the optimal image.

[0009] Furthermore, the method for obtaining the light temperature coefficient includes: Collect group D of historical environmental data, including historical ambient temperature, historical measured light intensity and historical predicted light intensity. Based on the historical environmental data, the ratio of historical measured light intensity to historical predicted light intensity is used as the historical actual light temperature coefficient, the historical ambient temperature is used as the input of the light intensity influence coefficient model, the predicted light temperature coefficient is used as the output of the light intensity influence coefficient model, and the historical actual light temperature coefficient is used as the prediction target of the light intensity influence coefficient model for model training. The light intensity influence coefficient model is constructed. The light intensity influence coefficient model is a linear regression model. The target loss function of the light intensity influence coefficient model is: ;in, is the mean square error, is the total amount of historical environmental data, No. The historical actual light temperature coefficient, It is The predicted light temperature coefficient is predicted; the training goal is to minimize the value of the target loss function and obtain the light intensity influence coefficient model with the minimum value of the target loss function; Based on the constructed light intensity influence coefficient model, the ambient temperature is used as the input of the light intensity influence coefficient model for prediction to obtain the light temperature coefficient.

[0010] Furthermore, the method of performing light calculation includes: The light calculation function is defined according to the light temperature coefficient. The formula of the light calculation function is: ,in, Represents the ambient light intensity, represents the light intensity, Represents the minimum value of the wavelength of light, represents the maximum value of the wavelength of light, represents the wavelength of light, represents the central light wavelength, represents the spectral distribution, represents the standard deviation of the spectral distribution, represents the light temperature coefficient, represents the reflectivity, Representatives Perform integration.

[0011] Furthermore, the method of performing graph analysis includes: The brightness of the optimal image is calculated to obtain the optimal brightness, and the light is reversed based on the optimal brightness to obtain the optimal ambient light intensity.

[0012] Furthermore, the method for obtaining the optimum brightness includes: The brightness conversion of the optimal image is performed, and the formula for brightness conversion is: ;in, Represents the best image The brightness of a pixel, Represents the best image The red component of the pixel, Represents the best image The blue component of pixels, Represents the best image The green component of pixels, Represents the best image pixels, , and Represents the weighting coefficient; the brightness of all pixels in the optimal image is averaged to obtain the optimal brightness.

[0013] Furthermore, the method for obtaining the optimal ambient light intensity includes: Based on the optimum brightness, light inversion is performed, and the calculation formula for light inversion is: ;in, Represents the optimal ambient light intensity, represents the irradiated area, Represents the angle between the irradiated light and the irradiated area, Represents the minimum angle between the irradiated light and the irradiated area. Represents the maximum angle between the irradiated light and the irradiated area. represents the reflectivity, Representatives Perform integration.

[0014] Furthermore, the lighting adjustment method includes: According to the comparison between the optimal ambient light intensity and the ambient light intensity, when When the light is not adjusted, Calculate the fill light intensity when , the power is calculated based on the fill light intensity, and the formula for power calculation is: ;in, represents the required power, Represents the light effect of the light source. Represents the area of ​​the irradiated area; based on the required power, the required voltage is obtained according to the voltage power calculation formula. The voltage power calculation formula is: ;in, represents the required voltage, Represents a fixed current; the voltage is adjusted according to the required voltage, thereby adjusting the light intensity.

[0015] The technical effects and advantages of the adaptive brightness adjustment system for the vehicle-mounted DMS camera fill light of the present invention are as follows: The present invention ensures the accuracy and reliability of ambient light data by eliminating invalid data, effectively eliminating the influence of noise and outliers; automatically identifies and retains the most suitable images for subsequent analysis through image screening, thereby improving the accuracy of image analysis; accurately predicts the light temperature coefficient according to the ambient temperature, making the light intensity calculation more scientific and practical, and effectively coping with the light intensity changes under different temperature conditions; automatically adjusts the brightness of the fill light through real-time calculation of the ambient light intensity and the optimal ambient light intensity, and accurately controls the light intensity, thereby ensuring the shooting quality, effectively reducing energy consumption, and extending the life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a vehicle-mounted DMS camera fill light brightness adaptive adjustment system of the present invention; Figure 2 The present invention is a schematic diagram of a method for adaptively adjusting the brightness of a vehicle-mounted DMS camera fill light. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Embodiment 1

[0018] See also Figure 1 As shown, the vehicle-mounted DMS camera fill light brightness adaptive adjustment system described in this embodiment includes: a data acquisition module: collecting ambient light data and character images; Data processing module: remove invalid data from ambient light data to obtain valid light data, and screen the character images to obtain the most suitable image; Impact factor module: obtain the light temperature coefficient based on the constructed light intensity impact coefficient model; Lighting adjustment module: performs light calculation based on effective light data and light temperature coefficient to obtain ambient light intensity, performs image analysis based on the most suitable image to obtain the optimal ambient light intensity, and adjusts the lighting based on the ambient light intensity and the optimal ambient light intensity; Each module is connected by wired and / or wireless means to achieve data transmission between modules; Ambient light data includes: light intensity, spectral distribution, ambient temperature, light wavelength, illuminated area and reflectivity; Ways to remove invalid data from ambient light data include: The ambient light data is normalized by the Min-Max normalization algorithm to obtain normalized light data, and the normalized light data is vectorized to obtain the ambient light vector. Calculate the data points Other data points and data points The similarity is calculated as follows: ;in, Representative The ambient light vector is The similarity of the ambient light vectors, Represents the bandwidth parameter, which is used to determine the attenuation range of similarity. represents the base of natural logarithms, Representative Ambient light vectors, Representative Ambient light vectors, and represents the index of the ambient light vector, and ; Constructing an undirected similarity graph based on similarity ,in, represents the ambient light vector, Represents a set of edges; preset similarity threshold , and define the edge weight, the weight calculation formula is: ;in, Representative The ambient light vector is The weight of the edge connecting the ambient light vectors; For the ambient light vector in the undirected similarity graph , all with The indices of other ambient light vectors that have directly connected edges Neighborhood index set , the ambient light vector corresponding to the index in the neighborhood index set and the ambient light vector Construct a local point set, and count the total number of edges that are actually connected to each ambient light vector in the local point set based on the undirected similarity graph , based on the total number of Calculating the ambient light vector The clustering coefficient is calculated as follows: ;in Representative The clustering coefficient of the ambient light vectors, Represents the total amount of data in the neighborhood index set; The basic outlier degree of each data point in the ambient light vector is calculated based on the degree of the ambient light vector in the undirected similarity graph (the number of edges directly connected to the ambient light vector). The calculation formula of the basic outlier degree is: ;in, Representative The basic outlier of the ambient light vector, No. The degree of the ambient light vector, Represents the maximum degree of the ambient light vector in the undirected similarity graph; The local outlier degree of each data point in the ambient light vector is calculated based on the neighborhood index set and the clustering coefficient. The calculation formula of the local outlier degree is: ;in, Representative The local outlier of the ambient light vector, Representative The neighborhood index set of ambient light vectors, Representative A similarity list of ambient light vectors; The comprehensive outlier degree is calculated based on the basic outlier degree and the local outlier degree. The formula for calculating the comprehensive outlier degree is: ;in, Representative The comprehensive outlier degree of the ambient light vectors; the abnormal threshold is set based on the comprehensive outlier degree ,in, represents the abnormal threshold, represents the mean of the comprehensive outlier degree, Represents the standard deviation of the comprehensive outlier degree; marks the ambient light data corresponding to the ambient light vector whose comprehensive outlier degree is greater than or equal to the abnormal threshold as abnormal data, and regards all ambient light data not marked as abnormal data as valid light data; ensures the accuracy and reliability of ambient light data and eliminates the influence of noise and outliers; The steps of image screening for person images include: Step 1: preset K groups of random person images, each group of person images is represented as a chromosome, and the genes in the chromosome correspond to the pixels in the person image; Step 2: For each chromosome, the character image is gray-processed by the quantization method to obtain a gray-scale image. The fitness of the chromosome is evaluated based on the gray-scale image. The calculation formula for fitness evaluation is: ;in, represents the fitness of the chromosome, represents the number of pixels of the chromosome, Represents the chromosome The gray value of a pixel, Represents the mean gray value of each pixel in the chromosome, Represents the index of the pixel in the chromosome; the chromosome with a larger fitness value is better, and the parent chromosome is selected by the elite selection method based on the fitness value; Step 3: Compare the genes in the parent chromosome one by one according to the grayscale value, use the parent chromosome as a template, exchange the genes with high grayscale values ​​in the mother chromosome with the genes in the parent chromosome to generate the daughter chromosome; Step 4: Randomly select U genes in the offspring chromosome for replacement through a random mutation algorithm, and replace them with genes at corresponding positions in the pre-stored chromosome (the preset initial optimal chromosome); Step 5, evaluate the fitness of the offspring chromosomes. When the fitness value of the offspring chromosomes is greater than the fitness value of the pre-stored chromosomes, replace the pre-stored chromosomes with the offspring chromosomes as new pre-stored chromosomes; when the maximum number of iterations is not reached and the fitness value of the pre-stored chromosomes is less than the preset fitness threshold, return to step 2 to continue iterating; when the maximum number of iterations is reached or the fitness value of the pre-stored chromosomes is greater than the preset fitness threshold, stop the iteration and output the pre-stored chromosomes as the most suitable images; screen and optimize the character images through multiple iterations, automatically identify and retain the images most suitable for subsequent analysis, and improve the accuracy of image analysis; Ways to obtain the light temperature coefficient include: Collect group D of historical environmental data, including historical ambient temperature, historical measured light intensity and historical predicted light intensity. Based on the historical environmental data, the ratio of historical measured light intensity to historical predicted light intensity is used as the historical actual light temperature coefficient, the historical ambient temperature is used as the input of the light intensity influence coefficient model, the predicted light temperature coefficient is used as the output of the light intensity influence coefficient model, and the historical actual light temperature coefficient is used as the prediction target of the light intensity influence coefficient model for model training. The light intensity influence coefficient model is constructed. The light intensity influence coefficient model is a linear regression model. The target loss function of the light intensity influence coefficient model is: ;in, is the mean square error, is the total amount of historical environmental data, No. The historical actual light temperature coefficient, It is The predicted light temperature coefficient is predicted; the training goal is to minimize the value of the target loss function and obtain the light intensity influence coefficient model with the minimum value of the target loss function; Based on the constructed light intensity influence coefficient model, the ambient temperature is used as the input of the light intensity influence coefficient model for prediction to obtain the light temperature coefficient; the light temperature coefficient is accurately predicted according to the ambient temperature, making the light intensity calculation more scientific and practical, and effectively coping with the changes in light intensity under different temperature conditions; The light calculation is based on the light temperature coefficient. The formula for light calculation is: ,in, Represents the ambient light intensity, represents the light intensity, Represents the minimum value of the wavelength of light, represents the maximum value of the wavelength of light, represents the wavelength of light, represents the central wavelength (the median wavelength of light), represents the spectral distribution, represents the standard deviation of the spectral distribution, represents the light temperature coefficient, represents the reflectivity, Representatives Perform integration; Methods for graph analysis based on the best fit image include: The brightness conversion of the optimal image is performed, and the formula for brightness conversion is: ;in, Represents the best image The brightness of a pixel, Represents the best image The red component of the pixel, Represents the best image The blue component of pixels, Represents the best image The green component of pixels, Represents the best image pixels, , and Represents the weighting coefficient, which reflects the sensitivity of the human eye to different colors and is determined based on the international standard CIE1931 colorimetric space; the brightness of all pixels in the optimal image is averaged to obtain the optimal brightness; Based on the optimum brightness, light inversion is performed, and the calculation formula for light inversion is: ;in, Represents the optimal ambient light intensity, represents the irradiated area, represents the angle between the irradiated light and the irradiated area (obtained by actual measurement with an X-ray goniometer), Represents the minimum angle between the irradiated light and the irradiated area. Represents the maximum angle between the irradiated light and the irradiated area. represents the reflectivity, Representatives Perform integration; Ways to adjust lighting based on ambient light intensity and optimal ambient light intensity include: According to the comparison between the optimal ambient light intensity and the ambient light intensity, when When the light is not adjusted, Calculate the fill light intensity when , the power is calculated based on the fill light intensity, and the formula for power calculation is: ;in, represents the required power, Represents the light effect of the light source (equipment production information acquisition), Represents the area of ​​the irradiated area; based on the required power, the required voltage is obtained according to the voltage power calculation formula. The voltage power calculation formula is: ;in, represents the required voltage, Represents a fixed current; adjusts the voltage according to the required voltage, thereby adjusting the light intensity; automatically adjusts the brightness of the fill light according to the comparison between the ambient light intensity and the optimal ambient light intensity. The precise control of the light intensity ensures the shooting quality, effectively reduces energy consumption, and extends the life of the equipment; This embodiment ensures the accuracy and reliability of ambient light data by eliminating invalid data, effectively eliminating the influence of noise and outliers; automatically identifies and retains the most suitable images for subsequent analysis through image screening, thereby improving the accuracy of image analysis; accurately predicts the light temperature coefficient based on the ambient temperature, making the light intensity calculation more scientific and practical, and effectively responding to changes in light intensity under different temperature conditions; automatically adjusts the brightness of the fill light through real-time calculation of the ambient light intensity and the optimal ambient light intensity, and accurately controls the light intensity, thereby ensuring shooting quality, effectively reducing energy consumption, and extending the life of the equipment. Embodiment 2

[0019] See also Figure 2 As shown, the part not described in detail in this embodiment is described in Example 1, and a method for adaptively adjusting the brightness of a fill light of a vehicle-mounted DMS camera is provided, including: S1, collect ambient light data and character images; S2, remove invalid data from ambient light data to obtain valid light data, and screen the character image to obtain the most suitable image; S3, obtaining the light temperature coefficient based on the constructed light intensity influence coefficient model; S4. Perform light calculation based on effective light data and light temperature coefficient to obtain ambient light intensity, perform graph analysis based on the optimal image to obtain optimal ambient light intensity, and adjust lighting based on the ambient light intensity and the optimal ambient light intensity. Embodiment 3

[0020] This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the operation mode of the above-mentioned method for adaptively adjusting the brightness of the fill light of a vehicle-mounted DMS camera is implemented.

[0021] Since the electronic device introduced in this embodiment is an electronic device used to implement a method for adaptively adjusting the brightness of a vehicle-mounted DMS camera fill light in the embodiment of this application, based on the method for adaptively adjusting the brightness of a vehicle-mounted DMS camera fill light introduced in the embodiment of this application, a person skilled in the art can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application will not be described in detail here. As long as a person skilled in the art implements the electronic device used in the method for adaptively adjusting the brightness of a vehicle-mounted DMS camera fill light in the embodiment of this application, it belongs to the scope of protection of this application.

[0022] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0023] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A vehicle-mounted DMS camera fill light brightness adaptive adjustment system, characterized in that: include: Data acquisition module: collects ambient light data and character images; Data processing module: remove invalid data from ambient light data to obtain valid light data, and screen the character images to obtain the most suitable image; Impact factor module: obtain the light temperature coefficient based on the constructed light intensity impact coefficient model; Lighting adjustment module: performs light calculation based on effective light data and light temperature coefficient to obtain ambient light intensity, performs graph analysis based on the most suitable image to obtain the optimal ambient light intensity, and adjusts the lighting based on the ambient light intensity and the optimal ambient light intensity.

2. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 1, characterized in that: The ambient light data includes: light intensity, spectral distribution, ambient temperature, light wavelength, illuminated area and reflectivity.

3. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 2, characterized in that: The method for removing invalid data includes: The ambient light data is normalized by the Min-Max normalization algorithm to obtain normalized light data, and the normalized light data is vectorized to obtain the ambient light vector. Calculate the data points Other data points and data points The similarity is calculated as follows: ;in, Representative The ambient light vector is The similarity of the ambient light vectors, represents the bandwidth parameter, represents the base of natural logarithms, Representative Ambient light vectors, Representative Ambient light vectors, and represents the index of the ambient light vector, and ; Constructing an undirected similarity graph based on similarity ,in, represents the ambient light vector, Represents a set of edges; preset similarity threshold , and define the edge weight, the weight calculation formula is: ;in, Representative The ambient light vector is The weight of the edge connecting the ambient light vectors; For the ambient light vector in the undirected similarity graph , all with The indices of other ambient light vectors that have directly connected edges Neighborhood index set , the ambient light vector corresponding to the index in the neighborhood index set and the ambient light vector Construct a local point set, and count the total number of edges that are actually connected to each ambient light vector in the local point set based on the undirected similarity graph , based on the total number of Calculating the ambient light vector The clustering coefficient is calculated as follows: ;in Representative The clustering coefficient of the ambient light vectors, Represents the total amount of data in the neighborhood index set; The basic outlier degree of each data point in the ambient light vector is calculated based on the degree of the ambient light vector in the undirected similarity graph. The calculation formula of the basic outlier degree is: ;in, Representative The basic outlier of the ambient light vector, No. The degree of the ambient light vector, Represents the maximum degree of the ambient light vector in the undirected similarity graph; The local outlier degree of each data point in the ambient light vector is calculated based on the neighborhood index set and the clustering coefficient. The calculation formula of the local outlier degree is: ;in, Representative The local outlier of the ambient light vector, Representative The neighborhood index set of ambient light vectors, Representative A similarity list of ambient light vectors; The comprehensive outlier degree is calculated based on the basic outlier degree and the local outlier degree. The formula for calculating the comprehensive outlier degree is: ;in, Representative The comprehensive outlier degree of the ambient light vectors; the abnormal threshold is set based on the comprehensive outlier degree ,in, represents the abnormal threshold, represents the mean of the comprehensive outlier degree, Represents the standard deviation of the comprehensive outlier degree; the ambient light data corresponding to the ambient light vector whose comprehensive outlier degree is greater than or equal to the abnormal threshold is marked as abnormal data, and all ambient light data not marked as abnormal data is regarded as valid light data.

4. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 3, characterized in that: The steps of performing graph screening include: Step 1: preset K groups of random person images, each group of person images is represented as a chromosome, and the genes in the chromosome correspond to the pixels in the person image; Step 2: For each chromosome, the character image is gray-processed by the quantization method to obtain a gray-scale image. The fitness of the chromosome is evaluated based on the gray-scale image. The calculation formula for fitness evaluation is: ;in, represents the fitness of the chromosome, represents the number of pixels of the chromosome, Represents the chromosome The gray value of a pixel, Represents the mean gray value of each pixel in the chromosome, Represents the index of the pixel in the chromosome; the parent chromosome is selected by elite selection method according to the fitness value; Step 3: Compare the genes in the parent chromosome one by one according to the grayscale value, use the parent chromosome as a template, exchange the genes with high grayscale values ​​in the mother chromosome with the genes in the parent chromosome to generate the daughter chromosome; Step 4: Using a random mutation algorithm, randomly select U genes in the offspring chromosome to replace them with genes at corresponding positions in the pre-stored chromosome; Step 5, evaluate the fitness of the offspring chromosome. When the fitness value of the offspring chromosome is greater than the fitness value of the pre-stored chromosome, replace the pre-stored chromosome with the offspring chromosome as the new pre-stored chromosome. When the maximum number of iterations has not been reached and the fitness value of the pre-stored chromosome is less than the preset fitness threshold, return to step 2 to continue iterating. When the maximum number of iterations is reached or the fitness value of the pre-stored chromosome is greater than the preset fitness threshold, stop the iteration and output the pre-stored chromosome as the optimal image.

5. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 4, characterized in that: The method for obtaining the light temperature coefficient includes: Collect group D of historical environmental data, including historical ambient temperature, historical measured light intensity and historical predicted light intensity. Based on the historical environmental data, the ratio of historical measured light intensity to historical predicted light intensity is used as the historical actual light temperature coefficient, the historical ambient temperature is used as the input of the light intensity influence coefficient model, the predicted light temperature coefficient is used as the output of the light intensity influence coefficient model, and the historical actual light temperature coefficient is used as the prediction target of the light intensity influence coefficient model for model training. The light intensity influence coefficient model is constructed. The light intensity influence coefficient model is a linear regression model. The target loss function of the light intensity influence coefficient model is: ;in, is the mean square error, is the total amount of historical environmental data, No. The historical actual light temperature coefficient, It is The predicted light temperature coefficient is predicted; the training goal is to minimize the value of the target loss function and obtain the light intensity influence coefficient model with the minimum value of the target loss function; Based on the constructed light intensity influence coefficient model, the ambient temperature is used as the input of the light intensity influence coefficient model for prediction to obtain the light temperature coefficient.

6. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 5, characterized in that: The method of performing light calculation includes: The light calculation function is defined according to the light temperature coefficient. The formula of the light calculation function is: ,in, Represents the ambient light intensity, represents the light intensity, Represents the minimum value of the wavelength of light, represents the maximum value of the wavelength of light, represents the wavelength of light, represents the central light wavelength, represents the spectral distribution, represents the standard deviation of the spectral distribution, represents the light temperature coefficient, represents the reflectivity, Representatives Perform integration.

7. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 6, characterized in that: The method of performing graph analysis includes: The brightness of the optimal image is calculated to obtain the optimal brightness, and the light is reversed based on the optimal brightness to obtain the optimal ambient light intensity.

8. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 7, characterized in that: The method for obtaining the optimal brightness includes: The brightness conversion of the optimal image is performed, and the formula for brightness conversion is: ;in, Represents the best image The brightness of a pixel, Represents the best image The red component of the pixel, Represents the best image The blue component of pixels, Represents the best image The green component of pixels, Represents the best image pixels, , and Represents the weighting coefficient; the brightness of all pixels in the optimal image is averaged to obtain the optimal brightness.

9. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 8, characterized in that: The method for obtaining the optimal ambient light intensity includes: Based on the optimum brightness, light inversion is performed, and the calculation formula for light inversion is: ;in, Represents the optimal ambient light intensity, represents the irradiated area, Represents the angle between the irradiated light and the irradiated area, Represents the minimum angle between the irradiated light and the irradiated area. Represents the maximum angle between the irradiated light and the irradiated area. represents the reflectivity, Representatives Perform integration.

10. The vehicle-mounted DMS camera fill light brightness adaptive adjustment system according to claim 9, characterized in that: The lighting adjustment method includes: According to the comparison between the optimal ambient light intensity and the ambient light intensity, when When the light is not adjusted, Calculate the fill light intensity when , the power is calculated based on the fill light intensity, and the formula for power calculation is: ;in, represents the required power, Represents the light effect of the light source. Represents the area of ​​the irradiated area; based on the required power, the required voltage is obtained according to the voltage power calculation formula. The voltage power calculation formula is: ;in, Represents the required voltage, Represents a fixed current; the voltage is adjusted according to the required voltage, thereby adjusting the light intensity.

Images