Car window light transmittance adjusting method and device and medium

Abstract

The invention provides a vehicle window light transmittance adjusting method and device and a medium, and the method comprises the steps: obtaining user images of users in a vehicle through a camera in a vehicle cabin, and determining the seat positions, identity information and fatigue degrees of the users in the vehicle according to the user images; according to the seat positions of the users in the vehicles, to-be-adjusted vehicle windows corresponding to the users in the vehicles are determined; acquiring environment illumination intensity, road scene information, passing road condition information and vehicle state information; for each in-vehicle user, according to the identity information of the in-vehicle user, determining a target color changing model corresponding to the in-vehicle user, inputting the fatigue degree, the environment illumination intensity, the road scene information, the traffic condition information and the vehicle state information of the in-vehicle user into the target color changing model to obtain a target light transmittance, and according to the target light transmittance, determining a target color changing model corresponding to the in-vehicle user; the to-be-adjusted car window corresponding to the user in the car is adjusted, the light transmittance of the car window is adjusted in a targeted mode, and the adjusting speed and effectiveness of the light transmittance of the car window are improved.

Description

Technical Field

[0001] The present application relates to the field of vehicle control technology, and in particular to a method, device and medium for adjusting the light transmittance of a vehicle window. Background Art

[0002] With the development of technology and people's constant pursuit of comfort and privacy, no matter high-end or low-end models, a layer of solar film will be pasted on the windows and windshields after purchase to achieve functions such as heat insulation, external radiation protection and privacy protection. Traditional solar film is a one-time product, and its light transmittance cannot be adjusted after pasting, and it may be damaged.

[0003] Photochromic glass has been gradually applied in the automotive field due to its adjustable light transmittance and strong UV shielding ability. However, the light transmittance of conventional photochromic glass is adjusted manually or adjusted by photosensitivity to external light, which cannot effectively adapt to the changing surrounding scene when the vehicle is driving, resulting in poor user experience. Summary of the invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, the present application aims to provide a method, device and medium for adjusting the light transmittance of vehicle windows, so as to achieve targeted adjustment of the light transmittance of vehicle windows according to driving scenarios and users in the vehicle, thereby improving the response speed, effectiveness and personalized effect of adjusting the light transmittance of vehicle window glass.

[0005] The present application provides a method for adjusting the light transmittance of a vehicle window, the method comprising: The user image of each user in the vehicle is obtained through the camera inside the vehicle cabin, and the seat position, identity information and fatigue level of each user in the vehicle are determined based on each user image; Determine the window to be adjusted corresponding to each user in the car according to the seat position of each user in the car; Obtain ambient light intensity, road scene information, traffic condition information, and vehicle status information; For each user in the car, a target color change model corresponding to the user in the car is determined according to the identity information of the user in the car, the fatigue level of the user in the car, the ambient light intensity, the road scene information, the traffic road condition information and the vehicle status information are input into the target color change model to obtain the target transmittance, and according to the target transmittance, the window to be adjusted corresponding to the user in the car is adjusted.

[0006] According to the technical solution provided in the embodiment of the present application, optionally, after adjusting the window to be adjusted corresponding to the user in the vehicle according to the target light transmittance, the method further includes: In response to the received window active adjustment parameter, at least one target window corresponding to the window active adjustment parameter and a corrected light transmittance corresponding to each target window are determined, each corresponding target window is adjusted based on each corrected light transmittance, and each target identity information corresponding to the target window is determined according to the target window; For each target identity information, the target color change model corresponding to the target identity information is updated according to the corrected light transmittance of each target vehicle window corresponding to the target identity information.

[0007] According to the technical solution provided in the embodiment of the present application, optionally, updating the target color change model corresponding to the target identity information according to the corrected light transmittance of each target window corresponding to the target identity information includes: Determine the target loss according to a preset loss function, the corrected light transmittance of each target window corresponding to the target identity information, and the target light transmittance; Determining a model parameter adjustment gradient according to the target loss; According to the model parameter adjustment gradient, the preset learning rate and each model parameter in the target color change model corresponding to the target identity information, each model parameter is updated; The target color change model corresponding to the target identity information is updated according to the updated model parameters.

[0008] According to the technical solution provided in the embodiment of the present application, optionally, the obtaining of ambient light intensity, road scene information, traffic condition information and vehicle status information includes: Determine the ambient light intensity based on the light sensor; Determine the current location information according to the positioning system, and determine the road scene information according to the current location information and a preset map; Acquire surrounding environment data according to the camera and the radar, and determine traffic condition information according to the surrounding environment data; wherein the traffic condition information includes congestion information and / or traffic light information; Obtain vehicle status information based on the in-vehicle infotainment system.

[0009] According to the technical solution provided in the embodiment of the present application, optionally, determining the target color change model corresponding to the user in the vehicle according to the identity information of the user in the vehicle includes: In response to the presence of a color change model that matches the identity information among the pre-stored color change models to be selected, the color change model that matches the identity information is used as a target color change model corresponding to the user in the vehicle; In response to the pre-stored color change models to be selected not matching the identity information, obtaining model parameters of the color change models to be selected and model parameters of the reference color change model; Determining average values ​​of the model parameters according to the model parameters of each to-be-selected color change model and the model parameters of the reference color change model; According to the average value of the model parameters and the reference color change model, a color change model to be selected corresponding to the identity information is determined and stored, and the color change model to be selected corresponding to the identity information is determined as a target color change model corresponding to the user in the vehicle.

[0010] According to the technical solution provided in the embodiment of the present application, optionally, the reference color change model is constructed based on the following method: Acquire each sample data and the sample transmittance corresponding to each sample data; wherein the sample data includes sample fatigue, sample ambient light intensity, sample road scene information, sample traffic condition information and sample vehicle status information; The initial color change model is trained according to each sample data and the sample transmittance corresponding to each sample data to obtain a reference color change model, and each model parameter in the reference color change model is determined.

[0011] According to the technical solution provided in the embodiment of the present application, optionally, the obtaining of each sample data and the sample transmittance corresponding to each sample data includes: Acquire each historical time series data and the historical light transmittance corresponding to each historical time series data; wherein the historical time series data includes historical fatigue, historical ambient light intensity, historical road scene information, historical traffic condition information and historical vehicle status information; Perform Gaussian filtering on each historical time series data to obtain each process data; Performing data normalization processing on each process data to obtain each normalized data; Based on the preset window, each normalized data is segmented to obtain each sample data; For each set of sample data, the historical transmittance corresponding to the start timestamp of the sample data is determined as the sample transmittance corresponding to the sample data.

[0012] According to the technical solution provided in the embodiment of the present application, optionally, determining the window to be adjusted corresponding to each user in the vehicle according to the seat position of each user in the vehicle includes: For each in-vehicle user, in response to the in-vehicle user being in the front driving seat, determining the driving-related window and all the front-row windows as the corresponding windows to be adjusted; wherein the driving-related windows include the front windshield window, the rear windshield window and the vehicle sunroof; In response to the user in the vehicle being located in the front passenger seat, determining that the corresponding window to be adjusted is empty; In response to the user in the vehicle being located in a rear passenger seat, determining a target distance between the user in the vehicle and each window in the same row corresponding to the seat row where the user in the vehicle is located; For each window in the same row, in response to the target distance between the user in the vehicle and the window in the same row being the minimum value among the target distances corresponding to the windows in the same row, the window in the same row is determined to be the window to be adjusted corresponding to the user in the vehicle.

[0013] The present application also provides an electronic device, the electronic device comprising: Processor and memory; The processor is used to execute the steps of the vehicle window transmittance adjustment method as described in any embodiment by calling the program or instruction stored in the memory.

[0014] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program or instruction, wherein the program or instruction enables a computer to execute the steps of the vehicle window transmittance adjustment method as described in any embodiment.

[0015] In summary, the present application proposes a method for adjusting the light transmittance of vehicle windows, which obtains the user image of each user in the vehicle through a camera inside the vehicle cabin, and determines the seat position, identity information and fatigue level of each user in the vehicle according to each user image, determines the window to be adjusted corresponding to each user in the vehicle according to the seat position of each user in the vehicle, and obtains the ambient light intensity, road scene information, traffic condition information and vehicle status information, so as to facilitate the subsequent comprehensive adjustment of the light transmittance of the window in combination with external factors and the user's own factors in the vehicle, and then, for each user in the vehicle, determines the target color change model corresponding to the user in the vehicle according to the identity information of the user in the vehicle, and adjusts the light transmittance of the window according to the target color change model of the user in the vehicle. The user's fatigue, ambient light intensity, road scene information, traffic condition information and vehicle status information are input into the target color change model to obtain the target transmittance, and according to the target transmittance, the window to be adjusted corresponding to the user in the car is adjusted, so as to achieve targeted adjustment of the transmittance of the window according to the driving scene and each user in the car, thereby improving the response speed, effectiveness and personalized effect of the light transmittance adjustment of the window glass. Compared with the manual adjustment and simple photochromic adjustment method in the prior art, the transmittance adjustment in various driving scenarios is taken into account, and personalized adjustment is also performed for different users, which is more intelligent and improves the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flow chart of a method for adjusting the light transmittance of a vehicle window provided in an embodiment of the present application; Figure 2 is a schematic diagram of a reference color change model provided in an embodiment of the present application; Figure 3is a flow chart of another vehicle window transmittance adjustment method provided in an embodiment of the present application; Figure 4 It is a structural schematic diagram of a vehicle window transmittance adjustment system provided in an embodiment of the present application; Figure 5 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0017] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0018] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0019] The car window glass can be made of photochromic glass. Common photochromic glass technologies include: PDLC technology (Polymer Dispersed Liquid Crystal), SPD technology (Suspended Particle Device) and EC technology (Electrochromic).

[0020] PDLC technology is made by dispersing liquid crystal droplets in a polymer matrix. When no voltage is applied, the arrangement of the liquid crystal molecules is random, causing light to scatter, making the glass translucent or opaque. When voltage is applied, the liquid crystal molecules are arranged neatly, allowing light to pass through, making the glass transparent. This technology is often used in privacy protection occasions. Its advantages are low cost and fast response, but its disadvantages are also obvious. This technology has high energy consumption, cannot be adjusted continuously, and white glass is not suitable for privacy glass in the back row.

[0021] SPD glass contains tiny optical particles suspended in a liquid. When no voltage is applied, these particles are randomly distributed, blocking light and making the glass appear dark or opaque. When voltage is applied, the particles are neatly arranged, allowing light to pass through, making the glass transparent. This technology has a strong ability to block visible light and ultraviolet rays, but it is expensive and requires 110V high voltage electricity to continuously drive.

[0022] Electrochromic glass works by sandwiching or coating electrochromic materials in the glass layer. When voltage is applied, these materials undergo a redox reaction, changing their optical properties (color or transmittance). Typically, EC glass can switch between transparency and different degrees of color, such as from transparent to dark (usually blue or gray). It has low energy consumption, strong ability to absorb visible light and ultraviolet rays, and has a power-off memory function. The disadvantage is that the color change speed is slower than other technologies, but it is also within an acceptable range.

[0023] The specific type of photochromic glass used to implement the vehicle window transmittance adjustment method of the present application can be selected based on requirements such as color and change rate.

[0024] Figure 1 is a flow chart of a method for adjusting the light transmittance of a vehicle window provided in an embodiment of the present application. Figure 1 , the vehicle window transmittance adjustment method specifically includes: S110, obtaining a user image of each user in the vehicle through a camera inside the vehicle cabin, and determining a seat position, identity information, and fatigue level of each user in the vehicle based on each user image.

[0025] Among them, the users in the car include the driver and the passengers. The user image is an image taken for each user in the car. The seat position includes the driver's seat and the passenger's seat, and includes the number of seats in the row. The identity information is used to distinguish different users in the car. It can be distinguished by identifying various facial features, and then numbering different users in the car. Fatigue is used to describe the fatigue level of the user in the car, which is mostly determined based on the analysis of the eye area of ​​the user image.

[0026] Specifically, the user image of each user in the car can be captured by a camera inside the car cabin. For example, the user image of each user in the car can be captured by a camera based on a system such as an OMS (Occupant Monitoring System) or a DMS (Driver Monitoring System), and identity recognition and fatigue detection can be performed on each user image to obtain the seat position, identity information and fatigue level of each user in the car.

[0027] S120: Determine the window to be adjusted corresponding to each user in the vehicle according to the seat position of each user in the vehicle.

[0028] The windows to be adjusted are windows associated with different users in the vehicle, that is, windows that can be controlled without perception by different users in the vehicle.

[0029] Specifically, according to the distribution of the seat positions of the users in the car, the car windows that can be individually controlled by each user in the car are determined.

[0030] It is understandable that there may be a situation where one or several users in the car cannot control the windows in a personalized manner. This is because other users in the car have a higher control priority over the windows than the user in the car.

[0031] Based on the above example, the window to be adjusted corresponding to each user in the car can be determined according to the seat position of each user in the car in the following manner: For each user in the vehicle, in response to the user in the vehicle being located in the front driving seat, determining the driving-related window and all the front windows as corresponding windows to be adjusted; In response to the user in the vehicle being located in the front passenger seat, determining that the corresponding window to be adjusted is empty; In response to the user in the vehicle being located in a rear passenger seat, determining a target distance between the user in the vehicle and each window in the same row corresponding to the seat row where the user in the vehicle is located; For each window in the same row, in response to the target distance between the user in the vehicle and the window in the same row being the minimum value among the target distances corresponding to the windows in the same row, the window in the same row is determined to be the window to be adjusted corresponding to the user in the vehicle.

[0032] Among them, the front driving seat is the driver's position of the first row of seats in the front direction of the vehicle. The front passenger seat is the other position in the first row of seats in the front direction except the driver's position. The driving-related windows are windows that affect the driver's driving operation of the vehicle, and the driving-related windows include the front windshield window, the rear windshield window and the vehicle sunroof. All front-row windows can be understood as all windows corresponding to the number of rows of seats where the front driving seat is located, that is, the windows on the left and right sides of the first row in the front direction of the vehicle. The rear passenger seats are the other rows of seats except the first row in the front direction of the vehicle. The windows in the same row are the windows corresponding to the number of rows of seats where the user in the car is located. It can be understood that if the user in the car is a passenger in the third row of seats, the windows in the same row are the left and right windows corresponding to the third row of seats. The target distance is the distance between the user in the car and the windows in the same row.

[0033] Specifically, for each in-car user, the window to be adjusted corresponding to the in-car user can be determined respectively. If the in-car user is located in the front driving seat, the driving-related window and all the front windows are used as the windows to be adjusted corresponding to the in-car user. These windows are strongly related to the driving situation and driving operation, and therefore are only affected by the driver. If the in-car user is located in the front passenger seat, since the driver of the front driving seat has personalized and invisible control over the driving-related window and all the front windows, the front passenger seat has no control over these windows, and the rear windows have little effect on the front passenger seat, so it is determined that the window to be adjusted corresponding to the in-car user is empty. If the in-car user is located in the rear passenger seat, the windows in the same row corresponding to the number of seats where the in-car user is located are first determined, and then the target distance between the in-car user and the windows in the same row is measured respectively, which can be to identify the distance between the in-car user and the windows in the same row through the captured in-car image. Furthermore, for each window in the same row, it is determined whether the target distance between the user in the vehicle and the window in the same row is the minimum value among the target distances corresponding to the windows in the same row, that is, it is determined whether the user in the vehicle is the one of the users in the same row closest to the window in the same row. If so, the window in the same row is the window to be adjusted corresponding to the user in the vehicle. If not, the window in the same row is not the window to be adjusted corresponding to the user in the vehicle.

[0034] For example, taking the second row as an example, if the users in the second row include the left user in the second row, the middle user in the second row, and the right user in the second row, then the windows in the same row are the left window in the second row and the right window in the second row. Since the target distance between the left user in the second row and the left window in the second row is smaller than the target distance between the middle user in the second row and the right user in the second row and the left window in the second row, the left window in the second row is the window to be adjusted corresponding to the left user in the second row. It can be seen from the above method that the right window in the second row is the window to be adjusted corresponding to the right user in the second row, and the window to be adjusted corresponding to the middle user in the second row is empty. If there is only one user in the second row, then the windows in the same row are all windows to be adjusted corresponding to the user in the car.

[0035] S130, obtaining ambient light intensity, road scene information, traffic condition information, and vehicle status information.

[0036] Among them, the ambient light intensity is the light intensity around the vehicle, which can be measured based on a light sensor. The road scene information is the road scene where the vehicle is located, such as: highways, expressways, urban roads, parking lots, service areas, toll booths, etc. The traffic condition information can include road congestion and traffic light conditions, such as: lane congestion, large traffic flow around, parking, waiting for traffic lights, etc. The vehicle status information is information used to describe the status of the vehicle and the driving status of the vehicle, such as: current vehicle speed, temperature in the vehicle, number of people in the vehicle, etc.

[0037] Specifically, through the sensors and intelligent systems installed on the vehicle, the ambient light intensity, road scene information, traffic condition information and vehicle status information can be collected or analyzed.

[0038] It is understandable that the intensity of ambient light will affect the light projected through the car windows. If the ambient light intensity is high, the transmittance can be lowered to avoid glaring light intake. If the ambient light intensity is low, the transmittance can be increased to facilitate observation of the situation outside the car. Therefore, the intensity of ambient light can affect the transmittance of the car windows to improve the comfort of users in the car. Road scene information will provide vehicle driving information to a large extent. For example, in scenes such as highways and expressways, vehicles usually travel fast, and the probability of privacy leakage through the windows is low, so there is no need to lower the transmittance. However, in scenes such as parking lots, service areas, and toll stations, vehicles are usually parked or driving slowly, and there is a possibility of privacy leakage through the windows. Therefore, it may be necessary to lower the transmittance to protect the privacy in the car. Traffic condition information and vehicle status information can also be used to describe the driving situation of the vehicle. For example, in the traffic condition information, such as parking, lane congestion, and waiting for traffic lights, there is a possibility of privacy leakage through the car window, and it may be necessary to adjust the window transmittance in combination with other information. The current speed of the vehicle status information is more specific. If the temperature inside the car is too high, the light can be blocked by lowering the transmittance to prevent the temperature inside the car from continuing to rise. Therefore, the ambient light intensity, road scene information, traffic condition information, and vehicle status information will all affect the window transmittance.

[0039] Therefore, the window transmittance adjustment function may be triggered in the following scenarios: 1) Changes in ambient light intensity: When the external sunlight intensity is too high, the glass should be adjusted to a darker state to reduce glare, ensure that the driver's field of vision is not disturbed by the sunlight, and reduce the temperature inside the car. When a tunnel scene is identified, the glass transmittance should be increased / decreased according to the timing of entering and exiting the tunnel to prevent the drastic changes in light intensity from affecting driving safety. 2) Privacy needs: When encountering scenes such as heavy traffic, heavy traffic, camping, waiting for traffic lights, and parking, the glass transmittance should be reduced to ensure the privacy of vehicle personnel. After the vehicle is parked, the visibility of the glass should also be reduced to ensure that the items in the car are not easily discovered and protect the safety of the vehicle and the items in the car. 3) Comfort needs: When the temperature changes rapidly and the temperature inside and outside the car is too high, the transmittance should be adjusted to increase / reduce the temperature accumulation caused by the light inside the car, improve the comfort of the people in the car, and reduce the frequency of air conditioning use. In addition, the light transmittance of the glass is dynamically adjusted according to the occupant information and fatigue monitored by OMS and DMS to achieve preference and comfort adjustment. For example, when the rear passengers are sleeping, the light transmittance of the corresponding windows should be reduced to provide a better sleeping environment, and when the driver's fatigue increases, the overall brightness should be appropriately increased to ensure driving safety.

[0040] Based on the above examples, the ambient light intensity, road scene information, traffic condition information and vehicle status information can be obtained in the following ways: Determine the ambient light intensity based on the light sensor; Determine the current location information according to the positioning system, and determine the road scene information according to the current location information and the preset map; Obtain surrounding environment data based on cameras and radars, and determine traffic conditions based on the surrounding environment data; Obtain vehicle status information based on the in-vehicle infotainment system.

[0041] Among them, the traffic condition information includes congestion information and / or traffic light information. The positioning system can be a system or module such as GPS (Global Positioning System) and Beidou system for determining the current position of the vehicle. The current position information is the current position information of the vehicle, such as latitude and longitude information. The preset map is a map with different road scenes marked, and the road scenes include highways, expressways, urban roads, parking lots, service areas, toll stations, etc. The surrounding environment data may include the number of surrounding vehicles, driving speed, pedestrian flow and traffic light conditions. The in-vehicle infotainment system (In-Vehicle Infotainment / On-Board Infotainment, IVI) can realize functions including three-dimensional navigation, real-time traffic conditions, IPTV, assisted driving, fault detection, vehicle information, and body control. Therefore, it can be used to obtain vehicle status information.

[0042] Specifically, the current sunlight intensity is obtained through the light sensor, in units of Lux (lumens), as the ambient light intensity. The current position of the vehicle can be obtained through positioning systems such as GPS, that is, the current position information. Then, the current position information is matched in the preset map to obtain the matched road scene information. Using a separately set camera and radar or cameras and radars equipped with various intelligent systems in the vehicle, the situation around the vehicle, that is, the surrounding environment data, can be collected and obtained. Then, the surrounding environment data is analyzed, mainly analyzing information related to congestion, pedestrian flow, and traffic light conditions to obtain traffic condition information. For example: data can be collected and analyzed through ADAS (Advanced Driving Assistance System, Advanced Driving Assistance System), that is, the environmental information around the vehicle can be obtained, that is, the surrounding environment data, and further analysis and extraction can obtain traffic condition information, such as lane congestion, large traffic flow around, parking, and waiting for traffic lights. Vehicle status information can be obtained based on the in-vehicle infotainment system. For example, the vehicle status information, such as the current vehicle speed, vehicle temperature, number of people in the vehicle, etc., can be comprehensively determined through IVI, OMS, DMS and other systems.

[0043] S140. For each in-vehicle user, a target color change model corresponding to the in-vehicle user is determined according to the in-vehicle user's identity information, and the in-vehicle user's fatigue level, ambient light intensity, road scene information, traffic condition information, and vehicle status information are input into the target color change model to obtain a target transmittance, and according to the target transmittance, the window to be adjusted corresponding to the in-vehicle user is adjusted.

[0044] The target color change model is a pre-stored model corresponding to the user in the car, and is used to adjust the window to be adjusted corresponding to the user in the car. The target light transmittance is the output result of the target color change model, and is used to adjust the corresponding window to be adjusted.

[0045] Specifically, for each in-car user, the target light transmittance of the window to be adjusted corresponding to the in-car user can be determined and adjusted accordingly. This is explained by taking one of the in-car users as an example. The identity information of the in-car user is used to search in the pre-stored color change models, and the model corresponding to the in-car user is determined as the target color change model. Then, the fatigue level of the in-car user, the ambient light intensity, the road scene information, the traffic condition information, and the vehicle status information are input into the target color change model, processed by the target color change model, and the target light transmittance is output. Further, the window to be adjusted corresponding to the in-car user is adjusted according to the target light transmittance.

[0046] Optionally, when the same identity information corresponds to a driver or a passenger, the corresponding target color change model may be different. That is, for the same identity information, one target color change model may be determined when corresponding to the driver's seat, and another target color change model may be determined when corresponding to the passenger's seat. Therefore, when establishing the above model, it may be established for the driver's seat and the passenger's seat respectively, or adjusted separately on the basis of the subsequent reference color change model.

[0047] It can be understood that S140 can be executed only for the in-vehicle user who has the corresponding window to be adjusted, that is: in response to the existence of the window to be adjusted corresponding to the in-vehicle user, the target color change model corresponding to the in-vehicle user is determined according to the identity information of the in-vehicle user, the fatigue level of the in-vehicle user, the ambient light intensity, the road scene information, the traffic condition information and the vehicle status information are input into the target color change model to obtain the target transmittance, and according to the target transmittance, the window to be adjusted corresponding to the in-vehicle user is adjusted.

[0048] Based on the above example, the target color change model corresponding to the user in the car can be determined according to the identity information of the user in the car in the following manner: In response to the presence of a color change model that matches the identity information among the pre-stored color change models to be selected, the color change model that matches the identity information is used as a target color change model corresponding to the user in the vehicle; In response to the pre-stored color change models to be selected not matching the identity information, obtaining model parameters of the color change models to be selected and model parameters of the reference color change model; Determine an average value of the model parameters according to the model parameters of each candidate color change model and the model parameters of the reference color change model; According to the average value of the model parameters and the reference color change model, a candidate color change model corresponding to the identity information is determined and stored, and the candidate color change model corresponding to the identity information is determined as a target color change model corresponding to the user in the vehicle.

[0049] The color change models to be selected are pre-stored color change models, and each color change model to be selected matches an identity information. The reference color change model is a model trained based on sample data, that is, a universal color change model. It can be understood that each color change model to be selected is adjusted based on the reference color change model.

[0050] Specifically, if there is a color change model to be selected that matches the identity information among the pre-stored color change models to be selected, it means that there is a personalized color change model to be selected corresponding to the identity information, therefore, the existing color change model to be selected that matches the identity information can be used as the target color change model corresponding to the corresponding in-vehicle user. If the pre-stored color change models to be selected do not match the identity information, it means that the corresponding target color change model has not been constructed for the identity information, therefore, the pre-stored color change models to be selected and the reference color change model can be processed comprehensively to obtain the target color change model, specifically: obtain the model parameters of the color change models to be selected and the model parameters of the reference color change model, calculate the average value of the model parameters of these models, and use the average value of the model parameters to replace the model parameters in the reference color change model, to obtain a new color change model to be selected, which is the target color change model corresponding to the in-vehicle user, so as to facilitate the subsequent search for the target color change model for the in-vehicle user, and the color change model to be selected can be stored corresponding to the identity information.

[0051] The previously collected data (fatigue, ambient light intensity, road scene information, traffic conditions information, and vehicle status information, etc.) is given as input to the target color change model. After being processed by the model, the window changes and transmittance of the window to be adjusted (such as adjusting the gear position) will be output and given as control instructions to the window color change actuator. The actuator will output the corresponding current to the corresponding window to be adjusted according to the received instructions, so that its transmittance changes, and finally achieve the effect of intelligent color change.

[0052] Based on the above example, a baseline color change model can be constructed in the following way: Obtaining each sample data and the sample transmittance corresponding to each sample data; The initial color change model is trained according to each sample data and the sample transmittance corresponding to each sample data to obtain a reference color change model, and each model parameter in the reference color change model is determined.

[0053] The sample data includes sample fatigue, sample ambient light intensity, sample road scene information, sample traffic condition information and sample vehicle status information. The sample transmittance is the transmittance of the appropriate window manually adjusted by the test user corresponding to each set of sample data. The initial color change model is a neural network model, which can be an initial model whose model parameters are not trained.

[0054] Specifically, each sample data and the sample transmittance corresponding to each sample data for model training are obtained. An initial color change model is constructed, and each sample data and the sample transmittance corresponding to each sample data are used to train the initial color change model. The trained model is used as a reference color change model, and each model parameter in the reference color change model can be obtained.

[0055] For example, if there is an open source data set for reference, it can be used directly. If there is no open source data set for reference, sample data can be obtained by manually designing and collecting data sets. A preset number (for example, 1,000) of test vehicles that can actively adjust the window transmittance can be distributed to different regions. Each test vehicle is assigned to different test users to drive and conduct tests in various scenarios. The test vehicle will record the operation of manually adjusting the window transmittance, and model training can begin after having enough sample data.

[0056] The input of the baseline color change model (neural network model) specifically includes the following: Sample environment light intensity L, for example: L∈[0,5000], when L=0Lux, it means it is in a dark environment, when L=5000Lux, it means it is in a strong light environment.

[0057] Sample road scene information P, for example: P∈[0,1,2,3,4,5], where 0 represents a parking lot, 1 represents a city road, 2 represents an expressway, 3 represents a highway, 4 represents a service area, and 5 represents a toll station.

[0058] Sample traffic condition information E, for example: E=[ ],in, Indicates the degree of traffic congestion. It means that the traffic is smooth. Indicates mild congestion. Indicates severe congestion; represents the density of human traffic, It means that there are few people. ; Indicates the current status of the vehicle. =0 means parking. =1 means the vehicle is moving, Indicates waiting at the traffic light; Indicates the traffic light status. =0 means green light, =1 means yellow light, =2 means red light; Indicates the real-time temperature outside the vehicle.

[0059] Sample vehicle status information and sample fatigue S, for example: ,in, Represent the passenger ID and driver ID respectively, which are used to load the preference settings in order to determine the corresponding target color change model. and ∈[0,5] represents the current fatigue level of the passenger and the driver, 0 represents wakefulness, 5 represents fatigue, represents the temperature inside the vehicle, and v represents the current vehicle speed.

[0060] The output of the baseline color-changing model (neural network model) is: Window transmittance control signal: contains the window number that needs to be changed and the corresponding transmittance. Model parameter θ: The model is automatically updated and generated, which can be used for subsequent intelligent feedback correction to optimize the model. The model parameter is the core control variable for model adjustment. These model parameters will determine how the model maps from input variables (such as ambient light intensity, road scene information, in-car temperature, etc.) to output variables (such as window transmittance). This model parameter should only appear in the hidden layer of the neural network model. Here, for the purpose of subsequent model updates, these model parameters are also printed out as output parameters.

[0061] Window transmittance control signal C, for example: C = [G, τ], G ∈ [0, 1, 2, 3, 4, 5], G = 0 means no adjustment, 1 means adjusting the front window transmittance, 2 means adjusting the rear window transmittance, 3, 4, 5 respectively indicate the adjustment of the left and right windows and the roof window, τ indicates the adjustment size of each transmittance, the transmittance of conventional photochromic glass is 5%~80%, so τ = 50% means adjusting the transmittance of the glass with the corresponding window number to 50%. Model parameter θ = [ ], where n+1 is the total number of model parameters.

[0062] Based on the above example, each sample data and the sample transmittance corresponding to each sample data can be obtained in the following manner: Obtain each historical time series data and the historical transmittance corresponding to each historical time series data; Perform Gaussian filtering on each historical time series data to obtain each process data; Performing data normalization processing on each process data to obtain each normalized data; Based on the preset window, each normalized data is segmented to obtain each sample data; For each set of sample data, the historical transmittance corresponding to the start timestamp of the sample data is determined as the sample transmittance corresponding to the sample data.

[0063] Among them, the historical time series data includes historical fatigue, historical ambient light intensity, historical road scene information, historical traffic condition information and historical vehicle status information, and various data in the historical time series data are aligned in time series. It can be understood that in this example, the historical time series data are the data collected during the test, and the sample data are the data obtained by data preprocessing based on the historical data. The process data is the data after Gaussian filtering of the historical time series data. Normalized data is the data after normalization of the process data. The preset window is a window determined according to a preset time length, which is used to split longer time series data into shorter time series data. The start timestamp is the timestamp corresponding to the first data in the sample data.

[0064] Specifically, each historical time series data and the historical transmittance corresponding to each historical time series data can be obtained through actual testing. Then, each historical time series data is subjected to Gaussian filtering, and the processed data is used as each process data. Further, each process data is subjected to data normalization processing to obtain each normalized data. Then, each normalized data is segmented according to a preset step size using a preset window, and each group of data obtained by the data segmentation is used as a sample data. For each group of sample data, the historical transmittance corresponding to the start timestamp of this group of sample data is determined as the sample transmittance corresponding to this group of sample data.

[0065] For example, after obtaining the historical time series data, further processing can be performed, as follows: First, data cleaning can be performed. Consider using Gaussian filtering to clean historical time series data. Gaussian filtering is a weighted average filter. Gaussian weights are distributed according to Gaussian distribution. The data points in the center of the field have a large weight, while the edge data have a small weight.

[0066] For a one-dimensional data x, the output of Gaussian filtering is for:

[0067] in:

[0068] are Gaussian weights, , k is the radius of the window, the window size is 2k+1, and i is the order.

[0069] Then, the process data obtained after data cleaning is normalized. The maximum-minimum value normalization method can be used to scale the process data to the range of 0-1. The formula is as follows:

[0070] in, To normalize the data, For process data, is the minimum value in the process data, is the maximum value in the process data.

[0071] It is understandable that historical time series data needs to be processed separately according to different dimensions.

[0072] Finally, the time series is segmented: the normalized data can be segmented into window data suitable for model input, namely sample data.

[0073] For example, the construction of the benchmark color change model (neural network model) can be to build a fully connected neural network with 2 hidden layers using a 4×8×4×2 network structure, and the activation functions are 2 Relu (linear rectification functions) and 1 tanh (hyperbolic tangent function), such as Figure 2 shown.

[0074] As a simple and efficient activation function, Relu directly maps the negative values ​​in the input x to 0, and the positive values ​​remain unchanged, which can effectively solve the problems of gradient disappearance and sparsity. Its formula is:

[0075] Tanh maps the input x to between -1 and 1, making the output It is centered on 0. Due to its symmetry, this activation function can effectively accelerate convergence and avoid deviation. Its formula is:

[0076] Select MSE (Mean-Square Error) as the loss, and the corresponding formula is:

[0077] Among them, y and are the true value and the predicted value of the model output, respectively, and n is the number of predicted values.

[0078] The specific training process is as follows: First, assume that the parameters of the input layer are They correspond to the parameters defined before the i-th one, such as: = sample ambient light intensity L, the output layer is They correspond to the window transmittance control signal and the model parameter θ respectively.

[0079] Because using Figure 2 The 4-layer network structure shown here consists of 3 weights and biases: weight matrix , with a shape of 4×8, connecting the input layer to hidden layer 1; bias vector , shape is 4, corresponding to the bias of the hidden layer; weight matrix , with a shape of 8×4, connecting hidden layer 1 and hidden layer 2; bias vector , shape is 8, corresponding to hidden layer 2 of hidden layer 1; weight matrix , with a shape of 4×2, connecting hidden layer 2 and the output layer; bias vector , the shape is 4, corresponding to the bias of the hidden layer.

[0080] The process of forward propagation is: 1. From the input layer to the hidden layer 1: pass the input x through the weight matrix and bias Perform a linear transformation to obtain , and is obtained by nonlinear transformation through activation function Relu : , =Activation( ).

[0081] 2. From hidden layer 1 to hidden layer 2: Through the weight matrix and bias Perform a linear transformation to obtain , and is obtained by nonlinear transformation through activation function Relu : , =Activation( ).

[0082] 3. From hidden layer 2 to output layer: Through the weight matrix and bias Perform a linear transformation to obtain , and is obtained by nonlinear transformation through activation function tanh : ,y^=Activation( ). Where y^ is the predicted value of the neural network model.

[0083] Then, the loss value L is calculated using the loss function defined previously.

[0084] Backward propagation is used to calculate the gradient of the loss function for each model parameter (w, b) in the neural network model, differentiate the weights and biases of the previous layer, and propagate the gradient to the front layer. The process is: Gradient from the output layer to hidden layer 2: First calculate the output layer error: ,in is the error term of the output layer, and then calculate the gradient of hidden layer 2, , .

[0085] Gradient from hidden layer 2 to hidden layer 1: Calculate the error of hidden layer 2: ,in, is the inverse of the activation function. Calculate the gradient of hidden layer 1 , .

[0086] Gradient from hidden layer 1 to input layer: Calculate the error of hidden layer 1: ,in, is the inverse of the activation function, calculating the gradient of hidden layer 1 , .

[0087] Then, parameter update is performed, specifically: Update weights and biases using the gradient descent algorithm: , ,in, is the learning rate, and l represents the number of layers of the neural network model (1, 2, 3).

[0088] The entire model training process includes multiple forward propagation and backward propagation cycles, each cycle is called an iteration, and when the entire dataset is used to train an iteration, it is called an epoch. The entire process can be summarized as the following steps: initialize weights and biases → forward propagation → calculate loss → backward propagation → parameter update → repeat until the desired effect is achieved.

[0089] By continuously adjusting the model parameters and multiple iterations, the weights and biases of the neural network model are gradually adjusted to an optimal state, and eventually a baseline color change model will be obtained. When data is input, the model will output a predicted value y^, which will be closest to the current expectation.

[0090] The window transmittance adjustment method provided in the embodiment of the present application obtains the user image of each user in the car through the camera inside the car cabin, and determines the seat position, identity information and fatigue level of each user in the car according to each user image, determines the window to be adjusted corresponding to each user in the car according to the seat position of each user in the car, and obtains the ambient light intensity, road scene information, traffic road condition information and vehicle status information, so as to facilitate the subsequent comprehensive adjustment of the window transmittance in combination with external factors and the user's own factors in the car, and then, for each user in the car, determines the target color change model corresponding to the user in the car according to the identity information of the user in the car, and adjusts the window transmittance of the user in the car according to the target color change model of the user in the car. Fatigue, ambient light intensity, road scene information, traffic condition information and vehicle status information are input into the target color change model to obtain the target transmittance, and according to the target transmittance, the windows to be adjusted corresponding to the users in the car are adjusted, so as to achieve targeted adjustment of the transmittance of the windows according to the driving scene and the users in the car, thereby improving the response speed, effectiveness and personalized effect of the transmittance adjustment of the window glass. Compared with the manual adjustment and simple photochromic adjustment method in the prior art, the transmittance adjustment in various driving scenarios is taken into account, and personalized adjustment is also performed for different users, which is more intelligent and improves the user experience.

[0091] Figure 3 This is a flow chart of another window transmittance adjustment method provided by an embodiment of the present application. Based on the above embodiments, a process of updating the target color change model corresponding to each user in the car based on the active adjustment of the window transmittance by each user in the car is added, and an exemplary description of the process is given. Figure 3 , the vehicle window transmittance adjustment method specifically includes: S210, obtaining a user image of each user in the vehicle through a camera inside the vehicle cabin, and determining a seat position, identity information, and fatigue level of each user in the vehicle based on each user image.

[0092] S220: Determine the window to be adjusted corresponding to each user in the vehicle according to the seat position of each user in the vehicle.

[0093] S230: Obtaining ambient light intensity, road scene information, traffic condition information, and vehicle status information.

[0094] S240. For each in-vehicle user, a target color change model corresponding to the in-vehicle user is determined according to the in-vehicle user's identity information, and the in-vehicle user's fatigue level, ambient light intensity, road scene information, traffic condition information, and vehicle status information are input into the target color change model to obtain a target light transmittance. According to the target light transmittance, the window to be adjusted corresponding to the in-vehicle user is adjusted.

[0095] S250. In response to the received active window adjustment parameters, at least one target window corresponding to the active window adjustment parameters and the corrected light transmittance corresponding to each target window are determined, and based on each corrected light transmittance, each corresponding target window is adjusted, and according to the target window, each target identity information corresponding to the target window is determined.

[0096] Among them, the window active adjustment parameter is the result parameter of active transmittance adjustment for each window of the vehicle, and the active adjustment method can be manual adjustment or voice adjustment. The target window is the window corresponding to the active transmittance adjustment, that is, the window whose personalized transmittance adjustment does not meet the user's needs. The corrected transmittance is the transmittance of the target window after active adjustment. The target identity information is the identity information corresponding to the target window during this adjustment.

[0097] Specifically, if the user in the car is not satisfied with the result of adjusting the light transmittance of the corresponding window to be adjusted, the user can actively adjust it. If the window active adjustment parameters are received, at least one target window and the corrected light transmittance corresponding to each target window included in the window active adjustment parameters are analyzed and determined. First, the corresponding target windows are adjusted according to the corrected light transmittances, and then, the target identity information corresponding to the target window needs to be determined, that is, the target window is affected by the user in the car corresponding to which identity information. This process can be similar to the process of determining the window to be adjusted corresponding to the user in the car, and will not be repeated here.

[0098] S260 . For each target identity information, update the target color change model corresponding to the target identity information according to the corrected light transmittance of each target vehicle window corresponding to the target identity information.

[0099] Specifically, for each target identity information, there is a situation where the adjusted target transmittance does not meet the demand. Therefore, the corrected transmittance of each target window corresponding to the target identity information is used to feedback adjust the target color change model corresponding to the target identity information to update the corresponding target color change model.

[0100] Based on the above example, the target color change model corresponding to the target identity information can be updated according to the corrected light transmittance of each target window corresponding to the target identity information in the following manner: Determine the target loss according to the preset loss function, the corrected light transmittance of each target window corresponding to the target identity information, and the target light transmittance; Determine the model parameter adjustment gradient based on the target loss; According to the model parameter adjustment gradient, the preset learning rate and each model parameter in the target color change model corresponding to the target identity information, each model parameter is updated; According to the updated model parameters, the target color change model corresponding to the target identity information is updated.

[0101] Among them, the preset loss function can be used to calculate the loss between the target transmittance and the difference between the corrected transmittance and the target transmittance. The specific loss function type can be set according to the needs. The target loss is the calculation result of the preset loss function. The model parameter adjustment gradient is the gradient of the target loss for each model parameter, which can be understood as the gradient of each model parameter (w, b) of the target color change model, where w is the weight and b is the bias. The preset learning rate is a pre-set learning rate that determines whether it can converge to the local minimum and when it converges to the minimum. The specific size can be determined according to the needs and is not limited here.

[0102] Specifically, the corrected light transmittance and the target light transmittance of each target window corresponding to the target identity information are input into a preset loss function for calculation to obtain a target loss. Further, the gradient of the target loss with respect to each model parameter is calculated to obtain a model parameter adjustment gradient. The model parameter adjustment gradient and the preset learning rate are used to adjust each model parameter in the target color change model corresponding to the target identity information to update each model parameter. Furthermore, each updated model parameter is applied to the target color change model to update the target color change model corresponding to the target identity information.

[0103] For example, the transmittance results after adjustment based on each target color change model, i.e., each target transmittance, and the results of secondary adjustments made manually within a short period of time (e.g., 60s), i.e., each corrected transmittance, can be recorded, indicating that the user in the car is not satisfied with the adjustment result of the current target color change model. Therefore, the previous target color change model is corrected in real time so that it will be more inclined to correspond to the preferences of the user in the car when used next time, making the transmittance adjustment more intelligent.

[0104] For the triggering timing of updating the target color change model: based on the transmittance results after adjustment of each target color change model, that is, each target transmittance is y1, the result of the secondary adjustment made by humans, that is, each corrected transmittance is y2, that is, the following formula is obtained: y is the transmittance after active adjustment.

[0105] When y=y1, it means there is no artificial secondary adjustment and the model update will not be triggered. The positive or negative value of y indicates the positive or negative offset.

[0106] Since it involves automatic updating of the model, an online adaptation model can be used to update the model in real time to dynamically adjust the relevant model parameters of each target color change model.

[0107] About problem definition and modeling: Input parameters: including x t:In-car user fatigue, ambient light intensity, road scene information, traffic conditions information, and vehicle status information, and can also include target light transmittance Output parameter y t : Parameters adjusted by the user for the second time. Model parameter θ t : The model parameter set of the target color change model, that is, the initial adjustment strategy.

[0108] About the workflow of automatic model update: First, determine the model parameters of the target color change model , and perform the first model adjustment, for example: Initial_adjustment= , t is the number of adjustments.

[0109] Then, the transmittance after the user manually adjusts the parameters for the second time is obtained , define the loss function: .

[0110] Next, calculate the model parameters θ t Gradient , that is, the model parameter adjustment gradient, which reflects which parameters can be adjusted to reduce the error. The specific formula is: .

[0111] Finally, use online gradient descent to update the model parameters so that the next intelligent adjustment is closer to the user's expectations: .in, It is the preset learning rate, which determines the step size of each update.

[0112] Through the above steps, the online gradient descent algorithm can be effectively used for online updating of the model, so that each target color change model can continuously self-learn and optimize, and gradually adapt to the preferences of each in-car user.

[0113] The above technical solution can automatically adjust the light transmittance according to the light intensity, temperature and other conditions of the external environment, thereby reducing the dependence on air conditioning and lighting systems and reducing energy consumption. In addition, the windows can be controlled to automatically darken under strong sunlight to reduce solar radiation entering the room, thereby reducing the load on the cooling system. Moreover, through the technical solution of this embodiment, the windows can be quickly switched between the transparent and opaque states, and the users in the car can also adjust the privacy protection level instantly as needed without the need for physical curtains or shielding objects. It can also be dynamically adjusted according to indoor activities, time or user needs, providing a flexible privacy protection solution. During driving, the window glass can automatically adjust the light transmittance to reduce strong light reflection or glare, thereby improving driving safety.

[0114] The above technical solution can effectively reduce the glare inside the car and provide a more comfortable visual experience by automatically adjusting the light transmittance of the car windows. It reduces the need for traditional shading devices such as curtains and blinds, thereby reducing the use and waste of related materials. Since no additional traditional shading equipment is required, it can be seamlessly integrated with the vehicle design to improve the overall aesthetics, making the vehicle design more innovative and providing greater space utilization possibilities. Usually, the material of this type of window glass that can adjust the light transmittance also has good sound insulation properties, which can reduce external noise and further improve the quietness and comfort of the car. In addition, this type of window glass usually has higher breakage resistance, providing additional safety protection. In addition, the energy-saving characteristics of the glass material help reduce the carbon footprint, which is in line with the trend of environmental protection and sustainable development.

[0115] The window transmittance adjustment method provided in the embodiment of the present application determines at least one target window corresponding to the window active adjustment parameter and the corrected transmittance corresponding to each target window in response to the received window active adjustment parameter, adjusts the corresponding target windows based on the corrected transmittances, and determines the target identity information corresponding to the target windows according to the target windows, and updates the target color change model corresponding to the target identity information for each target identity information according to the corrected transmittance of each target window corresponding to the target identity information, so that when the user is not satisfied with the result of the automatic adjustment of the window transmittance, the active adjustment is used to correct and adjust the target color change model, and the corrected transmittance is recorded for updating the target color change model, thereby improving the pertinence of the model and the subsequent use effect, making it closer to the preferences of the corresponding user.

[0116] Figure 4 This is a schematic diagram of the structure of a vehicle window transmittance adjustment system provided in an embodiment of the present application. The system uses a vehicle camera, a light sensor, a GPS system, etc. to perceive the current environment and scene of the vehicle in real time, and automatically adjusts the transmittance of the vehicle window in real time based on this information. In addition, it can receive active adjustment instructions from users in the vehicle, and on the basis of actively adjusting the transmittance of the vehicle window, provide feedback and update the model. The system includes: a data acquisition module, a vehicle window transmittance automatic adjustment module, a vehicle window transmittance active adjustment module, a feedback correction module, and an execution module.

[0117] Among them, the data acquisition module is responsible for collecting the status of the driver and passengers (identity, fatigue), ambient light intensity, road scene information, traffic condition information (entering and exiting tunnels, parking status, driving road conditions, etc.) and vehicle status information. It can include ADAS system, and use cameras, light sensors, GPS, etc. to obtain real-time ambient light intensity, road scene information, traffic condition information and vehicle status information. It can also include DMS and OMS systems, which can identify the identity information of the driver and passengers through the camera inside the cabin, and verify the fatigue, so as to facilitate the subsequent combination of feedback correction module and window transmittance automatic adjustment module to further adjust the transmittance of the corresponding window to be adjusted according to the preferences and fatigue of different users in the car, so as to improve driving safety and comfort.

[0118] The window transmittance automatic adjustment module can process data based on the input of the user's identity information, the user's fatigue level, the ambient light intensity, the road scene information, the traffic condition information and the vehicle status information, and output the transmittance corresponding to the window to be adjusted through the calculation of each corresponding target color change model (neural network model), generate a transmittance adjustment instruction, and finally give the instruction to the corresponding actuator in the execution module.

[0119] The feedback correction module is used to automatically adjust the transmittance of each window to be adjusted. If the user is not satisfied with the adjustment result, the user can actively adjust it and compare the transmittance to optimize the model, remember the preferences of the user in the car, and update the model in real time so that the next adjustment will be made according to the preferences of the user in the car, making the model more useful with each use.

[0120] The window transmittance active adjustment module is used to receive manual adjustment instructions and / or voice adjustment instructions from the user in the car, and the voice adjustment quality can be received based on the voice control system. The voice control system can receive the instructions proposed by the user through the voice receiver in the car, and transmit the results to the corresponding actuator in the execution module to achieve the effect of controlling the transmittance of the window through voice.

[0121] The execution module, when receiving the adjustment instruction from the window transmittance automatic adjustment module or the window transmittance active adjustment module, adjusts the amount of current passing through each window through the actuator corresponding to each window to achieve the effect of controlling the transmittance.

[0122] The triggering method of the window transmittance adjustment system provided in the embodiment of the present application is divided into two parts. The first part is the active part triggered by the user in the car issuing a command through voice or manual control, and the second part is the transmittance adjustment part automatically triggered by the neural network model (the target color change model corresponding to each user in the car). In addition, if the user in the car is not satisfied with the adjustment strength within a short period of time after the automatic adjustment and uses the active adjustment again, the model update of the feedback correction module will be triggered, so that adjustments can be made according to the preferences of the user in the car in the future, thereby improving the response speed and effectiveness of the window glass transmittance adjustment and the personalized adjustment effect of the window transmittance adjustment.

[0123] Figure 5 Schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 5 As shown, the electronic device 500 includes one or more processors 501 and a memory 502 .

[0124] The processor 501 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 500 to perform desired functions.

[0125] The memory 502 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 501 may run the program instructions to implement the vehicle window transmittance adjustment method of any embodiment of the present application described above and / or other desired functions. Various contents such as initial external parameters, thresholds, etc. may also be stored in the computer-readable storage medium.

[0126] In one example, the electronic device 500 may further include: an input device 503 and an output device 504, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown). The input device 503 may include, for example, a keyboard, a mouse, etc. The output device 504 may output various information to the outside, including early warning prompt information, braking force, etc. The output device 504 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0127] Of course, to simplify, Figure 5Only some of the components related to the present application in the electronic device 500 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device 500 may also include any other appropriate components.

[0128] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the vehicle window transmittance adjustment method provided in any embodiment of the present application.

[0129] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0130] In addition, the embodiments of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the vehicle window transmittance adjustment method provided in any embodiment of the present application.

[0131] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0132] It should be noted that the terms used in this application are only for describing specific embodiments, rather than limiting the scope of this application. As shown in the specification and claims of this application, unless the context clearly indicates an exception, the words "one", "a", "a kind of" and / or "the" do not specifically refer to the singular, but may also include the plural. The terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the presence of other identical elements in the process, method or device including the elements.

[0133] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0134] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and its core ideas of this application. The above is only the preferred implementation method of this application. It should be pointed out that due to the limitations of textual expression and the objective existence of infinite specific structures, ordinary technicians in this technical field can make several improvements, modifications or changes without departing from the principles of the present invention, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the scope of protection of this application.

Claims

1. A method for adjusting the light transmittance of a vehicle window, characterized in that: include: The user image of each user in the vehicle is obtained through the camera inside the vehicle cabin, and the seat position, identity information and fatigue level of each user in the vehicle are determined based on each user image; Determine the window to be adjusted corresponding to each user in the car according to the seat position of each user in the car; Obtain ambient light intensity, road scene information, traffic condition information, and vehicle status information; For each in-car user, a target color change model corresponding to the in-car user is determined according to the identity information of the in-car user, the fatigue degree of the in-car user, the ambient light intensity, the road scene information, the traffic road condition information and the vehicle status information are input into the target color change model to obtain the target transmittance, and according to the target transmittance, the window to be adjusted corresponding to the in-car user is adjusted.

2. The method according to claim 1, characterized in that: After adjusting the window to be adjusted corresponding to the user in the vehicle according to the target light transmittance, the method further includes: In response to the received window active adjustment parameter, at least one target window corresponding to the window active adjustment parameter and a corrected light transmittance corresponding to each target window are determined, each corresponding target window is adjusted based on each corrected light transmittance, and each target identity information corresponding to the target window is determined according to the target window; For each target identity information, the target color change model corresponding to the target identity information is updated according to the corrected light transmittance of each target vehicle window corresponding to the target identity information.

3. The method according to claim 2, characterized in that The updating of the target color change model corresponding to the target identity information according to the corrected light transmittance of each target window corresponding to the target identity information includes: Determine the target loss according to a preset loss function, the corrected light transmittance of each target window corresponding to the target identity information, and the target light transmittance; Determining a model parameter adjustment gradient according to the target loss; According to the model parameter adjustment gradient, the preset learning rate and each model parameter in the target color change model corresponding to the target identity information, each model parameter is updated; The target color change model corresponding to the target identity information is updated according to the updated model parameters.

4. The method according to claim 1, characterized in that The obtaining of ambient light intensity, road scene information, traffic condition information and vehicle status information includes: Determine the ambient light intensity based on the light sensor; Determine the current location information according to the positioning system, and determine the road scene information according to the current location information and a preset map; Acquire surrounding environment data according to the camera and the radar, and determine traffic condition information according to the surrounding environment data; wherein the traffic condition information includes congestion information and / or traffic light information; Obtain vehicle status information based on the in-vehicle infotainment system.

5. The method according to claim 1, characterized in that The step of determining a target color change model corresponding to the user in the vehicle according to the identity information of the user in the vehicle comprises: In response to the presence of a color change model that matches the identity information among the pre-stored color change models to be selected, the color change model that matches the identity information is used as a target color change model corresponding to the user in the vehicle; In response to the pre-stored color change models to be selected not matching the identity information, obtaining model parameters of the color change models to be selected and model parameters of the reference color change model; Determining average values ​​of the model parameters according to the model parameters of each to-be-selected color change model and the model parameters of the reference color change model; According to the average value of the model parameters and the reference color change model, a color change model to be selected corresponding to the identity information is determined and stored, and the color change model to be selected corresponding to the identity information is determined as a target color change model corresponding to the user in the vehicle.

6. The method according to claim 5, characterized in that The reference color change model is constructed based on the following method: Acquire each sample data and the sample transmittance corresponding to each sample data; wherein the sample data includes sample fatigue, sample ambient light intensity, sample road scene information, sample traffic condition information and sample vehicle status information; The initial color change model is trained according to each sample data and the sample transmittance corresponding to each sample data to obtain a reference color change model, and each model parameter in the reference color change model is determined.

7. The method according to claim 6, characterized in that The obtaining of each sample data and the sample transmittance corresponding to each sample data includes: Acquire each historical time series data and the historical light transmittance corresponding to each historical time series data; wherein the historical time series data includes historical fatigue, historical ambient light intensity, historical road scene information, historical traffic condition information and historical vehicle status information; Perform Gaussian filtering on each historical time series data to obtain each process data; Performing data normalization processing on each process data to obtain each normalized data; Based on the preset window, each normalized data is segmented to obtain each sample data; For each set of sample data, the historical transmittance corresponding to the start timestamp of the sample data is determined as the sample transmittance corresponding to the sample data.

8. The method according to claim 1, characterized in that The step of determining the window to be adjusted corresponding to each user in the vehicle according to the seat position of each user in the vehicle includes: For each in-vehicle user, in response to the in-vehicle user being in the front driving seat, determining the driving-related window and all the front-row windows as the corresponding windows to be adjusted; wherein the driving-related windows include the front windshield window, the rear windshield window and the vehicle sunroof; In response to the user in the vehicle being located in the front passenger seat, determining that the corresponding window to be adjusted is empty; In response to the user in the vehicle being located in a rear passenger seat, determining a target distance between the user in the vehicle and each window in the same row corresponding to the seat row where the user in the vehicle is located; For each window in the same row, in response to the target distance between the user in the vehicle and the window in the same row being the minimum value among the target distances corresponding to the windows in the same row, the window in the same row is determined to be the window to be adjusted corresponding to the user in the vehicle.

9. An electronic device, characterized in that: The electronic device comprises: Processor and memory; The processor is used to execute the steps of the vehicle window transmittance adjustment method as claimed in any one of claims 1 to 8 by calling the program or instruction stored in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute the steps of the vehicle window transmittance adjustment method as claimed in any one of claims 1 to 8.

Images