Self-adaptive light induction adjusting method and device, medium and program product

Abstract

The invention discloses a self-adaptive light induction adjusting method and device, a medium and a program product, and relates to the field of vehicle light control. The method comprises the steps that the environment condition of a vehicle driving area is acquired through a cloud server; inputting the environment condition into a preset light induction adjustment model to obtain light pre-adjustment information; pre-adjusting the light brightness and color temperature of headlamps of the vehicle according to the light pre-adjusting information; under the condition that the light brightness and the color temperature are pre-adjusted, a pre-adjusted light irradiation image in front of the vehicle is obtained; inputting the pre-adjusted lamplight irradiation image into a preset lamplight irradiation condition evaluation model to obtain a lamplight irradiation evaluation result; and if the lamplight irradiation evaluation result is unqualified, finely adjusting the lamplight brightness and color temperature of the headlamp according to the pre-adjusted lamplight irradiation image. The technical problem that a better lighting effect cannot be provided for a driver under the condition of different driving environments due to the lighting condition of the headlamp of the vehicle can be relieved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle lighting control, and in particular to an adaptive lighting sensing adjustment method, device, medium and program product. Background Art

[0002] With the continuous development of automobile technology, the adjustment method of vehicle headlights is also gradually improving. The existing light adjustment methods mainly include manual adjustment and automatic adjustment. The manual adjustment method relies on the driver to adjust the brightness and angle of the headlights based on his own experience and feeling through the control buttons or knobs in the car. However, this method has obvious defects, that is, the timeliness and accuracy of its adjustment are difficult to guarantee.

[0003] As a result, more and more vehicles are using automatic headlight adjustment technology. In related technologies, the automatic adjustment of vehicle headlights mainly uses sensors to collect ambient brightness information, and turns on the headlights when the ambient brightness is lower than the preset value. This makes the vehicle's perception of the external environment relatively simple. In different environments and weather conditions, traditional automatic light adjustment often cannot make more precise adjustments to the lights.

[0004] This may result in the light conditions illuminated by the vehicle headlights being unable to provide the driver with better lighting effects under different driving environments. Summary of the invention

[0005] In view of the above-mentioned technical problems and defects, the purpose of the present invention is to provide an adaptive lighting sensing adjustment method, device, medium and program product, which can alleviate the technical problem that the lighting conditions illuminated by vehicle headlights cannot provide better lighting effects for the driver under different driving environments.

[0006] To achieve the above-mentioned purpose, in a first aspect, the present invention provides an adaptive lighting sensing adjustment method, comprising: obtaining environmental conditions of a vehicle driving area through a cloud server, the environmental conditions including temperature information, humidity information, haze conditions, rain conditions and sunlight conditions; inputting the environmental conditions into a preset lighting sensing adjustment model to obtain lighting pre-adjustment information; pre-adjusting the light brightness and color temperature of the vehicle's headlights according to the lighting pre-adjustment information; obtaining a pre-adjusted lighting image in front of the vehicle when the light brightness and color temperature have been pre-adjusted; inputting the pre-adjusted lighting image into a preset lighting condition evaluation model to obtain a lighting evaluation result; if the lighting evaluation result is unqualified, finely adjusting the light brightness and color temperature of the headlights according to the pre-adjusted lighting image.

[0007] The present invention obtains a wealth of environmental conditions, including temperature, humidity, haze, rain and sunlight conditions, through a cloud server, and no longer relies on the subjective judgment of the driver, greatly improving the timeliness and accuracy of the adjustment. Secondly, for the problem that the traditional headlight automatic adjustment has a single perception of the external environment, this embodiment collects environmental information from multiple dimensions, and can fully understand the various conditions of the vehicle driving area, so that the light sensing adjustment model can generate more accurate light pre-adjustment information based on these complex environmental data, so that more precise adjustments can be made in different environments and weather conditions. Finally, when there is a situation where it may not be possible to provide the driver with a better lighting effect, by obtaining the pre-adjusted light illumination image and inputting it into the evaluation model, the light effect is objectively evaluated, and if it is unqualified, fine adjustment is performed to ensure that the light of the vehicle headlight can accurately and timely provide the driver with a good lighting effect. The present invention effectively alleviates the technical problem that the light conditions of the vehicle headlights cannot provide the driver with a better lighting effect under different driving environments through the above method steps, and breaks through the limitations of the existing vehicle light automatic adjustment technology that the external environment is single and cannot be finely adjusted.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the pre-adjusted lighting illumination image is input into a preset lighting illumination condition evaluation model to obtain a lighting illumination evaluation result, specifically including: determining the lighting illumination area in the pre-adjusted lighting illumination image through the lighting illumination condition evaluation model; determining the image key features of the lighting illumination area through the lighting illumination condition evaluation model, the image key features including light intensity, color distribution and clarity of the lighting area; when the difference between the image key features and the expected standard features is greater than a set threshold, determining the lighting illumination evaluation result as unqualified.

[0009] By adopting the technical solution of the above embodiment, the lighting area and key features are determined through the lighting condition evaluation model, and the lighting is judged as unqualified when the key features are greatly different from the expected standard features, which can ensure that the lighting effect of the headlights meets the actual needs. Accurately evaluating the light intensity, color distribution and clarity can timely discover lighting problems and make adjustments to improve driving safety. At the same time, the judgment method based on the difference value and threshold makes the evaluation results more objective and accurate, provides a clear basis for the precise adjustment of the lights, helps to achieve adaptive lighting adjustment, and improves the driving experience.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the brightness and color temperature of the headlights are finely adjusted based on the pre-adjusted lighting illumination image, including: determining the light fine adjustment information based on the difference between the key features of the image and the expected standard features; and finely adjusting the brightness and color temperature of the headlights based on the light fine adjustment information.

[0011] By adopting the technical solution of the above embodiment, the light precision adjustment information is determined according to the difference between the key features of the image and the expected standard features, and the brightness and color temperature of the headlights can be accurately adjusted. This method can make targeted adjustments to specific lighting problems and avoid the adverse effects of blind adjustments. By accurately adjusting the lighting parameters, the adaptability and lighting effects of the headlights in different environments can be improved, providing the driver with a clearer and more comfortable visual environment, thereby improving driving safety and comfort, while also reducing the risk of visual fatigue and misjudgment caused by inappropriate lighting.

[0012] In combination with some embodiments of the first aspect, in some embodiments, when the difference between the key features of the image and the expected standard features is greater than a set threshold, before determining that the lighting evaluation result is unqualified, it also includes: determining the expected standard features from a preset standard feature database based on the environmental conditions, and the standard feature database stores the correspondence between different environmental conditions and different standard features.

[0013] By adopting the technical solution of the above embodiment, the expected standard features are determined from the standard feature database according to the environmental conditions, making the lighting evaluation more scientific and reasonable. Different environments correspond to different standard features, ensuring that the lighting can achieve the best lighting effect in various complex situations. This helps to improve the accuracy and reliability of lighting evaluation and provides a clear goal and reference basis for lighting adjustment. At the same time, through the corresponding relationship in the database, it is possible to quickly adapt to different environmental changes, realize adaptive lighting adjustment, and improve driving safety and comfort.

[0014] In combination with some embodiments of the first aspect, in some embodiments, the method also includes: obtaining a facial image of the driver of the vehicle during driving of the vehicle; judging whether the driver of the vehicle is in a fatigue state based on the facial image of the driver; and if so, controlling the headlights to flash at a set frequency to remind the driver.

[0015] By adopting the technical solution of the above embodiment, the driver's facial image is obtained to determine the fatigue state and the headlights are controlled to flash as a reminder, which can effectively improve driving safety. Timely detection of the driver's fatigue state and giving obvious visual stimulation can prompt the driver to be more vigilant and avoid accidents caused by fatigue driving. This active safety measure helps to reduce traffic accidents, protect the lives and property of drivers and passengers, and also reflects the intelligent and humanized design of the vehicle system.

[0016] In combination with some embodiments of the first aspect, in some embodiments, the method also includes: during the driving of the vehicle, obtaining the speed and distance of the oncoming vehicle through the vehicle networking system; determining the low beam switching location based on the speed and distance of the oncoming vehicle and the vehicle's own driving speed; when the vehicle drives to the low beam switching location, switching the vehicle's headlights from high beam to low beam.

[0017] By adopting the technical solution of the above embodiment, the low beam switch location can be determined by obtaining the information of the oncoming vehicle through the Internet of Vehicles, which can avoid the glare interference of the oncoming vehicle driver and improve driving safety. Intelligent switching of the headlights according to factors such as speed and distance helps to optimize the road lighting environment and reduce traffic accidents caused by improper use of lights. At the same time, this adaptive light switching method improves the convenience and comfort of driving, so that the driver does not need to manually operate the lights frequently and can focus more on driving.

[0018] In combination with some embodiments of the first aspect, in some embodiments, the method also includes: obtaining performance index data of the headlights; inputting the performance index data of the headlights into a preset headlight fault prediction model to determine the failure probability of the headlights in the subsequent driving process; if the failure probability is greater than the set probability threshold, issuing an alarm prompt to the driver of the vehicle.

[0019] By adopting the technical solution of the above embodiment, the headlight performance index data is obtained and the headlight fault prediction model is used to determine the fault probability, which can provide early warning of headlight faults and ensure driving safety. An alarm prompt is issued to the driver in a timely manner so that the driver can take corresponding measures, such as checking and repairing the headlights as soon as possible or adjusting the driving style to avoid dangerous situations caused by sudden headlight failures. This preventive maintenance method helps to improve the reliability and stability of the vehicle and reduce the inconvenience and safety hazards caused by headlight failures.

[0020] In a second aspect, an embodiment of the present invention provides an electronic device, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation method of the first aspect or the second aspect.

[0021] In a third aspect, the present invention provides a computer-readable storage medium comprising instructions, which, when executed on the electronic device, causes the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation of the first aspect or the second aspect.

[0022] In a fourth aspect, the present invention provides a computer program product comprising instructions, which, when the computer program product is run on the electronic device, enables the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation of the first aspect or the second aspect.

[0023] It is understandable that the electronic device provided in the second aspect, the storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the method provided in the present invention. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided by the present invention have at least the following technical effects or advantages: 1. Obtain environmental conditions through the cloud server, use the light sensing adjustment model and the light exposure evaluation model to achieve adaptive adjustment of the brightness and color temperature of the vehicle headlights. In different environments such as temperature, humidity, haze, rain and sunlight, the system can automatically adjust the lighting parameters to ensure that the headlights provide the best lighting effect. For example, in haze weather, the system will lower the color temperature of the light to make it more warm to enhance penetration, while increasing the brightness to compensate for the light loss caused by haze. At night or when there is insufficient light, increase the light brightness and adjust the color temperature to ensure a good field of vision. This adaptive adjustment not only improves the driver's visibility in various environments, but also reduces the risk of visual fatigue and misjudgment caused by inappropriate lighting, greatly improving driving safety.

[0025] 2. The present invention integrates multiple intelligent functions, such as judging the fatigue state based on the driver's facial image and controlling the flashing of the headlights to remind, obtaining the information of oncoming vehicles through the Internet of Vehicles to determine the low beam switch location, and using the headlight failure prediction model to warn of headlight failure in advance. These functions provide drivers with a more convenient and comfortable driving experience. When the driver is tired, the flashing reminder of the headlights can prompt the driver to be vigilant and avoid fatigue driving accidents. When meeting oncoming vehicles, the automatic switching of the low beams avoids glare interference to the oncoming driver and improves driving safety. The headlight failure prediction model can notify the driver of possible headlight failures in advance, giving the driver enough time to take measures to avoid dangerous situations caused by sudden headlight failures. The integration of these intelligent functions makes the driving process easier and safer.

[0026] 3. Comprehensively manage vehicle headlights from multiple aspects. From obtaining environmental information for pre-adjustment and fine adjustment, to evaluating the lighting effect, to responding to the driver's status and predicting headlight failures, the entire system forms a closed-loop management system. This comprehensive management method ensures that the headlights are always in the best working condition and improves the reliability and stability of the vehicle. At the same time, through real-time monitoring of headlight performance indicators and fault prediction, potential problems can be discovered and repaired in a timely manner, reducing the inconvenience and safety hazards caused by headlight failures. In addition, the system's adaptive adjustment and intelligent functions also reduce the driver's manual operation of the lights, reduce the operating burden, and further improve the reliability and ease of use of the vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings herein are incorporated into and constitute a part of the specification, showing embodiments consistent with the present invention, and together with the specification, are used to explain the principles of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 is a flow chart of an adaptive lighting sensing adjustment method according to an embodiment of the present invention; Figure 2 is a flow chart of another adaptive lighting sensing adjustment method according to an embodiment of the present invention; Figure 3 It is a schematic diagram of the architecture of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to be limiting of the present invention. As used in the specification of the present invention, the singular expressions "a", "a", "above", "the" and "this" are intended to also include plural expressions, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present invention refers to any or all possible combinations comprising one or more of the listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood as implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, unless otherwise specified, the meaning of "plurality" is two or more.

[0030] It should also be noted that, unless otherwise clearly specified and limited, in the embodiments of the present invention, the terms such as "setting" and "connection" should be understood in a broad sense. For example, "connection" 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 an indirect connection through an intermediate medium, it can be the internal connection of two components; it can be a wired communication connection or a wireless communication connection. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The embodiments of the present invention are described in detail below.

[0031] The embodiment of the present invention provides an adaptive light sensing adjustment method, which effectively alleviates the technical problem that the light conditions of vehicle headlights cannot provide better lighting effects for drivers under different driving environments, and breaks through the limitations of existing vehicle light automatic adjustment technology that the perception of the external environment is single and cannot be finely adjusted. This embodiment first obtains a variety of environmental conditions through a cloud server, including temperature, humidity, haze, rain and sunlight conditions, etc., and no longer relies on the subjective judgment of the driver, which greatly improves the timeliness and accuracy of the adjustment. Secondly, for the problem that the traditional headlight automatic adjustment has a single perception of the external environment, this embodiment collects environmental information from multiple dimensions, which can fully understand the various conditions of the vehicle driving area, so that the light sensing adjustment model can generate more accurate light pre-adjustment information based on these complex environmental data, so that more fine adjustments can be made in different environments and weather conditions. Finally, when there is a situation where it may not be possible to provide the driver with a better lighting effect, by obtaining the pre-adjusted light illumination image and inputting it into the evaluation model, the light effect is objectively evaluated, and if it is unqualified, fine adjustment is performed to ensure that the light of the vehicle headlight can accurately and timely provide the driver with a good lighting effect.

[0032] Combine the following Figure 1 The method of this embodiment is specifically described, which can be applied to a vehicle lighting control system (hereinafter referred to as a system or control system), and specifically includes the following steps: Step 201, obtaining the environmental conditions of the vehicle driving area through the cloud server.

[0033] Among them, environmental conditions include temperature information, humidity information, haze conditions, rain conditions and sunlight conditions.

[0034] Specifically, for temperature information, data from meteorological monitoring stations distributed in different geographical locations can be used. These stations collect and upload temperature data to the cloud server in real time. The system establishes a connection with the cloud server to obtain the temperature values ​​recorded by meteorological stations near the driving area, thereby determining the current ambient temperature. For humidity information, the measurement results of meteorological monitoring equipment are also used. These devices upload humidity data to the cloud server, and the system retrieves humidity data for the corresponding area from the cloud server based on its own positioning. In terms of obtaining haze conditions, some professional environmental monitoring agencies will set up haze monitoring sensors in different regions. These sensors continuously detect indicators such as the concentration of particulate matter in the air to judge the degree of haze, and transmit data to the cloud server, from which the system can obtain the haze conditions in the driving area. The raining situation can be comprehensively judged by the data of meteorological radar and ground rainfall monitoring stations. The meteorological radar can monitor a large range of precipitation areas and intensity changes, and the rainfall monitoring station provides more accurate ground rainfall data. The two are combined and uploaded to the cloud server for system query. The sunlight conditions can be monitored using satellite remote sensing data and ground-based light sensor data. Satellites can monitor the intensity and distribution of light over a large area from high altitudes, while ground-based light sensors can more accurately measure the actual sunlight conditions at the vehicle's location. After these data are aggregated to the cloud server, the system can obtain the sunlight conditions in the driving area.

[0035] Step 202: Input the environmental conditions into a preset light sensing adjustment model to obtain light pre-adjustment information.

[0036] Among them, the light sensing adjustment model is a machine learning model that has been trained and optimized with a large amount of data. It can understand the actual meaning of parameters in different environmental conditions and their impact on vehicle headlight lighting.

[0037] When environmental data such as temperature information, humidity information, haze conditions, rain conditions, and sunlight conditions are input into the model, the light sensing and adjustment model will quickly analyze these data. For example, for temperature information, the model will consider the impact of different temperature ranges on light propagation and visual effects. High temperature may change the refraction characteristics of the air, thereby affecting the illumination path and intensity of the light. For humidity information, water vapor in the air under high humidity conditions may scatter the light, and the model will adjust the output parameters of the light accordingly to adapt to this situation. For haze conditions, the model will identify characteristics such as haze concentration and particle size. Because haze can seriously weaken the penetration of light, the model will adjust the color temperature of the light to a lower value accordingly, making it more warm to enhance the ability to propagate in haze, and may increase the brightness of the light to compensate for the light loss caused by haze. For rainy conditions, the model will consider the reflection of rain on the light and determine the appropriate light brightness and angle adjustment strategy based on the amount of rain. The sunlight exposure conditions determine the working mode of the headlights under different lighting conditions during the day and at night. When the sunlight is strong during the day, you may need to reduce the brightness of the headlights or even turn them off. At night or when there is insufficient light, you need to increase the brightness and adjust the color temperature to ensure a good view.

[0038] Through comprehensive analysis of these environmental conditions, the lighting sensing and adjustment model ultimately outputs lighting pre-adjustment information, providing precise guidance for subsequent headlight adjustment.

[0039] In some embodiments, the light sensing adjustment model training process is as follows: 1. Data preparation stage: Collect diverse environmental data: Rich and comprehensive data is collected from multiple professional data sources, including weather station data recorded in different regions, seasons and time periods, covering various temperature ranges, humidity changes, haze concentrations, rainfall intensity, and sunlight duration and intensity. At the same time, combined with first-hand environmental data collected by test vehicles equipped with professional monitoring equipment during actual driving, ensure that the data can accurately reflect the environmental conditions in real driving scenarios.

[0040] Collect corresponding lighting adjustment data: Professionals are arranged to evaluate the actual lighting effects of vehicle headlights in different environments, and manually adjust the light brightness and color temperature, and record the specific values ​​of light brightness and color temperature that can provide the best visual experience and lighting effects for each set of environmental conditions, which are used as the target output value (label data) during training. For example, in severe haze weather, after multiple tests and adjustments, it is determined that the light brightness that allows the driver to clearly see the road conditions ahead at a certain distance is a certain value, and the color temperature is a relatively warm value. This combination is matched one by one with environmental data such as the haze concentration at the time to build a complete training data set.

[0041] 2. Data preprocessing stage: Data cleaning: The collected environmental data and lighting adjustment data are carefully checked to eliminate data records with obvious errors, too many missing values, or that do not conform to actual logic. For example, if the temperature value of a data point exceeds the extreme value range that may appear in the normal environment of the earth, or the light brightness value has an unreasonable maximum or minimum value, it will be removed to ensure the quality and reliability of subsequent training data.

[0042] Data normalization / standardization: Since the different types of data collected have very different numerical ranges, for example, the temperature may vary from dozens of degrees below zero to dozens of degrees above zero, while the humidity ranges from 0 to 100%, and the light brightness also has its own specific value range. In order to enable the model to learn data features more efficiently and stably during training, normalization or standardization methods will be used to uniformly process these data into a specific range, such as mapping all data to the range of 0 to 1, or making it conform to the standard normal distribution with a mean of 0 and a variance of 1.

[0043] 3. Model selection stage: You can choose a suitable model architecture based on actual needs and data characteristics. Common choices are: Neural network models (such as multi-layer perceptron MLP): It has powerful nonlinear fitting capabilities and can handle complex input-output relationships. For input environmental data (such as temperature, humidity and other dimensions), it can be transmitted and calculated through multiple neurons layer by layer, and finally output the corresponding predicted values ​​of light brightness and color temperature. Moreover, by adjusting hyperparameters such as the number of neurons, the number of layers, and the activation function, it can flexibly adapt to different training effect requirements.

[0044] Convolutional Neural Networks (CNN) in Deep Learning: Although it is often used for image-related tasks, it can also play its advantages here. Environmental data can be structured according to a certain time series or spatial distribution (such as a combination of meteorological data in different regions, etc.) to make it similar to the two-dimensional structure of an image. CNN can automatically extract deep features, such as the correlation features between different environmental factors, to better predict the appropriate lighting adjustment parameters.

[0045] Recurrent Neural Networks (RNN) and their variants (such as LSTM, GRU): Taking into account the possible correlation between environmental data in time series, such as the impact of the continuous change in sunlight from day to night on lighting regulation, RNN models can effectively learn and process this sequence information, remember the environmental status at previous moments, and then more accurately predict the corresponding light brightness and color temperature at the current moment.

[0046] 4. Model training stage: Divide into training set, validation set and test set: Usually, the preprocessed data set is divided into training set, validation set and test set according to a certain ratio (such as the common 7:2:1 or 8:1:1, etc.). The training set is used to let the model learn the rules in the data and adjust its own parameters; the validation set is used to evaluate the performance of the model during the training process, prevent overfitting, and adjust the model's hyperparameters (such as learning rate, number of layers, etc.) according to the performance on the validation set; the test set is used to objectively and impartially evaluate the generalization ability and prediction accuracy of the model on actual unseen data after the model training is completed.

[0047] Determine the loss function and optimization algorithm: Choose a suitable loss function to measure the difference between the light brightness and color temperature values ​​predicted by the model and the actual optimal values. The common loss function is the mean square error (MSE), which can calculate the average of the squared errors between the predicted value and the true value, and directly reflect the accuracy of the prediction. At the same time, use an optimization algorithm to update the model parameters to minimize the value of the loss function, such as stochastic gradient descent (SGD) and its improved algorithms Adagrad, Adadelta, Adam, etc. These optimization algorithms can adaptively adjust the learning rate according to different gradient conditions, speed up the model convergence and improve the training effect.

[0048] Start the training process: Input the training set data into the selected model, and perform multiple rounds of iterative training according to the set loss function and optimization algorithm. In each round of training, the model tries to predict the light brightness and color temperature based on the input environment, calculates the loss value with the real label, and then back-propagates the error through the optimization algorithm, adjusts the weight parameters inside the model, and continuously optimizes the prediction results until the preset training rounds are reached or the performance indicators on the validation set no longer improve.

[0049] 5. Model evaluation and optimization stage: Model Evaluation: Use the test set data to evaluate the trained model, calculate evaluation indicators such as mean square error and mean absolute error, and check the accuracy of the model's prediction of light brightness and color temperature on data that has not participated in training. At the same time, you can also invite professional drivers to actually experience the lighting effects adjusted based on the model in various simulated environmental scenarios, collect subjective evaluation feedback, and measure the pros and cons of the model from the perspective of actual driving experience.

[0050] Model optimization: If the evaluation results show that the model performance does not meet expectations, such as large prediction errors or poor actual driving experience, you can analyze the reasons and try to adjust the model architecture, add more data, change the data preprocessing method or optimize hyperparameters, etc., re-train and evaluate, and continuously iterate and optimize until you get a high-performance lighting sensing and adjustment model that can accurately predict reasonable light brightness and color temperature based on environmental conditions.

[0051] Step 203: pre-adjust the light brightness and color temperature of the vehicle's headlights according to the light pre-adjustment information.

[0052] Specifically, after obtaining the light pre-adjustment information output from the light sensing adjustment model, the vehicle's headlight control system will respond quickly and perform corresponding adjustment operations. If the pre-adjustment information indicates that the light brightness needs to be increased, the control system will increase the light output intensity by adjusting parameters such as the current or power of the headlight. For example, in environmental weather conditions such as night or rainy days, in order to ensure that the driver can clearly see the road and obstacles ahead, it is necessary to increase the light brightness to enhance the lighting effect. As for the adjustment of color temperature, the control system will change the color temperature of the headlight light source to make it more in line with the current environmental requirements. For example, in haze weather, warm-toned lights with lower color temperatures have better penetration, can reduce the scattering of light by haze, and improve the driver's field of vision clarity and visibility of the path ahead.

[0053] Under different environmental conditions, by pre-adjusting the light brightness and color temperature, the vehicle headlights can quickly and initially adapt to changes in the surrounding environment, providing the driver with a safer and more comfortable driving experience.

[0054] In this embodiment, the vehicle headlights use high-quality light-emitting elements that can provide stable and high-intensity lighting. It is equipped with a precise current regulating device inside, which can quickly and accurately adjust the light brightness according to the system instructions. For color temperature adjustment, the headlights are built with switchable filters or use a combination of multi-color light-emitting chips, which can achieve accurate changes in color temperature under different environmental requirements. At the same time, the headlights are closely connected to the vehicle's sensor system and control system, and receive instructions from environmental perception sensors and light sensing adjustment models through high-speed data transmission lines to ensure that the real-time adjustment of the lights is synchronized with changes in the vehicle's driving environment. In addition, the outer shell of the headlight is made of a sturdy and durable material with good heat dissipation performance, which ensures long-term stable operation in various harsh environments and provides reliable lighting protection for the driver.

[0055] Step 204 , when the light brightness and color temperature have been pre-adjusted, obtain a pre-adjusted light illumination image in front of the vehicle.

[0056] Specifically, after the pre-adjustment of the light brightness and color temperature is completed, the system will obtain the pre-adjusted light illumination image in front of the vehicle through the high-definition camera installed in front of the vehicle. This camera has been calibrated and set up specifically to accurately capture the actual situation of the headlights illuminating the road ahead, objects and surrounding environment in the current pre-adjustment state. The camera will continuously capture images at a certain frame rate to ensure that the dynamic changes of the light in different scenes can be reflected in a timely manner.

[0057] The system uses image acquisition technology to transmit and process the pre-adjusted lighting images in real time, so as to accurately evaluate the lighting effect later. These pre-adjusted lighting images not only include the area illuminated by the light, but also show key information such as the intensity distribution, color performance, and lighting clarity of different objects, which provides an intuitive and important basis for the system to further judge whether the lighting adjustment has achieved the best effect.

[0058] Step 205: input the pre-adjusted lighting image into a preset lighting condition evaluation model to obtain a lighting condition evaluation result.

[0059] Among them, the lighting condition evaluation model has been trained with a large amount of labeled data and can extract multiple key features from the image to determine whether the lighting condition is good.

[0060] First, the lighting condition evaluation model analyzes the light intensity distribution in the image to determine whether there are areas that are too bright or too dark. If the brightness of some areas is too high, it may cause glare to other road users; if the brightness of some areas is too low, it may affect the driver's observation of the area.

[0061] Secondly, the lighting condition evaluation model will evaluate the color distribution in the image to ensure that the color temperature of the light meets the current environmental requirements. It should not be too high, causing the picture to be cold and affecting object recognition, nor too low, causing the picture to be yellowish and lack clarity.

[0062] In addition, the lighting evaluation model also checks whether there are obvious shadows or unevenly lit areas in the image, and whether the light can clearly illuminate key objects on the road ahead, such as traffic signs, pedestrians, and other vehicles.

[0063] Through comprehensive analysis of these characteristics, the model finally gives a lighting evaluation result, indicating whether the current lighting conditions are qualified.

[0064] In this embodiment, the lighting condition evaluation model training process is as follows: 1. Data collection stage: Collect diverse image data: In a variety of actual driving scenarios, a large number of lighting scene images are collected using a high-definition camera installed in front of the test vehicle. These scenarios cover different weather conditions, such as sunny, rainy, foggy, snowy, etc.; different times, including daytime, dusk, and night; and different road conditions, such as urban roads, rural roads, and highways. Ensure that the collected images can fully reflect the actual situation of vehicle headlights in various environments, and each image corresponds to the specific light brightness and color temperature settings at that time.

[0065] Annotated image data: Organize a professional evaluation team to annotate each collected lighting scene image according to the established standards. The annotation content mainly revolves around whether the lighting effect is qualified, and details the specific factors that affect the qualification, such as: Light intensity: mark whether there are overly bright areas in the image that may cause glare, or overly dark areas that affect line of sight; the uniformity of light will also be described to determine which areas have uneven light distribution.

[0066] Color temperature performance: Determine whether the overall color temperature of the picture is suitable for the current environment, such as whether it is a warm color with better penetration in foggy weather, or whether the color temperature is too high to see the road conditions clearly, and mark it out.

[0067] Object recognition: Check whether key objects on the road, such as traffic signs, pedestrians, and vehicles, can be clearly identified under lighting, and mark the objects with high or low recognition.

[0068] 2. Data preprocessing stage: Image cleaning and screening: Carefully check the collected images and eliminate those that are blurred due to shooting errors, blocked, or have poor image quality and cannot accurately judge the lighting effects. Ensure that the data images entering the subsequent training process are clear and can accurately analyze the lighting conditions.

[0069] Image normalization: In order to enable the model to better learn the features in the image, the image will be normalized. Usually, the pixel values ​​of the image are normalized to a specific range. For example, the pixel value of each channel of the RGB image is adjusted from the original 0255 range to the range of 01 through linear transformation, etc., so that images taken under different lighting conditions have better comparability and standardization at the data level, which facilitates model convergence and training.

[0070] 3. Model selection stage: Selection of deep learning model: Based on the characteristics of image data and the complex lighting conditions to be evaluated, convolutional neural networks (CNNs) and their related variants are often used as the basic model architecture. CNNs have a powerful ability to automatically extract image features and can mine deep features such as changes in light intensity and color distribution from raw images. For example, classic VGG networks, ResNet networks and other architectures can be used, or these architectures can be appropriately improved and adjusted according to actual needs to make them more suitable for lighting condition evaluation tasks.

[0071] Combine with other models or modules (optional): Sometimes we also consider combining other types of models or modules to further improve the evaluation effect. For example, adding a fully connected layer on the basis of CNN to form a complete neural network. The fully connected layer can integrate and further nonlinearly map the features extracted by CNN to output the final evaluation results. Or we can integrate some attention mechanism modules to allow the model to focus more on key areas when analyzing images, such as the glare situation in the area where oncoming vehicles are located, the light clarity at the location of road signs, etc., so as to improve the accuracy of the evaluation.

[0072] 4. Model training stage: Divide the dataset: The preprocessed labeled image data is divided into training set, validation set and test set in appropriate proportions, with common ratios such as 7:2:1 or 8:1:1. The training set is used to allow the model to learn the mapping relationship between image features and annotation evaluation results; the validation set is used to monitor the performance of the model during training to prevent overfitting, and to adjust the model's hyperparameters, such as learning rate and number of network layers, through the evaluation results on the validation set; the test set is used to objectively measure the model's generalization ability and evaluation accuracy on unseen data after training.

[0073] Determine the loss function and optimization algorithm: Select an appropriate loss function to measure the difference between the evaluation results predicted by the model and the actual labeled results. For such binary classification (qualified or unqualified) or multi-classification (for example, subdivided into multiple evaluation levels such as excellent, good, qualified, and unqualified), the commonly used loss function is the cross-entropy loss function (CrossEntropyLoss). At the same time, the corresponding optimization algorithm is used to update the parameters of the model to continuously reduce the value of the loss function, such as stochastic gradient descent (SGD) and its improved algorithms Adagrad, Adadelta, Adam and other optimization algorithms, which adaptively adjust the learning rate according to the gradient of the training data, so that the model can converge to a better parameter state more efficiently.

[0074] Start the training process: The training set image data is input into the selected model architecture, and multiple rounds of iterative training are performed according to the set loss function and optimization algorithm. In each round of training, the model attempts to predict the lighting evaluation results based on the input image, and then calculates the loss value between the actual annotation results, and transfers the error from the output layer to the input layer through the back propagation mechanism, adjusts the weight parameters of each layer of the model, and continuously optimizes the prediction effect until the preset training round limit is reached or the performance indicators on the validation set no longer improve.

[0075] 5. Model evaluation and optimization stage: Model Evaluation: The trained model is fully evaluated using the test set image data, and the relevant evaluation indicators are calculated. For binary classification tasks, the main indicators are accuracy, precision, recall, F1 value, etc. For multi-classification tasks, the macro average and micro average are also considered to determine the accuracy of the model's evaluation of lighting conditions on image data that has not participated in training. At the same time, professional drivers are invited to conduct subjective verification of the model's evaluation results in simulated actual driving scenarios to see whether the evaluation results are consistent with their feelings about the lighting effects during actual driving.

[0076] Model optimization: If the evaluation results show that the model performance does not meet expectations, such as low accuracy or poor subjective feedback from drivers, it is necessary to conduct in-depth analysis of the reasons, which may be caused by factors such as insufficient data volume, insufficient model complexity, or unreasonable hyperparameter settings. At this time, targeted measures can be taken, such as collecting more annotated image data, adjusting the model architecture to increase its complexity, changing the optimization algorithm, or re-fine-tuning the hyperparameters, and then training and evaluating again, and continuously iterating and optimizing until a high-performance model that can accurately and reliably evaluate lighting conditions is obtained.

[0077] Step 206: If the lighting evaluation result is unqualified, the headlight brightness and color temperature are finely adjusted according to the pre-adjusted lighting image.

[0078] Specifically, the system will analyze the specific problem areas in the image again. If it finds an area with low brightness, it will increase the brightness of the headlights appropriately, perhaps by increasing the current or adjusting the power of the light-emitting elements. If the brightness is too high, it will reduce the brightness to avoid glare.

[0079] Regarding the adjustment of color temperature, if the image shows a cold or warm color that is not suitable for the current environment, the system will adjust the color temperature value according to the specific situation to make it closer to the optimal lighting effect. For example, on a foggy day, if the image shows that the light color temperature is too high and the penetration is insufficient, the system will lower the color temperature to make the light more warm to enhance the penetration in the fog. Among them, yellow light with a lower color temperature (about 3000K) has better penetration in low-visibility environments such as haze, because its wavelength is longer and is not easily scattered and absorbed by particles in the air, while blue light or white light with a higher color temperature (about 5000K or above) has relatively weak penetration.

[0080] At the same time, the system will also take other environmental factors, such as temperature, humidity, etc. into comprehensive consideration to ensure that the adjusted light brightness and color temperature can provide the driver with the best lighting conditions in the current environment, thereby improving driving safety and comfort.

[0081] This embodiment adopts the above method steps. First, a comprehensive environmental condition is obtained through a cloud server, including temperature, humidity, haze, rain and sunlight conditions, etc., which provides a rich and accurate basis for light adjustment and makes the adjustment more scientific and reasonable. The preset light sensing adjustment model can quickly generate light pre-adjustment information based on these complex environmental information, greatly improving the timeliness and accuracy of adjustment, and avoiding the problem of imprecise adjustment caused by the single environmental perception of the traditional method. Then, after pre-adjustment, the pre-adjusted light illumination image in front of the vehicle is obtained, and the light illumination condition evaluation model is used for objective evaluation to ensure accurate judgment of the light effect. If the evaluation result is unqualified, fine adjustment can be performed according to the image, which further improves the accuracy and adaptability of light adjustment, provides the driver with the best lighting effect, and improves driving safety and comfort. At the same time, the whole method realizes fully automatic adjustment without manual intervention by the driver, reduces the driver's operating burden, and makes the driving process more convenient and efficient.

[0082] Combine the following Figure 2 To further illustrate the method of this embodiment, it specifically includes the following steps: Step 301, obtaining the environmental conditions of the vehicle driving area through the cloud server.

[0083] This step can refer to the description in the above embodiment and will not be described again here.

[0084] Step 302: Input the environmental conditions into a preset light sensing adjustment model to obtain light pre-adjustment information.

[0085] This step can refer to the description in the above embodiment and will not be described again here.

[0086] Step 303: pre-adjust the light brightness and color temperature of the vehicle's headlights according to the light pre-adjustment information.

[0087] This step can refer to the description in the above embodiment and will not be described again here.

[0088] Step 304 , when the light brightness and color temperature have been pre-adjusted, obtain a pre-adjusted light illumination image in front of the vehicle.

[0089] This step can refer to the description in the above embodiment and will not be described again here.

[0090] Step 305: determine the lighting area in the pre-adjusted lighting image by using the lighting condition evaluation model.

[0091] Specifically, the lighting condition evaluation model first determines the lighting area in the pre-adjusted lighting image. This process usually uses image analysis technology to determine the range illuminated by the headlights by identifying the parts of the image that are significantly brighter than the surrounding environment.

[0092] For example, a brightness threshold can be set to group together pixels with brightness higher than the threshold in the image, and the area formed by these pixels is initially identified as the lighting area. At the same time, combined with the edge detection algorithm of the image, the boundaries of the lighting area can be further accurately divided to exclude misjudged areas caused by factors such as reflected light from the surrounding environment, ensuring accurate locking of the area actually illuminated by the headlights, providing a clear target range for subsequent analysis.

[0093] Step 306: determine the key image features of the lighting area through the lighting condition evaluation model.

[0094] Among them, the key features of the image include light intensity, color distribution and clarity of the lighting area.

[0095] Light intensity refers to the quantitative expression of the brightness of light in the lighting area, which can usually be measured by the brightness value of the pixels in the area, and affects the driver's visual effect of the road and the surrounding environment. The lighting condition evaluation model can calculate the average brightness value of the pixels in the lighting area, or the statistical information of the brightness distribution, to measure the light intensity. If the average brightness value is too high or too low, it may affect the driver's visual effect and judgment of the road conditions.

[0096] The color distribution indicates the proportion and presentation status of different color components in the lighting area. The lighting evaluation model can be analyzed through color histograms, etc., which is related to the color temperature of the light and its adaptability to different environments. The color histogram of the lighting area can be analyzed to determine whether the color temperature of the light is appropriate. For example, if the color distribution tends to be too blue, it may mean that the color temperature is too high, which is not conducive to seeing objects clearly in certain environments; if it tends to be too yellow, the color temperature may be too low, which will also affect visual clarity.

[0097] The clarity of the illuminated area is an indicator that describes the degree of clarity of the edges and contours of objects under lighting. It is evaluated with the help of image sharpening and edge detection algorithms, and is related to whether the driver can accurately identify objects on the road. The lighting evaluation model can use image sharpening algorithms or edge detection algorithms to evaluate the clarity of the edges of objects under lighting. If the edges of objects are blurred, it means that the clarity of the illuminated area is not enough, which may affect the driver's recognition of objects on the road.

[0098] Step 307, determining the expected standard features from a preset standard feature database according to the environmental conditions.

[0099] The standard feature database stores the corresponding relationship between different environmental conditions and different standard features. This standard feature database is established through a large number of experiments and data accumulation.

[0100] Specifically, the system will query from the preset standard feature database according to the current environmental conditions. For example, when a specific temperature and humidity combination, as well as specific conditions such as haze, rain or sunlight are detected, the standard feature database will have a set of expected standard features corresponding to these specific conditions regarding light intensity, color distribution and clarity of the illuminated area.

[0101] For different environments, these expected standard features are determined based on research and actual testing of optimal lighting effects. For example, in dense fog, the expected standard features may include lower light intensity to avoid excessive scattering of light in the fog and cause glare, warmer color distribution to enhance the penetration of light in the fog, and higher illumination area clarity requirements to ensure that the driver can clearly identify objects in front.

[0102] In this way, the system can accurately determine the expected standard features that best suit the current environment from the database based on real-time environmental conditions, providing a scientific and reasonable reference for lighting evaluation.

[0103] Step 308: compare the key features of the image with the expected standard features to determine whether the difference between the two is greater than a set threshold.

[0104] Among them, for the light intensity, we must first obtain the actual light intensity value in the lighting area, which may be obtained by calculating the average value, median value of the pixel brightness in the area, or comprehensively considering the brightness distribution to obtain a quantitative value representing the overall light intensity. The light intensity value in the expected standard feature is an ideal value range pre-set according to different environmental conditions. For example, in a rural road environment without street lights at night, it may be determined that the appropriate light intensity average value is in a certain specific interval. Next, calculate the difference between the actual light intensity value and the expected standard light intensity value, use the absolute value of the difference to measure the degree of deviation between the two, and then compare this deviation with the set threshold. If the absolute value of the difference is greater than the set threshold, it means that the current actual light intensity deviates too much from the ideal state.

[0105] In terms of color distribution, the actual characteristics of color distribution, such as the proportion of different color channels and the distribution of hues, are obtained by analyzing the color histogram of the actual lighting area. The expected standard characteristics are the best color distribution patterns determined for different environments (such as foggy days, sunny days, dusk, etc.). For example, foggy environments are expected to have warmer tones and the proportions of each color have a corresponding ideal range. Then, appropriate mathematical methods are used, such as calculating distance metrics in color space (such as Euclidean distance, etc.) to quantify the difference between the actual color distribution and the expected standard color distribution. The difference value is then compared with the set threshold. If it is greater than the threshold, it means that the color distribution does not meet expectations and deviates from the color state required for the ideal lighting effect.

[0106] For the clarity of the illuminated area, the actual key feature is a quantitative indicator that can reflect the clarity of the edge of an object, such as the edge gradient value, calculated by using image sharpening algorithms, edge detection algorithms, etc. The expected standard feature is based on the standard clarity values ​​set for safe driving requirements and different scenarios. The actual clarity index is calculated by subtracting the expected standard clarity index from the actual clarity index, and the absolute value is also taken to obtain the difference value. When this difference value is greater than the corresponding threshold value set, it indicates that the clarity of the illuminated area does not meet the required standard, which may affect the driver's accurate recognition of road objects.

[0107] By calculating the difference values ​​and comparing the thresholds of the three key image features, namely light intensity, color distribution, and clarity of the illuminated area, with their respective expected standard features in the above manner, we can make an overall judgment on whether the current lighting conditions meet the requirements, providing a key basis for the subsequent accurate provision of lighting evaluation results.

[0108] When the difference between the key features of the image and the expected standard features is greater than the set threshold, it indicates that the current lighting conditions deviate significantly from the ideal state. This means that the light intensity may be too high or too low, affecting the driver's visual effect and judgment of the road conditions; the color distribution may not be suitable for the current environment, resulting in insufficient light penetration or reduced visual clarity; the clarity of the illuminated area may be poor, making the edges of objects blurred and difficult to accurately identify objects on the road. This large deviation will reduce driving safety and comfort, so the lighting evaluation result is determined to be unqualified.

[0109] Only when the difference value is within the set threshold range, the lighting condition is considered qualified, which can provide the driver with good lighting conditions and ensure driving safety.

[0110] If one or more of the above difference values ​​are greater than the set threshold, the lighting evaluation result is determined to be unqualified, and then the process proceeds to step 309 .

[0111] If the difference value is less than or equal to the set threshold, the lighting evaluation result is determined to be qualified, and the system will continue to maintain the current lighting condition of the vehicle headlights.

[0112] Step 309: Determine the light fine adjustment information according to the difference between the key features of the image and the expected standard features.

[0113] This step can refer to the description in the above embodiment and will not be described again here.

[0114] Step 310: Finely adjust the light brightness and color temperature of the headlights according to the light fine adjustment information.

[0115] This step can refer to the description in the above embodiment and will not be described again here.

[0116] In some embodiments, the method of this embodiment may also include: obtaining a facial image of the driver of the vehicle during driving of the vehicle; then determining whether the driver of the vehicle is in a fatigue state based on the driver's facial image; if so, controlling the headlights to flash at a set frequency to remind the driver.

[0117] Specifically, when the vehicle is driving, cameras are usually installed at appropriate locations in the vehicle, such as above the dashboard or near the rearview mirror. These cameras are carefully debugged and calibrated to ensure that the driver's facial image can be accurately captured. The camera will continue to shoot at a certain frame rate to obtain the driver's facial state at different driving moments in real time. In order to improve image quality and reduce interference, some image enhancement technologies may be used, such as automatically adjusting exposure and contrast, so that the driver's facial features are clearer and more recognizable. Through high-speed data transmission lines, the captured driver's facial image is transmitted to the vehicle's control system in real time for subsequent analysis and processing.

[0118] After acquiring the driver's facial image, the system uses image processing and machine learning algorithms to determine whether the driver is in a state of fatigue. First, the driver's facial image is preprocessed, including face detection and alignment, to determine the key areas of the driver's face, such as the eyes, mouth, forehead, etc. Then, the characteristics of these key areas are analyzed. For the eyes, the driver's fatigue level is judged by detecting indicators such as the degree of eye opening and closing, blinking frequency, and eye movement. For example, if the eyes are half-closed for a long time or the blinking frequency is significantly reduced, it may indicate that the driver is in a state of fatigue. For the mouth, observe whether there are signs of fatigue such as yawning. At the same time, other features such as facial expressions and head posture can also be combined for comprehensive judgment. The algorithm is optimized through a large amount of training data so that it can accurately identify different degrees of fatigue.

[0119] If the driver is judged to be in a state of fatigue based on the driver's facial image, the vehicle's control system will issue a command to control the headlights to flash at a set frequency. This frequency is usually determined through experiments and research. It should be able to attract the driver's attention, but not too frequent to cause interference. The control system will achieve the flashing function by adjusting the current of the headlights or the control circuit. For example, the headlight power can be turned on and off quickly within a certain time interval to make the light flash. The flashing light can serve as a strong visual stimulus to remind the driver to pay attention to his fatigue state and take measures such as rest in time to ensure driving safety. At the same time, the system can also combine other methods such as sound prompts to further enhance the reminder effect.

[0120] In some embodiments, the method of this embodiment may also include: during the driving of the vehicle, obtaining the speed and distance of the oncoming vehicle through the vehicle networking system; then determining the low beam switching location based on the speed and distance of the oncoming vehicle and the vehicle's own driving speed; when the vehicle drives to the low beam switching location, switching the vehicle's headlights from high beam to low beam.

[0121] Among them, the Internet of Vehicles is an intelligent network system that interconnects vehicles with vehicles, vehicles with people, vehicles with road infrastructure, etc. through in-vehicle networks, inter-vehicle networks, and in-vehicle mobile Internet to achieve information interaction, collaborative control, and improve traffic efficiency and driving experience.

[0122] In this embodiment, the vehicle system can obtain information about oncoming vehicles in real time through its own sensors, such as radar, camera, etc., as well as communication modules with surrounding vehicles. The radar can emit electromagnetic waves and receive reflected signals, and calculate the distance of the oncoming vehicle based on the round-trip time and intensity of the signal. The camera can use image recognition technology to analyze the position and size changes of the oncoming vehicle in the picture, and estimate the distance in combination with known camera parameters. At the same time, through the communication protocol between vehicles, the speed information sent by the oncoming vehicle is received, or the displacement change of the oncoming vehicle per unit time is analyzed by image analysis to infer its speed. These data will be transmitted to the control system of this vehicle in real time through the Internet of Vehicles system, providing accurate basic data for subsequent processing.

[0123] After obtaining the speed and distance of the oncoming vehicle and the driving speed of the vehicle, the system will use a specific algorithm to determine the location where the low beam lights are switched. First, consider the speed and distance of the oncoming vehicle and calculate the time and distance change rate of the two vehicles being relatively close. If the oncoming vehicle is faster and closer, the time of relative closeness will be shorter. At the same time, combined with the driving speed of the vehicle, determine the time required to switch from the current position to the position where the low beam lights need to be switched at the current speed. According to the principles of safe driving and relevant regulations, determine a suitable switching time, usually when the two vehicles are a certain distance apart and the relative approach speed is fast, then it is necessary to determine the switching location in advance. The determination of this location requires comprehensive consideration of multiple factors, such as road conditions, traffic flow, line of sight, etc., to ensure that after switching the low beam lights, the driver's field of vision can be guaranteed without causing glare interference to the oncoming vehicles.

[0124] When the vehicle reaches the determined low beam switching location, the control system will issue a command to switch the headlights from high beam to low beam. This process is usually achieved by controlling the circuit switch of the headlights. The control system will detect the vehicle's location information according to the preset program, and immediately trigger the headlight switching action when it is determined that the vehicle has reached the switching location. The switching action must be quick and stable to ensure that the transition of the lights is natural and does not cause sudden changes in the driver's vision. At the same time, the system will continue to monitor the vehicle's driving status and surrounding environment. If the situation changes, such as the oncoming vehicle moving away or the road conditions changing, the headlight status may be adjusted again as needed to adapt to different driving scenarios.

[0125] In some embodiments, the method of this embodiment may further include the following steps: obtaining performance index data of the headlights; then inputting the performance index data of the headlights into a preset headlight fault prediction model to determine the failure probability of the headlights during the subsequent driving process; if the failure probability is greater than the set probability threshold, an alarm is issued to the driver of the vehicle.

[0126] Specifically, the system obtains the performance index data of the headlights through the sensors installed in the headlight module and the vehicle's own electronic control unit (ECU). These sensors can monitor the performance index data of the headlights, such as current, voltage, power, brightness, color temperature, etc., and can also detect information such as the working temperature and bulb life of the headlights. Through the vehicle's internal communication network, these performance index data are transmitted to the control system in real time for subsequent analysis and processing. For example, the current sensor can measure the current size of the headlight when it is working, the voltage sensor can monitor the stability of the power supply voltage, and the brightness sensor can measure the light intensity emitted by the headlight. These data provide important basis for evaluating the performance status of the headlight.

[0127] After the performance index data of the headlights are obtained, these data will be input into the headlight fault prediction model. The headlight fault prediction model will analyze and process these data, extract key features, and compare and pattern match with historical fault data. Through comprehensive evaluation of different performance indicators, the headlight fault prediction model can predict the probability of headlight failure during the next driving process. For example, if the current of the headlight fluctuates abnormally, the brightness drops significantly, or the operating temperature is too high, the model may determine that the headlight has a high risk of failure. The model will continue to learn and update, and with the input of new data, the accuracy of the prediction will continue to improve.

[0128] When the failure probability determined by the headlight failure prediction model is greater than the set probability threshold, the vehicle's control system will issue an alarm to the driver. There can be many ways to alarm, such as flashing warning lights on the dashboard, sound prompts, text messages on the display screen, etc. This can promptly remind the driver to pay attention to potential headlight failures and take corresponding measures, such as checking and repairing the headlights as soon as possible or adjusting the driving style to ensure driving safety. For example, if it is predicted that the headlights may go out in a short time, the driver can slow down in advance and turn on the hazard warning lights to avoid dangerous situations caused by sudden loss of lighting. At the same time, the system can also transmit fault information to the vehicle's maintenance service center so that maintenance personnel can be arranged for inspection and repair in a timely manner.

[0129] In this embodiment, the training process of the headlight fault prediction model is as follows: First, a large amount of historical data on vehicle headlights is collected, including performance indicator data of headlights of vehicles of different brands and models under various usage conditions, such as current, voltage, power, brightness, color temperature, operating temperature, bulb life, etc. At the same time, information such as whether the corresponding headlights have faults and the time and type of faults are recorded.

[0130] The data are preprocessed, including data cleaning, removal of outliers and erroneous data, and data normalization, which unifies data of different dimensions into the same numerical range so that the model can learn better.

[0131] Select appropriate machine learning algorithms, such as neural networks, support vector machines, random forests, etc., to build the architecture of the headlight failure prediction model.

[0132] The preprocessed historical data is divided into training set, validation set and test set.

[0133] The model is trained using training set data and the model parameters are continuously adjusted so that the model can learn the potential relationship between headlight performance indicators and faults.

[0134] During the training process, the validation set data is used to monitor the performance of the model, prevent overfitting, and adjust the model's hyperparameters based on the performance on the validation set.

[0135] After the model training is completed, the test set data is used to perform a final evaluation of the model to test the model's predictive ability on unseen data.

[0136] If the performance of the model does not meet expectations, analyze the reasons and make adjustments, such as increasing the amount of data, improving the algorithm, adjusting hyperparameters, etc., and then retrain and evaluate until a high-performance model is obtained that can accurately predict the probability of headlight failure during the subsequent driving process.

[0137] In some embodiments, the lighting control system includes an electronic device. The electronic device in the embodiment of the present invention is described below from the perspective of hardware processing. Figure 3 , is a schematic diagram of a physical device structure of an electronic device in an embodiment of the present invention.

[0138] It should be noted that Figure 3 The structure of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0139] like Figure 3 As shown, the electronic device includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 to the random access memory (RAM) 403, such as executing the method described in the above embodiment. In RAM 403, various programs and data required for system operation are also stored. CPU 401, ROM 402 and RAM 403 are connected to each other through bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0140] The following components are connected to the I / O interface 405: an input section 406 including an audio input device, a button switch, etc.; an output section 407 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed so that a computer program read therefrom is installed into the storage section 408 as needed.

[0141] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 409, and / or installed from a removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, various functions defined in the present invention are performed.

[0142] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CDROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium that contains or stores a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0143] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0144] Specifically, the electronic device of this embodiment includes a processor and a memory, the memory is coupled to one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the electronic device to execute the method provided by the above embodiment.

[0145] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the electronic device, the electronic device implements the method provided in the above embodiment.

[0146] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

[0147] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0148] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. An adaptive lighting sensing adjustment method, characterized in that: include: Obtaining environmental conditions of the vehicle driving area through a cloud server, the environmental conditions including temperature information, humidity information, haze conditions, rain conditions, and sunlight conditions; Inputting the environmental conditions into a preset light sensing adjustment model to obtain light pre-adjustment information; pre-adjusting the light brightness and color temperature of the headlights of the vehicle according to the light pre-adjustment information; Acquiring a pre-adjusted light illumination image in front of the vehicle when the light brightness and color temperature have been pre-adjusted; Inputting the pre-adjusted lighting image into a preset lighting condition evaluation model to obtain a lighting condition evaluation result; If the lighting evaluation result is unqualified, the light brightness and color temperature of the headlight are finely adjusted according to the pre-adjusted lighting image.

2. The method according to claim 1, characterized in that The step of inputting the pre-adjusted lighting image into a preset lighting condition evaluation model to obtain a lighting condition evaluation result specifically includes: Determining the lighting illumination area in the pre-adjusted lighting illumination image by using the lighting illumination condition evaluation model; Determine the key image features of the lighting area through the lighting condition evaluation model, wherein the key image features include light intensity, color distribution and lighting area clarity; When the difference between the key feature of the image and the expected standard feature is greater than a set threshold, the lighting evaluation result is determined to be unqualified.

3. The method according to claim 2, characterized in that The finely adjusting the light brightness and color temperature of the headlight according to the pre-adjusted light illumination image comprises: Determining lighting fine adjustment information according to the difference between the key features of the image and the expected standard features; The light brightness and color temperature of the headlight are finely adjusted according to the light fine adjustment information.

4. The method according to claim 2, characterized in that: When the difference between the key feature of the image and the expected standard feature is greater than a set threshold, before determining that the lighting evaluation result is unqualified, the method further includes: The expected standard feature is determined from a preset standard feature database according to the environmental condition, and the standard feature database stores corresponding relationships between different environmental conditions and different standard features.

5. The method according to claim 1, characterized in that The method further comprises: Acquiring a facial image of a driver of the vehicle during driving of the vehicle; determining whether the driver of the vehicle is in a fatigue state according to the driver's facial image; If so, the headlights are controlled to flash at a set frequency to remind the driver.

6. The method according to claim 1, characterized in that The method further comprises: During the driving of the vehicle, the speed and distance of the oncoming vehicle are obtained through the vehicle networking system; Determining a low beam switch location according to the speed and distance of the oncoming vehicle and the vehicle's own speed; When the vehicle travels to the low beam switch location, the headlights of the vehicle are switched from high beam to low beam.

7. The method according to claim 1, characterized in that The method further comprises: Obtaining performance index data of the headlight; Inputting the performance index data of the headlight into a preset headlight failure prediction model to determine the failure probability of the headlight in the next driving process; If the failure probability is greater than a set probability threshold, an alarm is issued to the driver of the vehicle.

8. An electronic device, characterized in that: including one or more processors and memory; The memory is coupled to the one or more processors, and the memory is used to store computer program codes, wherein the computer program codes include computer instructions, and the one or more processors call the computer instructions to enable the electronic device to execute the method according to any one of claims 1 to 7.

9. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed on an electronic device, the electronic device is caused to execute the method as claimed in any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is executed on an electronic device, the electronic device is enabled to execute the method according to any one of claims 1 to 7.

Images