A method and system for adjusting color temperature of automobile headlamp in combination with weather identification

Abstract

The application relates to the technical field of automobile electronic control, and discloses a high-beam color temperature adjustment method and system combined with weather recognition, aiming to solve the problem that color temperature adjustment in the prior art only depends on light intensity and lacks accurate recognition and differentiated response to weather such as rain, fog and snow. The method comprises the following steps: fusing front-view camera, millimeter wave radar and ambient light sensor data, constructing a multi-modal weather discrimination model to identify the current weather type; calling a color temperature mapping table according to the weather type and combining light intensity dynamic compensation to determine a target color temperature value; and performing closed-loop correction on driver visual feedback based on infrared eye movement tracking. The system comprises a multi-source perception module, a weather recognition module, a color temperature decision module, an LED driving module and a visual feedback module. Through the above scheme, high-precision weather adaptive color temperature adjustment is realized, and the lighting penetration, road recognition and driving visual comfort under low visibility are improved.

Description

Technical Field

[0001] This invention relates to the field of automotive electronic control technology, specifically to a method and system for adjusting the color temperature of automotive headlights in conjunction with weather recognition. Background Technology

[0002] With the continuous evolution of intelligent vehicle technology, in-vehicle lighting systems have gradually developed from single-function lighting devices into intelligent subsystems that integrate environmental perception and human-machine interaction. Their performance not only affects nighttime driving safety, but is also increasingly becoming an important component for improving driving comfort and user experience.

[0003] Among them, the color temperature adjustment technology of automotive headlights, as one of the key directions of intelligent lighting systems, aims to dynamically adjust the color temperature of the light source according to external environmental conditions to optimize driver visual clarity and reduce visual fatigue. In existing technologies, some high-end models have introduced automatic dimming mechanisms based on ambient light intensity, but their adjustment logic mainly relies on light sensor data and lacks the ability to comprehensively identify and respond to complex weather conditions.

[0004] Current technologies for automotive headlight color temperature adjustment still have significant shortcomings: First, the color temperature adjustment strategy is overly simplistic, relying solely on linear adjustments based on illuminance without considering the varying impacts of different weather conditions such as rain, fog, and snow on light scattering and penetration. Second, it lacks the ability to actively recognize weather conditions, failing to effectively integrate meteorological information into lighting decisions. This leads to inappropriate color temperature selection in low-visibility conditions, exacerbating glare or reducing road visibility. Third, existing systems often use fixed thresholds or preset mode switching, resulting in poor adaptability and difficulty in handling rapidly changing weather scenarios. Finally, there is a lack of coordination between color temperature adjustment and other vehicle perception systems (such as cameras and radar), failing to construct a multi-source environmental perception closed loop centered on safe driving. These problems severely restrict the practicality and safety of current headlight color temperature adjustment technology in real and complex weather environments, necessitating an intelligent lighting solution that deeply integrates weather recognition and dynamic color temperature control. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for adjusting the color temperature of automotive headlights by incorporating weather recognition, which can effectively solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method and system for adjusting the color temperature of automotive headlights in conjunction with weather recognition includes the following specific steps: Step 1: Acquiring multi-source environmental perception data: Simultaneously acquiring images of the road ahead, meteorological echo signals, and light intensity data through an onboard forward-facing camera, millimeter-wave radar, and ambient light sensor; aligning the data with timestamps and unifying spatial coordinates to form a fused perception dataset; Step 2: Identifying the current weather type: Based on the fused perception dataset, extracting texture and contrast features from the road images using a convolutional neural network; combining the attenuation rate and scattering characteristics of the millimeter-wave radar echo signal to construct a multimodal weather discrimination model and outputting the current weather type, which includes sunny, rainy, foggy, snowy, and mixed weather scenarios; Step 3: Determining the target color temperature value: According to the weather type, calling a preset color temperature mapping table... The corresponding base color temperature value is obtained and dynamically compensated in conjunction with real-time light intensity. The color temperature for rainy and foggy days is within the first preset range, the color temperature for snowy days is within the second preset range, and the color temperature for clear nighttime days is within the third preset range. Step 4 generates a color temperature adjustment command: the target color temperature value is converted into LED driving current parameters, and the current ratio of warm white light and cool white light chips in the dual-color temperature LED light source is controlled by a pulse width modulation signal to achieve continuous stepless color temperature adjustment. Step 5 executes closed-loop feedback correction: after color temperature adjustment, the visual response signal of the driver's eye area is continuously monitored. The pupil contraction frequency and gaze stability index are obtained through an infrared eye tracking module. If visual fatigue or glare is detected, the color temperature value is finely adjusted and the adjustment command is updated until the visual comfort index stabilizes within the preset threshold range.

[0008] Preferably, in step 1, the forward-looking camera is a high-resolution global shutter CMOS image sensor with a predetermined frame rate and field of view; the millimeter-wave radar operates in a preset high-frequency band and has a predetermined detection distance; and the ambient light sensor's spectral response range covers the visible light band and has high sensitivity.

[0009] Preferably, in step 2, the convolutional neural network includes multiple convolutional layers, pooling layers, and fully connected layers. The input image has a predetermined size, and the training dataset contains a large number of real vehicle road images labeled with weather types. The attenuation rate calculation formula of the millimeter-wave radar echo signal is the exponential attenuation coefficient of the signal strength with distance. The scattering characteristics are quantified by the standard deviation of Doppler frequency shift. When the attenuation rate is greater than the first preset threshold and the scattering standard deviation is less than the second preset threshold, it is determined to be dense fog weather.

[0010] Preferably, in step 3, the color temperature mapping table is constructed based on the photopic and scotopic spectral sensitivity curves of the human eye under different weather conditions. A low color temperature is used on rainy days to enhance the penetration of red and yellow light; a medium-low color temperature is used on foggy days to reduce Rayleigh scattering of blue light; a medium-high color temperature is used on snowy days to improve the contrast between white road surfaces and snow accumulation; and a high color temperature is used on clear nights to maintain the visual experience of natural light. The dynamic compensation formula is as follows: ,in For reference light intensity, To measure the actual light intensity, This is a predetermined proportional coefficient.

[0011] Preferably, in step 4, the dual-color temperature LED light source is composed of a low color temperature warm white LED chip and a high color temperature cool white LED chip connected in parallel. The driving circuit adopts a constant current source architecture, the pulse width modulation frequency is a preset working frequency, the current adjustment has high resolution, and the color temperature adjustment accuracy reaches a predetermined tolerance range.

[0012] Preferably, in step 5, the infrared eye-tracking module has a predetermined sampling frequency, the pupil contraction frequency threshold is set as a preset physiological index, and the gaze stability index is calculated by the standard deviation of the gaze point offset. When the standard deviation of the offset is greater than the preset angle threshold and the duration exceeds the predetermined time period, it is determined to be visual discomfort, triggering color temperature fine-tuning. The fine-tuning step size is the predetermined color temperature increment, and the maximum number of adjustments is the preset upper limit.

[0013] Preferably, the multimodal weather discrimination model supports an online learning mechanism. When the vehicle enters a new meteorological area and the weather recognition results are inconsistent with the user's manual mode switching, the current environmental data is automatically collected and uploaded to the cloud training platform to update the local model weights. The model update cycle does not exceed a predetermined time interval.

[0014] Preferably, the system also includes a data interaction interface with the vehicle's ADAS domain controller. When foggy or rainy weather is detected, the system automatically activates the fog lights and reduces the illumination angle of the adaptive high beams, while sending a low visibility warning message to the instrument panel, thereby achieving coordinated linkage between the lighting system and the active safety system.

[0015] Preferably, the color temperature adjustment process is equipped with a safety lock mechanism. When the vehicle speed is greater than the preset speed threshold and the weather type is foggy, the setting of the color temperature higher than the preset upper limit value is prohibited to prevent the high color temperature light from producing strong backscattering in dense fog, which may cause the driver to be momentarily blinded.

[0016] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0017] 1. Precise meteorological perception and response capabilities

[0018] Breaking the limitations of reliance on single illumination: By fusing camera image texture, millimeter-wave radar scattering characteristics, and ambient light intensity, a multimodal weather recognition model is constructed. The recognition accuracy for low-visibility weather such as rain, fog, and snow is significantly better than existing systems that rely solely on light sensors. Differentiated color temperature strategy: Based on the differences in the physical characteristics of light propagation under different weather conditions, a scientific color temperature mapping rule is established. Low color temperatures within a first preset range are used to improve penetration in rainy and foggy weather, while medium to high color temperatures within a second preset range are used to enhance contrast in snowy weather, avoiding glare or reduced recognition caused by traditional linear adjustment.

[0019] 2. Dynamic closed-loop regulation and human factors optimization

[0020] Visual comfort closed-loop control: Introducing infrared eye-tracking technology to monitor the driver's physiological reactions in real time, achieving an upgrade of color temperature adjustment from "environmental adaptation" to "human-cause adaptation", significantly reducing the incidence of visual fatigue; High-precision stepless adjustment: Using dual-color temperature LEDs and high-resolution current control, the color temperature adjustment resolution reaches the predetermined accuracy, and the response time is less than the preset response time limit, meeting the real-time needs in rapidly changing weather scenarios.

[0021] 3. System-level security collaboration and intelligent evolution

[0022] Multi-system safety linkage: Deeply integrated with ADAS domain controllers, it automatically coordinates fog lights, high beams, and human-machine interaction systems in severe weather to build a lighting decision-making closed loop with safety as the core; Continuous learning capability: Supports online model updates based on user behavior and environmental data, enabling the system to continuously optimize recognition accuracy and adjustment strategies during long-term use and adapt to regional meteorological differences. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall process of the adjustment method and system proposed in this invention;

[0024] Figure 2 This is a flowchart illustrating the multimodal weather discrimination model in this invention;

[0025] Figure 3 This is a schematic diagram of the target color temperature determination and dynamic compensation logic flow in this invention;

[0026] Figure 4 This is a schematic diagram of the multi-level interaction relationship and data flow between visual comfort closed-loop feedback correction and system safety coordination in this invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Example 1

[0029] This embodiment applies to passenger vehicles equipped with Level 2 or higher intelligent driving assistance systems. Its core objective is to achieve high-precision, adaptive, and safe closed-loop adjustment of the color temperature of automotive headlights under complex and variable weather conditions by integrating multi-source perception data and human factor feedback mechanisms. This system must not only meet optical performance requirements but also achieve automotive-grade standards in terms of functional safety, real-time performance, and human-machine collaboration.

[0030] First, a complete hardware and software collaborative platform is constructed at the system architecture level. The system deployed in this embodiment includes the following core functional modules: a multi-source environmental perception module, a multimodal weather discrimination module, a color temperature decision and dynamic compensation module, an LED drive control module, a visual comfort feedback module, a safety collaborative execution module, and a cloud-based online learning interface. Each module forms a hybrid communication network via in-vehicle Ethernet (100BASE-T1) and CAN FD bus, ensuring parallel transmission of high-bandwidth data streams and low-latency control commands.

[0031] The multi-source environmental perception module consists of three physical sensor units: the first is a forward-looking camera, which uses a 1.2-megapixel global shutter CMOS image sensor (such as the Sony IMX390), has a 120° horizontal field of view (HFOV), a stable frame rate of 30 frames per second, and supports HDR mode, used to capture the texture, contrast, and visibility information of the road scene ahead; the second is a 77GHz millimeter-wave radar, with a detection range of up to 150 meters, and 4D imaging capabilities (distance, speed, azimuth, and elevation). Its antenna array outputs raw I / Q signals, which are processed internally by FFT to generate point clouds and echo intensity maps, used to quantify the attenuation and scattering characteristics of electromagnetic waves by raindrops, fog droplets, or snow crystals; the third is an ambient light sensor, which uses a silicon photodiode array, with a spectral response range covering the visible light band from 380nm to 780nm, a sensitivity of 0.1 lux (lx), and a dynamic range of 0.1 lx to 100,000 lx, used to monitor ambient light intensity in real time. All three types of sensors are timestamped using a hard synchronization signal (a 10MHz clock emitted by the main control ECU) and spatially registered using the vehicle coordinate system (with the center of the front axle of the vehicle body as the origin) to form a fusion perception dataset under a unified spatiotemporal reference.

[0032] The multimodal weather discrimination module is implemented by a dedicated AI acceleration unit (such as the DSP+NPU coprocessor integrated in the NXP S32G2) within the vehicle domain controller. This module runs a lightweight convolutional neural network (CNN) with a network structure comprising five convolutional layers (kernel sizes of 7×7, 5×5, 3×3, 3×3, and 3×3 respectively), two max-pooling layers (stride 2, window 2×2), and one fully connected output layer. The input image is preprocessed and cropped to 224×224 pixel RGB format. During the training phase, over 100,000 real-vehicle collected and manually annotated road images are used, covering nine scenarios including sunny, light rain, moderate rain, heavy rain, light fog, dense fog, light snow, heavy snow, and mixed weather (such as sleet, fog + haze). Simultaneously, millimeter-wave radar echo signals are used to extract two key physical features: one is the signal attenuation rate α, defined as the coefficient by which the echo intensity decreases exponentially with distance, calculated using the formula... ,in Distance The echo intensity at the location; and the scattering characteristic indicators. , or the standard deviation of the Doppler frequency shift, reflects the degree of disorder in particle motion. When > 0.8 and When the value is less than 0.2, the system identifies it as dense fog. The image classification probability output by the CNN and the radar feature vector are combined using a weighted fusion strategy (such as Bayesian fusion or attention mechanism) to generate the final weather type label, achieving an accuracy of 96.3% in real vehicle testing.

[0033] The color temperature decision and dynamic compensation module runs in the application layer software of the same domain controller. Its implementation logic is based on a preset color temperature mapping table and a dynamic compensation algorithm. This mapping table is constructed based on the CIE 1931 standard colorimetric system and the photopic and scotopic spectral sensitivity curves of the human eye: In rainy and foggy weather, because water droplets or particles cause stronger Rayleigh scattering of short-wavelength blue light, a low color temperature of 2700K–3500K is used to enhance the red and yellow light components and improve penetration; in snowy weather, because white snow has high reflectivity, a medium-high color temperature of 4000K–4500K is required to enhance the contrast of road textures and obstacles; on clear nights, a high color temperature of 5000K–6000K is used to simulate sunlight to maintain a natural visual experience. Based on this, a dynamic compensation mechanism is introduced, with the compensation formula as follows:

[0034]

[0035] in, The base color temperature value is obtained from the mapping table. , For reference light intensity, This is the measured value from the ambient light sensor. When... When the environment is dark, the color temperature is appropriately lowered to reduce glare; conversely, it is slightly increased, but this is subject to a safety locking mechanism.

[0036] The LED driver control module consists of a dual-channel constant current drive circuit, directly integrated into the headlight module. This module drives a pair of parallel dual-color temperature LED chips: a warm white LED with a color temperature of 2700K (typical CCT=2700±100K, CRI>80) and a cool white LED with a color temperature of 6500K (CCT=6500±100K, CRI>75). The two drive currents are independently controllable, employing a digital constant current source architecture with a current adjustment range of 0–350mA and a resolution of 1mA. The pulse width modulation (PWM) signal is generated by the main control MCU (such as the Infineon AURIX TC397) at a fixed frequency of 1000Hz to avoid flicker perceptible to the human eye. The target color temperature value is converted into the current ratio of the two LEDs using a lookup table method or interpolation algorithm. For example, a target color temperature of 3500K corresponds to , , The entire adjustment process has a response time of less than 0.5 seconds, and the color temperature control accuracy reaches ±50K.

[0037] The visual comfort feedback module is implemented by an embedded infrared eye-tracking unit, installed above the dashboard or in the rearview mirror base, and includes an 850nm infrared LED light source and a high-sensitivity near-infrared CMOS sensor (640×480 resolution, 60Hz frame rate). This module continuously tracks the center position of the driver's pupils, changes in pupil diameter, and the trajectory of the gaze point. The system calculates two key physiological indicators: first, the pupil constriction frequency, defined as the number of times the pupil diameter changes beyond a threshold (e.g., 0.2mm) per unit time, with a normal value of less than 12 times / minute; second, gaze stability, assessed by calculating the standard deviation (in degrees) of the gaze point offset angle over 5 consecutive seconds. If this value is greater than 1.5° and persists for more than 5 seconds, it is considered visual discomfort (e.g., glare or difficulty in recognition). Once discomfort is triggered, the system initiates closed-loop correction: fine-tuning the target color temperature in 100K increments (e.g., reducing it from 3500K to 3400K), regenerating the drive command, and monitoring whether the indicators return to the threshold within the next 5 seconds. A maximum of three fine-tuning attempts are allowed. If the result is still not satisfactory, the current optimal color temperature will be locked and the user will be prompted to intervene manually.

[0038] The safety collaborative execution module establishes bidirectional communication with the ADAS domain controller via the CAN FD bus. When the weather judgment module outputs a "foggy" or "heavy rain" label, the system automatically sends control commands: activates the front and rear fog lights (if not already on), lowers the illumination angle of the adaptive high beam (ADB) by 5°–10° to reduce glare from ground reflections, and sends a "Low visibility, drive with caution" warning text to the instrument panel HMI via CAN message. Furthermore, the system has built-in safety locking logic: when the vehicle speed signal (from the wheel speed sensor) is greater than 80 km / h and the weather type is "foggy," it prohibits the output of color temperature commands higher than 4000K. Even if the dynamic compensation algorithm calculates a higher value, it forcibly clamps the color temperature to 4000K to prevent strong backscattering of high color temperature blue light in dense fog, which could cause momentary blindness.

[0039] With the above system architecture as support, the workflow of this embodiment is as follows:

[0040] First, after the vehicle starts, the multi-source environmental perception module begins to collect data synchronously. The forward-facing camera outputs a 1.2-megapixel image every 33.3ms, the millimeter-wave radar outputs a point cloud and echo intensity map every 20ms, and the ambient light sensor updates the illumination intensity value at a 10ms cycle. All data are aligned with timestamps (error <1ms) and transformed in spatial coordinates (using the pitch and roll angles provided by the vehicle's IMU for image-radar registration) to form a fused perception dataset, which is then sent to the multimodal weather discrimination module.

[0041] Subsequently, the multimodal weather discrimination module processes image and radar data in parallel. The CNN extracts features from the images and outputs the probability distributions for various weather conditions (e.g., fog 0.85, rain 0.10, sunny 0.05). Simultaneously, the radar signal processing unit calculates the attenuation rate α and the scattering standard deviation σ_d. If α = 0.85 and σ_d = 0.15, the radar strongly supports the "dense fog" judgment. After fusion, the system finally outputs the "dense fog" weather type.

[0042] Next, the color temperature decision module looks up the base color temperature based on the "dense fog" label. , If the ambient light sensor reading at this time... , Substituting into the dynamic compensation formula, we get:

[0043]

[0044] The value is within the permissible range of 2700K–3500K, and the current vehicle speed is 60 km / h (lock not triggered), so it is legal.

[0045] Then, the LED driver control module converts 3450K into current parameters. Using a pre-calibrated color temperature-current mapping curve, the warm white light current is determined to be 260mA, and the cool white light current to be 140mA. The MCU generates two 1000Hz PWM signals to control two constant current sources respectively, achieving stepless smooth transition.

[0046] After adjustment, the visual comfort feedback module begins monitoring. If the driver's pupil constriction frequency reaches 15 times / minute within 5 seconds, and the standard deviation of the fixation point deviation is 1.8°, the system determines it to be slight glare. The first fine-tuning is then initiated: the color temperature is lowered to 3350K, and the current is redistributed (270mA for warm light, 130mA for cool light). If, after 2 seconds, the indicators improve to a pupil frequency of 10 times / minute and a standard deviation of 1.2°, the correction is successful, and the closed loop ends.

[0047] Simultaneously, the safety coordination module sends a "fog" event to the ADAS domain controller, which then activates the rear fog lights, lowers the high beam angle, and displays a warning on the instrument panel. The entire process, from perception to execution, takes approximately 400ms, meeting the ISO26262 ASIL-B functional safety requirements.

[0048] Furthermore, the system supports an online learning mechanism. If a vehicle enters a high-altitude area and misclassifies "dust + fog" as "sunny weather" five times consecutively, and the user manually switches to "fog weather" mode, the system automatically captures image, radar, and lighting data from the current 5 seconds, tags it with "user correction," and uploads it to the cloud training platform via a 4G / 5G module with encryption. After incrementally training the new model in the cloud, an OTA update package is pushed out within 24 hours, and the local NPU loads the new weights to improve the robustness of subsequent recognition.

[0049] In summary, this embodiment achieves intelligent, safe, and comfortable adjustment of automotive headlight color temperature under complex weather conditions through precise hardware selection, rigorous multimodal fusion algorithms, human-factor closed-loop feedback, and system-level safety collaboration. This is significantly superior to existing single-modal adjustment schemes that rely solely on light intensity.

[0050] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for adjusting the color temperature of automotive headlights by incorporating weather recognition, characterized in that, Includes the following steps: Multi-source environmental perception data is acquired by simultaneously collecting images of the road ahead, meteorological echo signals, and light intensity data through a vehicle-mounted forward-looking camera, millimeter-wave radar, and ambient light sensor. The data is then time-stamped and spatially coordinated to form a fused perception dataset. To identify the current weather type, based on the fused sensing dataset, a convolutional neural network is used to extract texture and contrast features from road images. Combined with the attenuation rate and scattering characteristics of millimeter-wave radar echo signals, a multimodal weather discrimination model is constructed to output the current weather type, which includes sunny, rainy, foggy, snowy, and mixed weather scenarios. Determine the target color temperature value, retrieve the corresponding base color temperature value by calling the preset color temperature mapping table according to the weather type, and perform dynamic compensation in combination with real-time light intensity; A color temperature adjustment command is generated, the target color temperature value is converted into LED driving current parameters, and the current ratio of warm white light and cool white light chips in the dual-color temperature LED light source is controlled by a pulse width modulation signal to achieve continuous stepless color temperature adjustment. The system performs closed-loop feedback correction and continuously monitors the visual response signal of the driver's eye area after color temperature adjustment. It obtains pupil contraction frequency and gaze stability index through infrared eye tracking module. If visual fatigue or glare reaction is detected, the color temperature value is finely adjusted and the adjustment command is updated until the visual comfort index is stable within the preset threshold range.

2. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, The forward-facing camera is a high-resolution global shutter CMOS image sensor with a predetermined frame rate and field of view. The millimeter-wave radar operates in the 77GHz band and has a predetermined detection distance. The ambient light sensor's spectral response range covers the visible light band and has high sensitivity.

3. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, The convolutional neural network includes multiple convolutional layers, pooling layers, and fully connected layers. The input image has a predetermined size, and the training dataset contains a large number of real vehicle road images labeled with weather types. The attenuation rate of the millimeter-wave radar echo signal is calculated by the exponential attenuation coefficient of the signal strength with distance, and the scattering characteristics are quantified by the standard deviation of Doppler frequency shift. When the attenuation rate is greater than a first preset threshold and the scattering standard deviation is less than a second preset threshold, it is determined to be dense fog weather.

4. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, The color temperature mapping table is constructed based on the photopic and scotopic spectral sensitivity curves of the human eye under different weather conditions. The color temperature for rainy and foggy days is in the range of 2700K to 3500K, the color temperature for snowy days is in the range of 4000K to 4500K, and the color temperature for clear nighttime days is in the range of 5000K to 6000K. Dynamic compensation is adjusted by the deviation ratio between the base color temperature value and the measured light intensity.

5. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, The dual-color-temperature LED light source is composed of a low color-temperature warm white LED chip and a high color-temperature cool white LED chip connected in parallel. The driving circuit adopts a constant current source architecture, with a pulse width modulation frequency of 1000Hz, current adjustment with 1mA resolution, and color temperature adjustment accuracy of ±50K.

6. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, The infrared eye-tracking module has a sampling frequency of 60Hz, a pupil contraction frequency threshold set to 12 times / minute, and a gaze stability index calculated by the standard deviation of the gaze point offset angle. When the offset standard deviation is greater than 1.5 degrees and the duration exceeds 5 seconds, it is determined to be visual discomfort, triggering color temperature fine-tuning. The fine-tuning step size is 100K, and the maximum number of adjustments is 3.

7. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, The multimodal weather discrimination model supports an online learning mechanism. When a vehicle enters a new meteorological area and the weather recognition results are inconsistent with the user's manual mode switching, the model automatically collects the current environmental data and uploads it to the cloud training platform to update the local model weights. The model update cycle does not exceed 24 hours.

8. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, It also includes data interaction steps with the vehicle's ADAS domain controller. When foggy or rainy weather is detected, the fog lights are automatically activated and the illumination angle of the adaptive high beams is reduced. At the same time, a low visibility warning message is sent to the instrument panel, realizing the coordinated linkage between the lighting system and the active safety system.

9. The method for adjusting the color temperature of automotive headlights in conjunction with weather recognition according to claim 1, characterized in that, The color temperature adjustment process is equipped with a safety lock mechanism. When the vehicle speed is greater than 80km / h and the weather is foggy, the color temperature setting is prohibited from being higher than 4000K to prevent high color temperature light from producing strong backscattering in dense fog, which could cause momentary blindness to the driver.

10. A car headlight color temperature adjustment system incorporating weather recognition, characterized in that, It includes a multi-source environment perception module, a multi-modal weather discrimination module, a color temperature decision and dynamic compensation module, an LED drive control module, a visual comfort feedback module, a safety collaborative execution module, and a cloud-based online learning interface; The multi-source environmental perception module is used to simultaneously acquire images of the road ahead, meteorological echo signals, and light intensity data to form a fused perception dataset; the multimodal weather discrimination module is used to identify the current weather type based on the fused perception dataset; the color temperature decision and dynamic compensation module is used to determine the target color temperature value according to the weather type; the LED drive control module is used to generate color temperature adjustment commands to control the dual-color temperature LED light source; the visual comfort feedback module is used to monitor the driver's visual response signal and perform closed-loop feedback correction; the safety collaborative execution module is used to interact with the ADAS domain controller to achieve safety linkage; and the cloud-based online learning interface is used to support the online updating of the multimodal weather discrimination model.

Images