Vehicle-mounted display screen image enhancement method and system

Abstract

The invention provides a vehicle-mounted display screen image enhancement method and system, and relates to the technical field of vehicle-mounted display, and the method comprises the steps: obtaining the illumination information of the surface of a display screen, dividing the illumination information into a plurality of illumination sub-regions, obtaining the real-time speed, the driving direction and the environment information of a vehicle, predicting the illumination change trend of each illumination sub-region, and carrying out the image enhancement of the vehicle-mounted display screen. Whether a driver wears polarized sunglasses is judged through mode recognition, the influence of the driver on a display screen image is pre-judged according to an optical model, and corresponding image enhancement parameters are selected and applied for all the illumination sub-areas according to the illumination information of the illumination sub-areas, the illumination change trend and the influence of the polarized sunglasses; according to the vehicle-mounted display screen, refined and personalized image enhancement is realized, so that the image visibility, the contrast ratio and the color accuracy of the vehicle-mounted display screen in various complex light environments are remarkably improved, and the problems that key information cannot be seen clearly due to glare, overexposure or shadow in the background technology, and color distortion and information black holes are caused under a polariscope are effectively solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle-mounted display, in particular to a vehicle-mounted display screen image enhancement method and system. BACKGROUND

[0002] As an important interface for human interaction in intelligent vehicles, the display effect of vehicle-mounted display screens is often greatly compromised in various complex light environments. For example, when the sun shines directly, there is not enough light at night, or when it rains or snows, the screen may have problems such as insufficient contrast, unclear details, color distortion, etc., which makes it difficult for the driver to quickly and accurately obtain key information. In order to solve these problems, the industry has been working hard to study how to improve the image display quality of vehicle-mounted display screens.

[0003] In the prior art, the vehicle-mounted display system usually relies on an ambient light sensor above the instrument panel to measure the overall light intensity in the cockpit, and according to this single light reading, the backlight brightness and picture contrast of the entire display screen are uniformly adjusted. This global uniform image adjustment strategy can still ensure the visibility of information in a uniform light change scenario (such as when the vehicle enters or exits a tunnel).

[0004] However, in actual driving, the complexity of the light environment far exceeds this simple processing method. For example, on a sunny afternoon, the vehicle is driving on urban roads, and the sun may shine into the vehicle from a certain angle, forming both a local highlight area and a deep shadow area on the surface of the central display screen. At this time, the light value collected by the ambient light sensor on the instrument panel is only a blurred average value of the light in the entire cockpit, and cannot accurately reflect the extreme non-uniformity of the light on the screen surface. The display system performs global uniform image adjustment based on this inaccurate "average information", resulting in the fact that the key information such as navigation and road conditions in the shadow area cannot be seen clearly due to insufficient brightness, while the safety alerts such as collision warning and blind area warning in the highlight area are completely ineffective due to glare and overexposure, so that the driver cannot obtain all the necessary information at the same time on the same screen.

[0005] More complex is that when the driver wears polarized sunglasses, due to the interaction between its optical properties and the polarization principle of the liquid crystal display screen, it will further exacerbate the above problems, causing unpredictable color distortion and local information black holes. For example, the blind area warning icon, which has already faded in the sunlight, may become an opaque black spot or show strange rainbow stripes under polarized glasses; the color and contrast of the navigation map in the shadow may also be distorted, making the already blurred route even more difficult to identify. The existing vehicle-mounted system cannot perceive or compensate for this visual deviation introduced by the external observation medium (sunglasses), and still performs invalid or even harmful global adjustment based on incorrect sensor data. SUMMARY

[0006] The application provides a vehicle-mounted display screen image enhancement method and system, aiming to solve the problems of poor image display effect, difficult information acquisition and the inability of the prior art to effectively perceive and compensate for these visual deviations caused by uneven illumination, the wearing of polarized sunglasses by the driver and other factors under complex lighting environments.

[0007] In one aspect, the application provides a vehicle-mounted display screen image enhancement method, comprising: obtaining illumination information of the display screen surface, and dividing the display screen into a plurality of illumination sub-regions according to the illumination information; obtaining the real-time speed, driving direction and environmental information of the vehicle, and predicting the illumination change trend of each illumination sub-region; judging whether the driver wears polarized sunglasses according to the pattern recognition of the optical characteristics of the eye region in the driver's face image, and predicting the influence of the polarized sunglasses on the display screen image according to a preset optical model of the interaction between the polarized sunglasses and the display screen; selecting and applying corresponding image enhancement parameters for each illumination sub-region according to the illumination information of the illumination sub-region, the illumination change trend and the influence of the polarized sunglasses on the display screen image.

[0008] Optionally, the step of selecting and applying corresponding image enhancement parameters for each illumination sub-region according to the illumination information of the illumination sub-region, the illumination change trend and the influence of the polarized sunglasses on the display screen image comprises: obtaining the physiological baseline data of the driver; after a small disturbance of the preset mode is performed on the image enhancement parameters of the illumination sub-region, collecting the instant physiological response data of the driver; comparing the instant physiological response data with the physiological baseline data to determine the influence direction of the disturbance of the image enhancement parameters on the visual comfort of the driver; iteratively adjusting the image enhancement parameters of the illumination sub-region according to the influence direction until the physiological indicators of the driver tend to be stable.

[0009] Optionally, the step of selecting and applying corresponding image enhancement parameters for each illumination sub-region according to the illumination information of the illumination sub-region, the illumination change trend and the influence of the polarized sunglasses on the display screen image comprises: identifying the information type displayed in the illumination sub-region, and performing priority sorting on the information displayed in the illumination sub-region according to the information type to obtain a priority sorting result; selecting the visibility image enhancement parameters of the information type with the highest priority from a preset parameter set according to the priority sorting result; Adjust the selected image enhancement parameters to optimize the visibility of other information types with lower priority, without significantly affecting the visibility of the information type with the highest priority, and apply the adjusted image enhancement parameters to the illumination sub-region.

[0010] Optionally, the step of obtaining the illumination information of the display screen surface and dividing the display screen into a plurality of illumination sub-regions according to the illumination information comprises: Obtaining image data of the display screen surface through an image sensor and pre-processing to remove transient highlight or dark spot interference information caused by dynamic reflection or local occlusion; Obtaining image data of the surrounding environment through an image sensor to obtain the illumination characteristics of the non-display screen area to assist in judging the overall ambient light type and intensity in the vehicle; According to the image data of the display screen surface and the illumination characteristics of the non-display screen area, separating the area of the display content itself from the area affected by the ambient light in the image data of the display screen surface; Analyzing the brightness, contrast and color distribution of the area affected by the ambient light to identify the local strong light direct radiation area and the deep shadow area; According to the boundaries and characteristics of the local strong light direct radiation area and the deep shadow area, the display screen is divided into a plurality of illumination sub-regions.

[0011] Optionally, the step of obtaining image data of the display screen surface through an image sensor and pre-processing comprises: Obtaining multiple frames of original images of the display screen surface with different exposure parameters through a high dynamic range image sensor, and performing exposure fusion processing on the multiple frames of original images of the display screen surface to generate a high dynamic range image; Performing local tone mapping processing on the high dynamic range image to compress the brightness range while preserving the details of the highlight and shadow areas, and identifying and separating the glare area and the reflection area of the display screen surface according to the image after local tone mapping processing; Performing image inpainting processing on the glare area and the reflection area to remove glare and reflection interference, and taking the image after image inpainting processing as the image data of the display screen surface.

[0012] Optionally, the step of obtaining the real-time speed, driving direction and environmental information of the vehicle to predict the illumination change trend of each illumination sub-region comprises: Obtaining the environmental information of the vehicle to identify whether there is a tunnel entrance, a tunnel exit, an area under a viaduct or a high-rise occlusion area in front of the vehicle, the environmental information including real-time image data of the road and surrounding environment in front of the vehicle; When it is judged that the vehicle is about to enter or leave the above-mentioned area, the type of the illumination change scene to be faced is determined; Based on the light change scene type and the real-time speed, driving direction of the vehicle, a corresponding light change prediction model is called and configured to predict the light change trend of each light sub-region.

[0013] Optionally, the step of acquiring the environmental information of the vehicle to identify whether there is a tunnel entrance, a tunnel exit, an under-bridge area, or a high-rise sheltered area in front of the vehicle includes: acquiring real-time image data of the road and the surrounding environment in front of the vehicle, performing semantic segmentation or instance segmentation on the real-time image data, and identifying the geometric features of buildings, mountains, bridge structures, and tunnel entrances; According to the relative position, contour and size of the geometric features of buildings, mountains, bridge structures and tunnel entrances identified by segmentation, it is judged whether they will form a large range of top shelter in front of the vehicle; If it is judged that there is a large range of top shelter, according to the longitudinal information and entrance characteristics of the shelter, it is determined to be a tunnel entrance, a tunnel exit, an under-bridge area, or a high-rise sheltered area.

[0014] Optionally, the step of calling and configuring a corresponding light change prediction model based on the light change scene type and the real-time speed, driving direction of the vehicle, to predict the light change trend of each light sub-region includes: According to the light change scene type, a corresponding light change prediction model is selected from a model library; wherein for a tunnel entrance scene, a prediction model focusing on brightness drop is selected, and for a high-rise sheltered scene, a prediction model focusing on shadow movement and color temperature change is selected; According to the real-time speed of the vehicle, when the vehicle is driving at high speed, the weight of the light change prediction model on the front long-distance environmental features is increased, and the weight of the light change prediction model on the instantaneous local environmental features is reduced; According to the driving direction of the vehicle, especially when in a sharp turn state, the weight of the light change prediction model on the lateral environmental features is increased, and the deformation prediction parameters of the light sub-region are adjusted according to the turning radius; The rate limit is applied to the weight updating process of the light change prediction model to ensure the smoothness of the prediction output.

[0015] Optionally, the step of acquiring the physiological baseline data of the driver includes: Determine whether the driver is in an abnormal physiological state by facial expression, eye movement features and voice tone; When it is judged that the driver is in an abnormal physiological state, the collection of the physiological baseline data is suspended, and when the driver's state returns to normal, the physiological baseline data is collected; In the process of collecting the physiological baseline data, the physiological index fluctuation of the driver is monitored; When the physiological index fluctuation exceeds a preset threshold, the physiological baseline data is re-collected.

[0016] In another aspect, the present application provides a vehicle-mounted display screen image enhancement system, which comprises: An illumination information acquisition module is configured to acquire illumination information of the display screen surface and divide the display screen into a plurality of illumination sub-regions according to the illumination information; An illumination change prediction module is configured to acquire real-time speed, driving direction and environmental information of the vehicle, and predict the illumination change trend of each illumination sub-region; A polarized sunglasses influence prediction module is configured to determine whether the driver wears polarized sunglasses according to pattern recognition of the optical characteristics of the eye region in the driver's face image, and predict the influence of the polarized sunglasses on the display screen image according to a preset optical model of the interaction between the polarized sunglasses and the display screen; An image enhancement parameter application module is configured to select and apply corresponding image enhancement parameters for each illumination sub-region according to the illumination information of the illumination sub-region, the illumination change trend and the influence of the polarized sunglasses on the display screen image.

[0017] The vehicle-mounted display screen image enhancement method and system provided by the present application can accurately identify the local uneven illumination phenomenon on the screen surface by acquiring the illumination information of the display screen surface and dividing it into a plurality of illumination sub-regions, overcoming the limitation that a single ambient light sensor in the prior art cannot reflect complex illumination environments. At the same time, the method predicts the illumination change trend of each illumination sub-region by acquiring the real-time speed, driving direction and environmental information of the vehicle, realizes the prediction of future illumination changes, and avoids the decline in information visibility caused by the lag response of traditional methods. In addition, the present application also innovatively determines whether the driver wears polarized sunglasses through pattern recognition, and predicts the influence of the polarized sunglasses on the display screen image according to an optical model, effectively compensating for the visual deviation introduced by the polarized sunglasses, and solving the problem that the existing system cannot perceive or compensate for the influence of such external observation media. Finally, the method can select and apply corresponding image enhancement parameters for each illumination sub-region according to the illumination information of the illumination sub-region, the illumination change trend and the influence of the polarized sunglasses, realize fine and personalized image enhancement, and significantly improve the image visibility, contrast and color accuracy of the vehicle-mounted display screen in various complex light environments, effectively solve the problem that key information cannot be clearly seen due to glare, overexposure or shadow, and color distortion and information black hole under polarized glasses in the background technology, and greatly improve the efficiency and driving safety of the driver in obtaining information. BRIEF DESCRIPTION OF DRAWINGS

[0018] To illustrate this application more clearly, the accompanying drawings used in the embodiments will be briefly described below. Obviously, those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0019] Figure 1 The diagram above illustrates a flowchart of an image enhancement method for an in-vehicle display screen. Figure 2 The diagram above illustrates a schematic of the structure of an image enhancement system for an in-vehicle display screen.

[0020] Figure reference numerals: 100, in-vehicle display image enhancement system; 10, illumination information acquisition module; 20, illumination change prediction module; 30, polarized sunglasses influence prediction module; 40, image enhancement parameter application module. Detailed Implementation

[0021] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0023] Traditional in-vehicle display systems often rely on a single ambient light sensor for global brightness adjustment in complex and changing lighting environments. This results in significantly reduced screen display quality in areas of direct sunlight or deep shadows, making it difficult to identify critical information. Especially when drivers wear polarized sunglasses, the display image may exhibit color distortion and information gaps, seriously impacting driving safety and the driving experience.

[0024] like Figure 1 The diagram illustrates an exemplary flowchart of an image enhancement method for an in-vehicle display screen. This application proposes an image enhancement method for an in-vehicle display screen, comprising: S10, obtain the illumination information of the display screen surface, and divide the display screen into a plurality of illumination sub-regions according to the illumination information; The illumination information refers to the brightness, color temperature, direction, and other physical properties of the ambient light received by each region of the display screen surface. The illumination sub-region refers to a local region with relatively uniform illumination characteristics, which is divided from the entire display screen surface according to the differences in illumination information.

[0025] S20, obtain the real-time speed, driving direction, and environmental information of the vehicle, and predict the illumination change trend of each illumination sub-region; S30, according to the pattern recognition of the optical characteristics of the eye region in the driver's face image, determine whether the driver wears polarized sunglasses, and according to the pre-set optical model of the interaction between the polarized sunglasses and the display screen, predict the influence of the polarized sunglasses on the display screen image; S40, according to the illumination information of the illumination sub-region, the illumination change trend, and the influence of the polarized sunglasses on the display screen image, select and apply corresponding image enhancement parameters for each illumination sub-region.

[0026] The image enhancement parameters refer to various parameters used to adjust the display effect of the image, such as brightness, contrast, gamma value, hue, saturation, etc. The adjustment of these parameters aims to optimize the visibility of the image under specific illumination conditions.

[0027] First, obtain the illumination information of the display screen surface, and divide the display screen into a plurality of illumination sub-regions according to the illumination information. As an implementation manner, a plurality of micro light sensor arrays can be deployed on the display screen surface or its vicinity to directly measure the illumination intensity and color temperature at different positions. These sensors can be distributed in a grid shape, with each sensor corresponding to one or a group of pixel regions, thereby obtaining fine illumination distribution data of the display screen surface. According to these data, clustering algorithms or threshold segmentation methods can be used to merge regions with similar illumination characteristics to form different illumination sub-regions. For example, regions with brightness higher than a certain threshold can be divided into high-brightness sub-regions, and regions with brightness lower than another threshold can be divided into shadow sub-regions.

[0028] In another implementation, a wide-angle camera or a dedicated image sensor installed inside the vehicle can be utilized to periodically capture images of the display surface. By analyzing these images, for example, by calculating the average pixel brightness, contrast, and color distribution of different regions of the image, the lighting information of the display surface can be indirectly inferred. Subsequently, based on these analysis results, the display surface can be divided into multiple lighting sub-regions, for example, through edge detection and region growing algorithms. For example, when image analysis shows that a certain region has a significant glare or reflection, this region can be identified as a high-lighting sub-region; while when a certain region has a significantly low brightness and insufficient contrast, it can be identified as a shadow lighting sub-region.

[0029] Secondly, the real-time speed, driving direction, and environmental information of the vehicle are obtained, and the lighting change trend of each lighting sub-region is predicted. As an implementation, the real-time speed and driving direction of the vehicle can be obtained through the vehicle CAN bus system. The environmental information can be obtained through sensors such as cameras, radars, or lidars outside the vehicle, for example, identifying whether there is a tunnel, viaduct, high-rise building, or other scenes that may cause a dramatic change in lighting in front of the vehicle. Combined with these information, a prediction model can be established. For example, when the vehicle is driving at high speed towards the entrance of a tunnel, the prediction model will judge that the brightness of all lighting sub-regions will rapidly decrease; when the vehicle is slowly driving in the city and passing through an area with high-rise buildings, the prediction model may predict that certain lighting sub-regions will periodically enter the shadow or be directly exposed to sunlight.

[0030] In another implementation, the GPS data of the vehicle and the preloaded high-precision map information can be utilized. By analyzing the vehicle's position and driving path on the map, combined with the environmental features such as buildings, terrain, road structure, etc. marked in the map, the lighting environment that the vehicle will enter or leave soon can be predicted in advance. For example, when the GPS data shows that the vehicle is about to enter a road section covered by a viaduct, it can be predicted that the lighting sub-regions above the display screen will gradually become dark. At the same time, combined with the real-time speed and driving direction of the vehicle, the prediction results can be further refined, for example, to predict the rate and duration of lighting changes.

[0031] Again, according to the pattern recognition of the optical features of the eye region in the driver's facial image, it is determined whether the driver wears polarized sunglasses, and the influence of the polarized sunglasses on the display screen image is predicted according to a preset optical model of the interaction between the polarized sunglasses and the display screen. As an implementation manner, the facial image of the driver can be captured in real time by an infrared camera or a visible light camera installed in the cockpit. The eye region of the driver can be identified by using image processing and pattern recognition technology. By analyzing the optical features of the eye region, such as whether there is a reflection of polarized light at a specific angle, the brightness variation pattern of the pupil region, etc., it can be determined whether the driver wears polarized sunglasses. Once it is determined that the driver wears polarized sunglasses, the preset optical model is called. The model can be a lookup table or a mathematical function, which describes the degree of attenuation or distortion of image brightness, contrast and color under different display screen polarization angles and polarized sunglasses polarization angles. For example, when the polarization direction of the display screen is close to perpendicular to the polarization direction of the polarized sunglasses, the model predicts that some areas of the display screen image will have significant brightness attenuation or even "black screen" phenomenon.

[0032] In another implementation manner, in addition to analyzing the optical features of the eye region, the head posture and line-of-sight direction information of the driver can also be combined. For example, when the head posture and line-of-sight direction of the driver indicate that he is staring at the display screen, the detection of the polarized sunglasses will be performed more frequently. The preset optical model can be a simulation model established based on physical optical principles, which can simulate the absorption and transmission characteristics of polarized sunglasses to light of different polarization states, and the visual effect of the interaction between the polarized light emitted by the liquid crystal display screen and the polarized sunglasses. Through the model, the influence of the polarized sunglasses on the brightness, contrast, color saturation and possible rainbow effect of the display screen image can be more accurately predicted.

[0033] Finally, according to the illumination information of the illumination sub-region, the illumination variation trend and the influence of the polarized sunglasses on the display screen image, corresponding image enhancement parameters are selected and applied to each illumination sub-region. As an implementation manner, an image enhancement parameter library can be maintained, which stores a set of preset parameters for different illumination conditions, illumination variation trends and influences of polarized sunglasses. For example, when a certain illumination sub-region is under direct sunlight and its brightness is predicted to further increase, and the driver wears polarized sunglasses, a set of parameters that can reduce brightness, increase contrast and compensate for color distortion caused by polarized sunglasses will be selected from the parameter library. These parameters are then applied to the illumination sub-region to optimize its display effect.

[0034] In another implementation, an adaptive algorithm can be employed to dynamically generate the image enhancement parameters. For example, a machine learning-based decision model can be established, which takes the illumination information, illumination change trend of the illumination sub-regions, and the impact of polarized sunglasses as inputs, and outputs the optimal image enhancement parameters. This model can be trained through a large amount of driving scene data, learning how to adjust the parameters to maximize the driver's visual comfort and information visibility under various complex conditions. For example, when it is detected that a certain illumination sub-region is about to enter the shadow, and the driver is not wearing polarized sunglasses, the model may output a set of parameters that increase the brightness and gamma value to ensure that the information in the shadow area is still clearly visible.

[0035] The vehicle-mounted display screen image enhancement method proposed in this application aims to solve the problem of poor display effect of traditional vehicle-mounted display systems in complex lighting environments, especially when the driver wears polarized sunglasses. This method obtains the illumination information of the display screen surface in detail and divides it into multiple illumination sub-regions, thereby being able to specifically handle the problem of local uneven illumination. For example, when direct sunlight causes local high-light areas on the display screen, the traditional method may only make uniform brightness adjustments to the entire screen, resulting in overexposure of the high-light areas and still overdarkness of the shadow areas. However, this application can identify high-light sub-regions and shadow sub-regions and apply different image enhancement parameters to them, such as reducing the brightness and increasing the contrast of the high-light sub-regions, while increasing the brightness of the shadow sub-regions, thereby ensuring that all information on the entire display screen is clearly visible.

[0036] In addition, this application also innovatively introduces the prediction of illumination change trends. For example, when the vehicle is about to enter a tunnel, it can predict in advance that the brightness of all illumination sub-regions of the display screen will rapidly decrease, and adjust the image enhancement parameters in advance, avoiding the response lag of traditional methods when the illumination changes dramatically, resulting in the driver's inability to see the screen information in a short period of time. This forward-looking adjustment significantly improves the timeliness and safety of the driver's information acquisition.

[0037] More importantly, this application takes into account the special impact of polarized sunglasses on the display screen image. This application determines whether the driver is wearing polarized sunglasses through pattern recognition and predicts its impact based on an optical model, thereby being able to specifically adjust the image enhancement parameters. For example, when it is determined that the driver is wearing polarized sunglasses and a certain illumination sub-region of the display screen may become dark or even "black screen" due to the polarization effect, the brightness of this region will be actively increased, the color saturation will be adjusted, and even the rendering method of the display content will be changed to ensure that even under polarized sunglasses, critical information can be clearly perceived.

[0038] In summary, the present application dynamically selects and applies customized image enhancement parameters for each illumination sub-region by comprehensively considering three key factors: the local illumination of the display screen surface, the future illumination trend, and whether the driver wears polarized sunglasses. This multi-dimensional, adaptive image enhancement strategy significantly outperforms the single, global adjustment method in the prior art, effectively solving the visibility problem of vehicle-mounted display screens in complex driving environments, and greatly improving the visual comfort and driving safety of drivers.

[0039] In some embodiments, the step of selecting and applying corresponding image enhancement parameters for each illumination sub-region according to the illumination information of the illumination sub-region, the illumination trend, and the influence of polarized sunglasses on the display screen image includes: acquiring physiological baseline data of the driver; after a small perturbation of the image enhancement parameters of the illumination sub-region in a preset mode, collecting the instantaneous physiological response data of the driver; comparing the instantaneous physiological response data with the physiological baseline data to determine the influence direction of the perturbation of the image enhancement parameters on the visual comfort of the driver; iteratively adjusting the image enhancement parameters of the illumination sub-region according to the influence direction until the physiological indicators of the driver tend to be stable.

[0040] Specifically, acquiring physiological baseline data of the driver means continuously monitoring and recording a series of physiological indicators of the driver, such as heart rate, heart rate variability, blink frequency, pupil size, skin conductance, and brain wave activity, when the driver is in a normal and relaxed state. These data are used to establish a physiological characteristic reference model representing the driver in a comfortable state, so as to evaluate the influence of image enhancement parameter adjustment on the visual comfort of the driver in the subsequent.

[0041] Wherein, after a small perturbation of the image enhancement parameters of the illumination sub-region in a preset mode, the instantaneous physiological response data of the driver is collected, which can be understood as a small amplitude and directional adjustment of the brightness, contrast, color temperature or gamma value of the specific illumination sub-region based on the selected and applied image enhancement parameters. For example, the brightness can be increased or decreased by a preset step, or the color temperature can be fine-tuned. After each perturbation, the instantaneous physiological response data of the driver is immediately collected, which reflects the instantaneous physiological response of the driver to the current image parameter change, such as pupil contraction or expansion, and change in blink frequency.

[0042] In practical applications, the difference between the real-time physiological response data and the pre-established physiological baseline data is analyzed by an algorithm to determine the influence direction of the disturbance of the image enhancement parameters on the driver's visual comfort. For example, if the pupil size tends to be stable, the blink frequency decreases, or the heart rate variability increases, it may indicate that the driver feels more comfortable with the disturbance, and the influence direction is positive. Conversely, if the pupil frequently fluctuates, the blink frequency increases, or the heart rate accelerates, it may indicate that the disturbance causes discomfort, and the influence direction is negative.

[0043] Further, according to the determined influence direction, the image enhancement parameters are continuously and gradually adjusted. For example, if the influence is determined to be positive, the parameters are fine-tuned in the same direction; if the influence is determined to be negative, the parameters are adjusted in the opposite direction. This process will continue until the physiological indicators of the driver decrease in fluctuation and approach or stabilize at the comfort level reflected by the physiological baseline data, thereby ensuring that the finally applied image enhancement parameters can maximize the driver's visual comfort.

[0044] The technical solution of the present application effectively solves the problem of individual comfort differences that may exist when only relying on environmental and driver state for parameter selection by introducing a closed-loop adjustment mechanism of driver physiological feedback. Specifically, after initially applying the image enhancement parameters, instead of simply maintaining these parameters, the real-time physiological response of the driver is actively detected by slightly disturbing the parameters. The physiological baseline data of the driver provides an objective comfort reference point, while the real-time physiological response data serves as direct and unconscious feedback to the parameter disturbance. By comparing the real-time response with the baseline, it can be accurately determined whether the current parameter adjustment has increased or decreased the driver's visual comfort. This influence direction determination based on physiological feedback enables the image enhancement parameters to be gradually optimized in an iterative manner until a balance point is found that meets the external environmental conditions and maximizes the individual visual comfort of the driver. Thus, the present technical solution converts subjective visual comfort into quantifiable physiological indicators, achieving personalized and dynamic optimization of image enhancement parameters.

[0045] By the technical solution, the individuality and adaptability of image enhancement of the vehicle-mounted display screen can be significantly improved. The technical solution not only considers objective factors such as external light and whether the driver wears polarized sunglasses, but also further integrates real-time physiological feedback of the driver, thereby ensuring that the applied image enhancement parameters can best match the individual visual comfort needs of the driver. This effectively avoids visual fatigue and discomfort caused by improper parameters, and improves the visual experience and driving safety of the driver in different driving environments. In addition, through iterative adjustment until the physiological indicators are stable, the technical solution can realize fine parameter optimization, so that the display screen image can provide the best visual comfort for the driver at any time, thereby being significantly superior to the traditional method of selecting parameters based on a preset model.

[0046] In some optional embodiments, assuming that the vehicle is in the process of driving, the vehicle-mounted display screen has preliminarily applied a set of image enhancement parameters according to information such as ambient light and whether the driver wears polarized sunglasses. In order to further optimize the visual comfort of the driver, first, the physiological baseline data of the driver is obtained, for example, the average heart rate and blink frequency measured in the relaxed state of the driver. Subsequently, the brightness parameter of a certain light sub-region of the display screen is slightly increased. After the disturbance occurs, the real-time physiological response data of the driver is immediately collected through the eye tracking device and heart rate monitor installed in the vehicle. If it is found that the blink frequency of the driver slightly decreases and the heart rate variability increases, it may indicate that the increase in brightness reduces the visual stress of the driver and improves the comfort. The brightness will continue to be increased in a small amount according to this positive influence direction. On the contrary, if the increase in brightness leads to an increase in the blink frequency of the driver or frequent pupil contraction, it indicates that the adjustment in this direction causes discomfort, and the adjustment will be reversed, that is, the brightness will be tried to be reduced. This iterative process will continue until the physiological indicators (such as blink frequency, pupil size, heart rate, etc.) of the driver tend to be stable, and a best comfortable state is reached, at which time the image enhancement parameters are determined and applied to the light sub-region. In this way, the image enhancement effect of the display screen can dynamically adapt to the real-time physiological state of the driver, providing a highly personalized visual experience.

[0047] In some embodiments, the step of selecting and applying corresponding image enhancement parameters for each light sub-region according to the light information of the light sub-region, the light change trend, and the influence of polarized sunglasses on the display screen image comprises: identifying the information type displayed in the light sub-region, and performing priority sorting on the information displayed in the light sub-region according to the information type to obtain a priority sorting result; selecting the visibility image enhancement parameter of the information type with the highest priority from a preset parameter set according to the priority sorting result; adjust the selected image enhancement parameters to optimize the visibility of other information types with lower priority, and apply the adjusted image enhancement parameters to the illumination sub-region.

[0048] Specifically, the information type displayed in the illumination sub-region is identified by determining the category of the content currently displayed in the illumination sub-region through image content analysis, metadata reading or application program interface (API) calling, etc. For example, the information type can be navigation information, vehicle status information, entertainment information, communication information or warning information, etc. The information displayed in the illumination sub-region is prioritized according to the information type, which can be based on a pre-set rule base, driver preference settings or real-time driving situation. For example, warning information or navigation instructions directly related to driving safety are usually given the highest priority, while entertainment or auxiliary information has a lower priority. Thus, a priority ranking result reflecting the importance of each information can be obtained.

[0049] Further, the image enhancement parameters for the visibility of the information type with the highest priority are selected from a pre-set parameter set according to the priority ranking result. The pre-set parameter set contains a combination of image enhancement parameters optimized for different information types and different environmental conditions, such as brightness, contrast, color saturation, sharpness, etc. The purpose of selecting these parameters is to ensure that the information with the highest priority has the best visibility under the current lighting conditions, i.e. it can be clearly presented even in strong direct light or when wearing polarized sunglasses.

[0050] On this basis, the selected image enhancement parameters are adjusted to optimize the visibility of other information types with lower priority without significantly affecting the visibility of the information type with the highest priority, and the adjusted image enhancement parameters are applied to the illumination sub-region. This means that the display effect of the most important information will be guaranteed first, and then the display parameters of other secondary information will be fine-tuned to improve its readability and visual comfort as much as possible without interfering with the main information. For example, the brightness or saturation of the background or non-critical elements can be appropriately reduced to reduce visual interference while maintaining their basic recognizability.

[0051] By the above technical solution, the application can significantly improve the information transmission efficiency of the vehicle-mounted display screen and the visual comfort of the driver under complex lighting conditions and when the driver wears polarized sunglasses. Specifically, by prioritizing the display information, it can ensure that critical driving information (such as navigation instructions, warning information) has the highest clarity and readability in any situation, avoiding the problem of critical information being blurred or difficult to identify due to environmental lighting or polarized sunglasses. At the same time, while ensuring the visibility of high-priority information, low-priority information is optimized and adjusted, making the entire display interface more visually coordinated and reducing the cognitive load and visual fatigue of the driver, thereby improving driving safety.

[0052] As a specific implementation, assume that during vehicle driving, a certain lighting sub-region of the vehicle-mounted display screen displays navigation path instructions, current vehicle speed, and background music playing information at the same time. At this time, the lighting sub-region is being strongly irradiated by lateral sunlight, and the driver is wearing polarized sunglasses.

[0053] First, the type of information displayed in the lighting sub-region will be identified: the navigation path instructions are identified as high-priority information, the current vehicle speed is identified as medium-priority information, and the background music playing information is identified as low-priority information.

[0054] Next, according to the priority sorting result, a set of image enhancement parameters will be selected and applied to the navigation path instructions first, such as increasing its brightness to the maximum, adjusting its contrast to the highest, and possibly using high-saturation colors to ensure that it remains clear and visible under the influence of strong light and polarized sunglasses.

[0055] Subsequently, while ensuring the visibility of the navigation path instructions, the image enhancement parameters of the current vehicle speed will be adjusted. For example, the speed display may be given a slightly lower brightness than the navigation instructions, but still maintain sufficient contrast to ensure readability.

[0056] Finally, for the background music playing information, its image enhancement parameters will be optimized without interfering with the navigation and speed display. For example, its brightness or saturation can be appropriately reduced so that it exists as background information, neither distracting the driver nor maintaining basic recognizability.

[0057] Through this hierarchical and regional image enhancement strategy, the driver can clearly see the navigation path instructions, quickly obtain the speed information, and the background music information is also presented softly, thereby obtaining the best visual experience under complex lighting conditions.

[0058] In some embodiments, the step of obtaining the lighting information of the display screen surface and dividing the display screen into multiple lighting sub-regions according to the lighting information comprises: acquire image data of the display screen surface through an image sensor and pre-process it to remove transient highlight or dark spot interference information caused by dynamic reflection or partial occlusion; acquire image data of the surrounding environment through an image sensor to obtain illumination characteristics of non-display screen areas to assist in judging the overall ambient light type and intensity in the vehicle; separate the display content itself luminous area and the area affected by ambient light in the image data of the display screen surface according to the image data of the display screen surface and the illumination characteristics of non-display screen areas; analyze the brightness, contrast and color distribution of the area affected by ambient light to identify local strong light direct area and deep shadow area; divide the display screen into multiple illumination sub-areas according to the boundaries and characteristics of the local strong light direct area and deep shadow area.

[0059] Specifically, the visual information of the display screen surface is captured by using a vehicle-mounted image sensor, such as a high dynamic range (HDR) camera, to realize the acquisition of image data of the display screen surface. The pre-processing aims to eliminate various interference in the image data, such as transient highlight area or dark shadow area caused by window reflection, instrument table reflection or driver body occlusion, etc. If these interference information is not removed, it will seriously affect the accurate analysis of subsequent illumination information. Among them, the pre-processing can include exposure fusion, local tone mapping, and image repair of glare and reflection area, etc. to ensure that the acquired image data can truly reflect the illumination condition of the display screen surface.

[0060] Further, real-time images of the vehicle interior or exterior environment are captured by using a vehicle-mounted camera to realize the acquisition of image data of the surrounding environment. The illumination characteristics of non-display screen areas can include overall brightness, color temperature, illumination direction and whether there is a strong light source, etc. These ambient light characteristics are used to assist in judging the type and intensity of the overall ambient light in the vehicle, such as sunny, cloudy, evening, night, and whether the illumination is uniform or there is strong local light.

[0061] On this basis, according to the image data of the display screen surface and the illumination characteristics of non-display screen areas, the display content itself luminous area and the area affected by ambient light in the image data of the display screen surface can be separated. Among them, the display content itself luminous area refers to the light emitted by the display screen pixel points, and the area affected by ambient light mainly refers to the reflection, glare or shadow of ambient light on the display screen surface. This separation can be realized by image processing algorithm, such as by analyzing the brightness distribution, color characteristics of the image or using deep learning model for pixel-level classification, so as to accurately distinguish the actual display content of the display screen and external light interference.

[0062] Subsequently, the brightness, contrast, and color distribution of the areas affected by ambient light are analyzed to identify local direct sunlight areas and deep shadow areas. Local direct sunlight areas are typically characterized by extremely high brightness and sharp contrast reduction, possibly caused by direct sunlight or intense external light sources. Deep shadow areas are characterized by extremely low brightness and difficulty in identifying details, possibly caused by internal structures or the driver's body blocking the light. By quantitatively analyzing the characteristics of these areas, such as setting brightness thresholds, contrast variation rates, or color shift amounts, these special lighting areas can be accurately identified and located.

[0063] Finally, according to the boundaries and characteristics of the local direct sunlight areas and deep shadow areas, the display screen is divided into multiple lighting sub-areas. For example, the direct sunlight area, the deep shadow area, and other normal lighting areas can be divided into different lighting sub-areas. Each lighting sub-area has its unique lighting conditions and visual challenges, laying the foundation for subsequent targeted image enhancement processing.

[0064] Through the above technical solution, the problem of inaccurate acquisition of display screen lighting information and rough partitioning in traditional methods can be overcome. The technical solution can accurately identify and remove various lighting disturbances through multi-step image data processing and analysis, and divide the display screen into multiple sub-areas with different lighting characteristics according to the actual lighting conditions. This enables the subsequent selection and application of image enhancement parameters to more accurately match the actual needs of each sub-area, avoiding local overexposure or underexposure caused by global uniform enhancement, and significantly improving the visual comfort and readability of the driver to the vehicle display screen information in complex and variable lighting environments.

[0065] In some embodiments, the step of obtaining image data of the display screen surface through an image sensor and preprocessing includes: obtaining multiple frames of original images of the display screen surface with different exposure parameters through a high dynamic range image sensor, and performing exposure fusion processing on the multiple frames of original images of the display screen surface to generate a high dynamic range image; performing local tone mapping processing on the high dynamic range image to compress the brightness range while preserving the details of highlight and shadow areas, and identifying and separating the glare area and the reflection area of the display screen surface according to the image after local tone mapping processing; performing image inpainting processing on the glare area and the reflection area to remove glare and reflection interference, and taking the image after image inpainting processing as the image data of the display screen surface.

[0066] Specifically, the above step of obtaining image data of the display screen surface through an image sensor and preprocessing can be further refined as follows.

[0067] A plurality of original images of the display screen surface with different exposure parameters are acquired by a high dynamic range image sensor, and exposure fusion processing is performed on the plurality of original images of the display screen surface to generate a high dynamic range image. The high dynamic range image sensor can capture a wider light range than a traditional image sensor. By acquiring a plurality of original images under different exposure times, details of extremely bright (such as direct sunlight) and extremely dark (such as deep shadows) regions that can appear on the display screen surface can be effectively covered. Exposure fusion processing refers to algorithmically combining these images with different exposures to generate a high dynamic range image that has good detail representation in all brightness regions.

[0068] Local tone mapping processing is performed on the high dynamic range image to compress the brightness range while preserving the details of the highlight and shadow regions. According to the image after local tone mapping processing, the glare region and the reflection region of the display screen surface are identified and separated. The local tone mapping processing aims to map the wide brightness range of the high dynamic range image to the range that can be presented by a standard display device, while maintaining the contrast and details of the image through local adjustment. For example, a tone mapping method based on the gradient domain or the frequency domain can be used. After this processing, since glare and reflection usually appear as regions with abnormally high brightness or specific texture patterns in the image, image processing algorithms such as threshold segmentation, edge detection, or machine learning-based classifiers can be used to identify and accurately separate these glare regions and reflection regions.

[0069] Image inpainting processing is performed on the glare region and the reflection region to remove glare and reflection interference, and the image after image inpainting processing is used as the image data of the display screen surface. Image inpainting processing, for example, can use algorithms based on image inpainting or texture synthesis to intelligently fill or replace the interference regions by analyzing the valid image information around the glare and reflection regions, thereby visually eliminating the occlusion and distortion of the display screen image caused by glare and reflection, and ultimately obtaining pure and accurate display screen surface image data.

[0070] The above technical solution can significantly improve the accuracy and robustness of the acquisition of image data of the display screen surface of the vehicle, especially in complex and variable vehicle lighting environments. The technical solution effectively solves the problem that traditional methods cannot accurately acquire real lighting information of the display screen surface under strong light, glare, or reflection interference, and avoids false judgments and inaccurate region division caused by interference information. Therefore, the accuracy of subsequent lighting sub-region division and the effectiveness of image enhancement parameter selection are ensured, thereby providing the driver with a clearer and more comfortable visual experience.

[0071] In some embodiments, the step of obtaining real-time speed, driving direction and environment information of the vehicle, and predicting the light change trend of each light sub-region comprises: obtaining environment information of the vehicle, identifying whether there is a tunnel entrance, a tunnel exit, an under-bridge area or a high-rise sheltered area in front of the vehicle, wherein the environment information comprises real-time image data of the road and surrounding environment in front of the vehicle; determining the type of light change scene to be faced when judging that the vehicle is about to enter or leave the above-mentioned area; based on the type of light change scene and the real-time speed and driving direction of the vehicle, calling and configuring a corresponding light change prediction model to predict the light change trend of each light sub-region.

[0072] Specifically, obtaining environment information of the vehicle aims to provide key context data for light change prediction. The environment information can be collected in real time by a vehicle-mounted sensor system (such as a camera, lidar, etc.), and mainly includes real-time image data of the road and surrounding environment in front of the vehicle. By analyzing these real-time image data, it can be identified whether there is a specific area in front of the vehicle that may cause dramatic light changes, such as a tunnel entrance, a tunnel exit, an under-bridge area or a high-rise sheltered area. These areas usually cause significant changes in light intensity, color temperature and distribution. When it is judged that the vehicle is about to enter or leave the above-mentioned identified specific area, the type of light change scene to be faced by the vehicle needs to be further determined. For example, entering a tunnel entrance means that the light will quickly become dark, while leaving a tunnel exit means that the light will quickly become bright. The under-bridge area or high-rise sheltered area may cause the formation and movement of local shadows, as well as changes in color temperature. Accurate identification of these scene types is the basis for selecting the appropriate prediction model. In practical applications, once the type of light change scene to be faced is determined, a corresponding light change prediction model will be called and configured based on the scene type, combined with the real-time speed and driving direction of the vehicle. Different light change scene types may correspond to different physical light change laws, so specially designed prediction models are needed. For example, for a tunnel entrance scene, a prediction model focusing on brightness drop may be needed; while for a high-rise sheltered scene, a prediction model focusing on shadow movement and color temperature change may be needed. The real-time speed and driving direction of the vehicle are used to adjust the prediction parameters of the model, such as the rate and range of prediction changes, to more accurately predict the light change trend of each light sub-region.

[0073] By the technical solution, the adaptability and prediction accuracy of the vehicle-mounted display screen image enhancement method to complex dynamic light environment can be improved. Specifically, by identifying a specific light change scene (such as a tunnel, an overpass, etc.) in advance and combining the real-time motion state of the vehicle, the most suitable light change prediction model for the current scene can be selected and applied, thereby realizing accurate prediction of the light change trend of each light sub-region. This refined prediction mechanism enables the image enhancement parameters of the display screen to be adjusted more timely and accurately, effectively avoiding driver visual discomfort or information recognition difficulty caused by sudden light changes, thereby improving the driver's visual comfort and driving safety.

[0074] In some optional embodiments, assume that a car is driving on a highway and is about to enter a long tunnel. Real-time image data of the road and surrounding environment in front of the vehicle is obtained through the front camera. The image processing module analyzes these image data and identifies that there is a tunnel entrance in front. Based on this identification result, it is judged that the light change scene type that the vehicle will face is “entering a tunnel”. At the same time, the real-time speed (for example, 80 kilometers / hour) and driving direction (straight ahead) of the vehicle are obtained. According to the scene type of “entering a tunnel” and the motion state of the vehicle, a light change prediction model specifically for brightness drop is called from the preset model library, and the prediction rate parameter of the model is adjusted according to the vehicle speed. The model will predict how the light intensity of each light sub-region on the display screen surface will rapidly decrease in the next few seconds, as well as the possible color temperature change. For example, it is predicted that the light intensity of the left region of the display screen will decrease from 5000 lux to 500 lux in 2 seconds, and the right region will complete similar changes in 3 seconds. Based on these predicted light change trends, the image enhancement parameter application module can select and apply the corresponding image enhancement parameters for each light sub-region in advance, such as increasing brightness, adjusting contrast, or changing color temperature, so as to start image optimization before the vehicle actually enters the tunnel, ensuring that the driver can clearly and comfortably view the display screen information when the light changes dramatically.

[0075] Specifically, the above steps of obtaining the environmental information of the vehicle, identifying whether there is a tunnel entrance, a tunnel exit, an overpass area, or a high-rise sheltered area in front of the vehicle include: obtaining real-time image data of the road and surrounding environment in front of the vehicle, performing semantic segmentation or instance segmentation on the real-time image data, and identifying the geometric features of buildings, mountains, bridge structures, and tunnel entrances; judging whether the buildings, mountains, bridge structures, and tunnel entrance geometric features identified by segmentation will form a large range of top shelter in front of the vehicle according to the relative position, contour, and size of the buildings, mountains, bridge structures, and tunnel entrance geometric features identified by segmentation; If a large area of ​​overhead obstruction is determined, based on the depth information of the obstruction and the characteristics of the entrance, it can be identified as a tunnel entrance, tunnel exit, area under an overpass, or area obstructed by a tall building.

[0076] Specifically, acquiring real-time image data of the road ahead and the surrounding environment can be achieved using onboard image sensors (e.g., visible light cameras, infrared cameras, or multispectral cameras). This real-time image data is used to capture visual information ahead of the vehicle's path. Semantic segmentation or instance segmentation of this real-time image data refers to using advanced image processing techniques, such as deep learning-based convolutional neural network models (e.g., U-Net, Mask R-CNN, or YOLO), to classify and identify each pixel or individual object in the image. In this way, objects with specific geometric features, such as buildings, mountains, bridge structures, and tunnel entrances, can be accurately distinguished and identified from complex environmental images. The aim is to accurately extract structural information from the environment that may lead to significant changes in lighting.

[0077] The process of determining whether identified buildings, mountains, bridge structures, and tunnel entrances will cause significant overhead occlusion in front of a vehicle, based on their relative positions, outlines, and dimensions, can be understood as performing geometric attribute analysis on the identified environmental objects. Relative position refers to the geographical relationship of these objects relative to the vehicle, such as their distance from the vehicle and their azimuth relative to the vehicle's direction of travel; outline refers to the boundary shape of these objects; and size refers to the actual size of these objects in space. By comprehensively analyzing these geometric attributes, it can be assessed whether these objects are large enough and whether their spatial position can significantly and extensively obstruct the light above the vehicle's path. For example, if a bridge structure with a width and height exceeding a preset threshold is identified ahead, and this structure is located directly above the vehicle's path, it can be determined that it will cause extensive overhead occlusion.

[0078] In practical applications, if a large-scale top occlusion is identified, it can be determined as a tunnel entrance, tunnel exit, area under an overpass, or area occupied by a tall building, based on the depth information and entrance features of the occlusion. Depth information refers to the length of the occluded area along the vehicle's travel direction, which can be obtained through depth estimation using vehicle-mounted LiDAR, stereo vision systems, or multi-view images. Entrance features refer to the unique geometric shape of the starting part of the occluded area. For example, tunnel entrances typically present as arched or rectangular openings, areas under overpasses may exhibit a structure open on both sides but occluded at the top, while tall building occlusion areas are usually composed of vertical walls. By comprehensively analyzing this depth information and entrance features, different types of occlusion scenes can be accurately distinguished, thus providing accurate scene type input for subsequent illumination change prediction models.

[0079] The above technical solution enables accurate identification of scene types involving changes in ambient lighting ahead of the vehicle. Compared to fuzzy judgments based solely on general environmental information, this solution, by introducing semantic segmentation or instance segmentation techniques combined with geometric feature analysis, can more meticulously and accurately identify specific occlusion structures and their types. This significantly improves the accuracy of identifying upcoming lighting change scene types, providing high-quality input for subsequent lighting change prediction models, thereby enhancing the adaptability and effectiveness of in-vehicle display image enhancement, and ultimately optimizing the driver's visual comfort.

[0080] In some embodiments, the step of calling and configuring the corresponding illumination change prediction model based on the illumination change scene type and the vehicle's real-time speed and driving direction, and predicting the illumination change trend of each illumination sub-region includes: Based on the type of lighting change scenario, a corresponding lighting change prediction model is selected from the model library; specifically, for the tunnel entrance scenario, a prediction model focusing on sudden brightness drop is selected, and for the tall building occlusion scenario, a prediction model focusing on shadow movement and color temperature change is selected. Based on the vehicle's real-time speed, when the vehicle is traveling at high speed, the weight of the illumination change prediction model on the distant environmental features ahead is increased, while the weight of the illumination change prediction model on the instantaneous local environmental features is decreased. Based on the vehicle's direction of travel, especially when making a sharp turn, the weight of the illumination change prediction model on lateral environmental features is increased, and the deformation prediction parameters of the illumination sub-region are adjusted according to the turning radius. A rate limit is imposed on the weight update process of the illumination change prediction model to ensure the stationarity of the prediction output.

[0081] Specifically, for different ambient lighting changes, the most suitable prediction algorithm for the current scenario is selected from a pre-established set of models, choosing the corresponding lighting change prediction model. For example, when a vehicle is about to enter a tunnel, the light intensity will drop sharply and significantly. In this case, a prediction model specifically optimized for sudden drops in brightness will be selected, which may focus more on the ability to predict rapid response and large-scale brightness adjustments. Conversely, when a vehicle is driving in an urban area with tall buildings, facing shadow movement and color temperature changes caused by building obstruction, a model that focuses on predicting shadow boundaries, movement trajectories, and ambient color temperature drift will be selected. This model may focus more on fine changes in local areas and color correction.

[0082] Specifically, when a vehicle is traveling at a higher speed, the model determines that distant environmental features (such as large buildings in the distance or the entrance to an upcoming tunnel) have a more significant impact on future changes in illumination, thus increasing the weight of these distant features in the prediction model. Conversely, instantaneous local environmental features (such as temporary shade from roadside trees or reflections from small billboards) have a smaller long-term impact on the illumination of high-speed vehicles, therefore their weight in the prediction model is reduced. This approach aims to enable the prediction model to more effectively capture illumination change trends that have a long-term impact on high-speed vehicles.

[0083] In practical applications, the weighting of lateral environmental features in the illumination change prediction model is increased based on the vehicle's direction of travel, especially when the vehicle is making a sharp turn. This is because the vehicle's lateral field of vision becomes particularly important during a sharp turn, as the lateral environment (e.g., buildings, mountains, or vegetation outside the curve) quickly enters the vehicle's forward field of vision and affects the illumination of the display screen. Furthermore, the deformation prediction parameters for the illumination sub-regions are adjusted based on the turning radius. For example, during a sharp turn, the driver's field of vision changes significantly, and the ambient lighting effects on different illumination sub-regions on the display screen may no longer be simple translations but may involve deformation. Therefore, the prediction parameters need to be adjusted to more accurately simulate this deformation.

[0084] Furthermore, a rate limit is imposed on the weight update process of the illumination change prediction model to ensure the stability of the prediction output. This means that when adjusting the model weights, large jumps in weights within a short period of time are not allowed; instead, adjustments are made gradually at a controlled rate. This effectively avoids instability in prediction results caused by fluctuations in environmental information or sensor noise, thereby preventing frequent or drastic changes in display image enhancement parameters and ensuring the consistency and comfort of the driver's visual experience.

[0085] Through the aforementioned technical solutions, the image enhancement method for in-vehicle displays can achieve more accurate and robust predictions of changes in ambient lighting. This refined model selection and parameter configuration allows the application of image enhancement parameters to better adapt to various complex driving environments. For example, when rapidly entering or exiting tunnels, navigating between tall buildings, or making sharp turns, the image enhancement effect of the display can be adjusted more promptly, accurately, and smoothly. This significantly improves the driver's visual comfort under different lighting conditions and ensures the clear readability of the display information, thereby enhancing driving safety.

[0086] In some embodiments, the step of acquiring the driver's physiological baseline data includes: Determine whether the driver is in an abnormal physiological state by observing the driver's facial expressions, eye movements, and tone of voice. When it is determined that the driver is in an abnormal physiological state, the collection of the physiological baseline data is suspended. When the driver's state returns to normal, the physiological baseline data is collected again. During the collection of the aforementioned physiological baseline data, fluctuations in the driver's physiological indicators are monitored; When the fluctuation of the physiological indicators exceeds the preset threshold, the physiological baseline data is re-collected.

[0087] Specifically, determining whether a driver is in an abnormal physiological state can be done by analyzing various physiological and behavioral signals, such as facial expressions, eye movements, and tone of voice. For example, facial expressions may indicate fatigue, drowsiness, irritability, or lack of concentration; eye movements may include blinking frequency, eye movement trajectory, and pupil size changes; abnormal eye movement patterns may indicate driver fatigue or distraction; tone of voice can reflect the driver's emotional state or cognitive load. These features can be collected in real time using sensors such as in-vehicle cameras and microphones, and analyzed using pattern recognition algorithms. When a driver is determined to be in an abnormal physiological state, the collection of physiological baseline data will be paused. This is to avoid collecting abnormal data when the driver is in a poor state, thus affecting the accuracy of the baseline data. Physiological baseline data collection will only resume when the driver's state returns to normal, for example, through rest or emotional adjustment. Furthermore, during the collection of the physiological baseline data, fluctuations in the driver's physiological indicators will be continuously monitored. These physiological indicators may include heart rate, skin conductance, and electroencephalogram (EEG), which reflect the driver's physiological stress level and cognitive state. By monitoring these indicators in real time, abnormal changes in the driver's physiological state can be detected promptly. In practical applications, when the fluctuations of these physiological indicators exceed preset thresholds, such as a sudden increase or decrease in heart rate or abnormal fluctuations in skin conductance, it indicates that the currently collected physiological data may be affected by external interference or changes in the driver's own state, and is no longer representative. At this time, the physiological baseline data will be re-collected to ensure that the obtained baseline data is stable and reliable, and can accurately reflect the driver's physiological response under normal conditions.

[0088] By employing the aforementioned technical solution, this application effectively avoids the problem of inaccurate physiological baseline data caused by abnormal driver physiological states. Compared to basic technical solutions that directly collect physiological baseline data, this application significantly improves the reliability and representativeness of physiological baseline data by introducing mechanisms for driver state assessment, data acquisition pause and resumption, and physiological indicator fluctuation monitoring and re-acquisition. This allows for a more accurate determination of the true impact of parameter perturbations on driver visual comfort when comparing the collected real-time physiological response data with high-quality physiological baseline data after minor perturbations to image enhancement parameters. Consequently, the iterative adjustment process of image enhancement parameters becomes more precise and efficient, ultimately achieving image enhancement effects that better meet the individual driver's visual comfort needs, thereby improving the driving experience and driving safety.

[0089] This application also proposes an image enhancement system for an in-vehicle display screen, such as... Figure 2 As shown, an image enhancement system 100 for an in-vehicle display screen includes: The illumination information acquisition module 10 is used to acquire illumination information on the surface of the display screen and divide the display screen into multiple illumination sub-regions based on the illumination information. The illumination change prediction module 20 is used to acquire the vehicle's real-time speed, driving direction and environmental information, and predict the illumination change trend of each illumination sub-region. The polarized sunglasses impact prediction module 30 is used to determine whether the driver is wearing polarized sunglasses based on pattern recognition of the optical features of the eye area in the driver's facial image, and to predict the impact of polarized sunglasses on the display screen image based on a preset optical model of the interaction between polarized sunglasses and the display screen. The image enhancement parameter application module 40 is used to select and apply corresponding image enhancement parameters for each illumination sub-region based on the illumination information of the illumination sub-region, the illumination change trend, and the influence of polarized sunglasses on the display screen image.

[0090] The in-vehicle display image enhancement system proposed in this application aims to solve the problem of poor display performance of traditional in-vehicle display systems under complex lighting conditions, especially when the driver is wearing polarized sunglasses. This system uses a lighting information acquisition module to finely acquire the lighting information of the display surface and divide it into multiple lighting sub-regions, thereby addressing the problem of uneven local lighting. For example, when direct sunlight causes bright areas on the display, traditional systems may only adjust the brightness of the entire screen uniformly, resulting in overexposed bright areas while shadow areas remain too dark. The system in this application, however, can identify the bright and shadow sub-regions and apply different image enhancement parameters to them respectively by the image enhancement parameter application module. For example, it reduces the brightness and increases the contrast of the bright sub-regions, while simultaneously increasing the brightness of the shadow sub-regions, thus ensuring that the information on the entire display is clearly visible.

[0091] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for enhancing images on a vehicle-mounted display screen, characterized in that, include: Acquire illumination information on the surface of the display screen, and divide the display screen into multiple illumination sub-regions based on the illumination information; By acquiring real-time vehicle speed, driving direction, and environmental information, the illumination change trend of each illumination sub-region can be predicted. Based on pattern recognition of the optical features of the eye area in the driver's facial image, it is determined whether the driver is wearing polarized sunglasses, and the impact of polarized sunglasses on the display screen image is predicted based on a preset optical model of the interaction between polarized sunglasses and the display screen. Based on the illumination information of the illumination sub-region, the illumination change trend, and the impact of polarized sunglasses on the display screen image, corresponding image enhancement parameters are selected and applied for each illumination sub-region.

2. The image enhancement method for a vehicle-mounted display screen according to claim 1, characterized in that, The step of selecting and applying corresponding image enhancement parameters for each illumination sub-region based on the illumination information of the illumination sub-region, the illumination change trend, and the impact of polarized sunglasses on the display screen image includes: Obtain the driver's physiological baseline data; After making slight perturbations to the image enhancement parameters of the illuminated sub-region in a preset mode, the driver's real-time physiological response data is collected. By comparing the real-time physiological response data with the physiological baseline data, the direction of the impact of the image enhancement parameter perturbation on the driver's visual comfort is determined. Based on the direction of influence, the image enhancement parameters of the illuminated sub-region are iteratively adjusted until the driver's physiological indicators tend to stabilize.

3. The image enhancement method for a vehicle-mounted display screen according to claim 1, characterized in that, The step of selecting and applying corresponding image enhancement parameters for each illumination sub-region based on the illumination information of the illumination sub-region, the illumination change trend, and the impact of polarized sunglasses on the display screen image includes: Identify the type of information displayed within the illuminated sub-region, and prioritize the information displayed within the illuminated sub-region according to the information type to obtain a priority ranking result; Based on the priority sorting results, select image enhancement parameters for the visibility of the information type with the highest priority from a preset parameter set; Without significantly affecting the visibility of the highest priority information type, the selected image enhancement parameters are adjusted to optimize the visibility of other lower priority information types, and the adjusted image enhancement parameters are applied to the illumination sub-region.

4. The image enhancement method for a vehicle-mounted display screen according to claim 1, characterized in that, The step of acquiring illumination information of the display screen surface and dividing the display screen into multiple illumination sub-regions based on the illumination information includes: Image data of the display surface is acquired by an image sensor and preprocessed to remove interference information such as instantaneous bright spots or dark spots caused by dynamic reflections or partial occlusion. Image data of the surrounding environment is acquired by an image sensor to obtain the illumination characteristics of non-display areas, which helps to determine the overall illumination type and intensity of the vehicle interior. Based on the image data of the display screen surface and the illumination characteristics of the non-display screen area, the areas of the display content that emit its own light and the areas affected by ambient light in the image data of the display screen surface are separated. Analyze the brightness, contrast, and color distribution of areas affected by ambient light to identify areas of direct, strong light and deep shadow. Based on the boundaries and characteristics of the local strong light direct illumination area and the deep shadow area, the display screen is divided into multiple illumination sub-areas.

5. The image enhancement method for a vehicle-mounted display screen according to claim 4, characterized in that, The step of acquiring image data of the display screen surface through an image sensor and performing preprocessing includes: The high dynamic range image sensor acquires multiple frames of original images of the display surface with different exposure parameters, and performs exposure fusion processing on the multiple frames of original images of the display surface to generate a single frame of high dynamic range image. The high dynamic range image is subjected to local tone mapping processing to compress the brightness range while preserving details in the highlight and shadow areas. Based on the image after local tone mapping processing, the glare area and the reflection area on the display surface are identified and separated. Image restoration processing is performed on the glare area and the reflection area to remove glare and reflection interference, and the image after image restoration processing is used as the image data of the display screen surface.

6. The image enhancement method for a vehicle-mounted display screen according to claim 1, characterized in that, The steps of acquiring the vehicle's real-time speed, driving direction, and environmental information, and predicting the illumination change trend of each illumination sub-region include: The vehicle acquires environmental information to identify whether there are tunnel entrances, tunnel exits, areas under overpasses, or areas obstructed by tall buildings in front of the vehicle. The environmental information includes real-time image data of the road in front of the vehicle and the surrounding environment. When it is determined that the vehicle is about to enter or leave the above-mentioned area, the type of lighting change scenario that it will encounter is identified. Based on the type of lighting change scenario and the vehicle's real-time speed and direction of travel, the corresponding lighting change prediction model is invoked and configured to predict the lighting change trend of each lighting sub-region.

7. The image enhancement method for a vehicle-mounted display screen according to claim 6, characterized in that, The step of acquiring environmental information about the vehicle and identifying whether there are tunnel entrances, tunnel exits, areas under overpasses, or areas obstructed by tall buildings in front of the vehicle, wherein the environmental information includes real-time image data of the road in front of the vehicle and the surrounding environment, includes: Real-time image data of the road ahead and the surrounding environment of the vehicle are acquired, and semantic segmentation or instance segmentation is performed on the real-time image data to identify the geometric features of buildings, mountains, bridge structures and tunnel entrances. Based on the relative position, outline, and size of the geometric features of buildings, mountains, bridge structures, and tunnel entrances identified by segmentation, it is determined whether they will form a large area of ​​overhead obstruction in front of the vehicle. If a large area of ​​overhead obstruction is determined, based on the depth information of the obstruction and the characteristics of the entrance, it can be identified as a tunnel entrance, tunnel exit, area under an overpass, or area obstructed by a tall building.

8. The image enhancement method for a vehicle-mounted display screen according to claim 6, characterized in that, The step of predicting the illumination change trend of each illumination sub-region by calling and configuring the corresponding illumination change prediction model based on the illumination change scene type and the vehicle's real-time speed and driving direction includes: Based on the type of lighting change scenario, a corresponding lighting change prediction model is selected from the model library; specifically, for the tunnel entrance scenario, a prediction model focusing on sudden brightness drop is selected, and for the tall building occlusion scenario, a prediction model focusing on shadow movement and color temperature change is selected. Based on the vehicle's real-time speed, when the vehicle is traveling at high speed, the weight of the illumination change prediction model on the distant environmental features ahead is increased, while the weight of the illumination change prediction model on the instantaneous local environmental features is decreased. Based on the vehicle's direction of travel, especially when making a sharp turn, the weight of the illumination change prediction model on lateral environmental features is increased, and the deformation prediction parameters of the illumination sub-region are adjusted according to the turning radius. A rate limit is imposed on the weight update process of the illumination change prediction model to ensure the stationarity of the prediction output.

9. The image enhancement method for a vehicle-mounted display screen according to claim 2, characterized in that, The steps for obtaining the driver's physiological baseline data include: Determine whether the driver is in an abnormal physiological state by observing the driver's facial expressions, eye movements, and tone of voice. When it is determined that the driver is in an abnormal physiological state, the collection of the physiological baseline data is suspended. When the driver's state returns to normal, the physiological baseline data is collected again. During the collection of the aforementioned physiological baseline data, fluctuations in the driver's physiological indicators are monitored; When the fluctuation of the physiological indicators exceeds the preset threshold, the physiological baseline data is re-collected.

10. An image enhancement system for a vehicle-mounted display screen, characterized in that, The system includes: The illumination information acquisition module is used to acquire illumination information on the surface of the display screen and divide the display screen into multiple illumination sub-regions based on the illumination information; The illumination change prediction module is used to acquire the vehicle's real-time speed, driving direction, and environmental information, and predict the illumination change trend of each illumination sub-region. The polarized sunglasses impact prediction module is used to determine whether the driver is wearing polarized sunglasses based on pattern recognition of the optical features of the eye area in the driver's facial image, and to predict the impact of polarized sunglasses on the display screen image based on a preset optical model of the interaction between polarized sunglasses and the display screen. The image enhancement parameter application module is used to select and apply corresponding image enhancement parameters for each illumination sub-region based on the illumination information of the illumination sub-region, the illumination change trend, and the influence of polarized sunglasses on the display screen image.

Images