A touch display screen image enhancement method and system

Abstract

The application relates to the technical field of image enhancement, and provides a touch display screen image enhancement method and system, which comprises the following steps: acquiring local light intensity data from a plurality of light ray sensing units arranged on the surface of a display screen or in the vicinity of the display screen, and acquiring interaction event information from a user interaction sensing unit; identifying a persistent light difference area and a local shielding area existing on the display screen according to the local light intensity data and the interaction event information; constructing a regional light state representation of the display screen based on the identification result, wherein the regional light state representation divides the display screen into a plurality of display areas and contains light state information of each display area; and applying corresponding display parameter adjustment to each display area of the display screen according to the regional light state representation, so as to optimize the display effect of each display area. The application can realize fine image enhancement of the touch display screen under a complex light environment.

Description

Technical Field

[0002] This application relates to the field of image enhancement technology, and more specifically, to a method and system for enhancing images on a touch display screen. Background Technology

[0003] Existing touchscreen displays struggle to cope effectively with complex and variable lighting conditions, especially in automotive environments. This often leads to issues such as unbalanced image contrast, reduced color saturation, and sluggish display response, severely impacting user experience. For example, when a vehicle is driving sideways towards the sun, sunlight may obliquely illuminate a portion of the central control screen, creating a high-intensity glare area, while another part of the screen remains in shadow. In this situation, a single, fixed-position light sensor can only provide an average reading of the overall brightness inside the vehicle, failing to distinguish these extreme localized lighting differences on the screen. Furthermore, when the driver interacts with the screen, their hands and arms cast moving, irregularly shaped, momentary shadows. These localized shadows introduced by user interaction further interfere with the screen's display quality. Summary of the Invention

[0004] This application provides a touch screen image enhancement method and system to solve at least one of the above-mentioned technical problems.

[0005] To achieve the above objectives, this application adopts the following technical solution:

[0006] In a first aspect, this application discloses a touch screen image enhancement method, comprising:

[0007] Acquire local light intensity data from multiple light sensing units arranged on or near the display screen surface, as well as interaction event information from user interaction sensing units;

[0008] Based on local light intensity data and interactive event information, identify areas of persistent illumination difference and local occlusion on the display screen;

[0009] Based on the recognition results, a regional illumination state representation of the display screen is constructed. This regional illumination state representation divides the display screen into multiple display areas and includes the light state information of each display area.

[0010] Based on the regional illumination status, the corresponding display parameters are adjusted for each display area of ​​the screen to optimize the display effect of each display area.

[0011] Secondly, this application also discloses a touch screen image enhancement system, comprising:

[0012] The data acquisition module is used to acquire local light intensity data from multiple light sensing units arranged on or near the display screen surface, as well as interaction event information from the user interaction sensing unit;

[0013] The area recognition module is used to identify areas of persistent illumination difference and partial occlusion on the display screen based on local light intensity data and interactive event information.

[0014] The state representation construction module is used to construct a regional illumination state representation of the display screen based on the recognition results. This regional illumination state representation divides the display screen into multiple display areas and contains the light state information of each display area.

[0015] The display adjustment module is used to adjust the corresponding display parameters of each display area of ​​the screen according to the regional lighting status, so as to optimize the display effect of each display area.

[0016] Compared with the prior art, this application has at least the following beneficial effects:

[0017] This application can avoid overexposure or underexposure in local areas caused by global adjustments, as well as display instability caused by user interaction. Attached Figure Description

[0018] Figure 1 A flowchart illustrating a touch screen image enhancement method provided in this application;

[0019] Figure 2 This is a schematic diagram of the structure of a touch screen image enhancement system provided in this application. Detailed Implementation

[0020] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0021] like Figure 1 As shown in the figure, this application proposes a touch screen image enhancement method, including:

[0022] Acquire local light intensity data from multiple light sensing units arranged on or near the display screen surface, as well as interaction event information from user interaction sensing units;

[0023] Based on local light intensity data and interactive event information, identify areas of persistent illumination difference and local occlusion on the display screen;

[0024] Based on the recognition results, a regional illumination state representation of the display screen is constructed. The regional illumination state representation divides the display screen into multiple display areas and includes the light state information of each display area.

[0025] Based on the regional illumination status, the corresponding display parameters are adjusted for each display area of ​​the screen to optimize the display effect of each display area.

[0026] Light sensing units are sensors capable of detecting ambient light intensity or spectral information, such as photoresistors, photodiode arrays, or multispectral sensors. These units are strategically placed on or near the display surface to acquire local light intensity data covering different areas of the display, rather than single global illumination data. User interaction sensing units are sensors used to detect user interaction with the display, such as touch sensors, proximity sensors, infrared sensors, or small cameras. Their purpose is to acquire interaction event information, such as touch location, gestures, or object proximity.

[0027] Local light intensity data refers to numerical values ​​reflecting the brightness of the area where each light sensing unit is located, independently collected by each unit. Interaction event information refers to data generated when a user interacts with the display screen, such as touch coordinates, press time, and gesture type. Persistent illumination difference areas refer to regions on the display screen where the light intensity is significantly different from the surrounding area over a long period, such as areas under direct sunlight or areas in fixed shadow. Partial occlusion areas refer to areas where the user's hand or other objects temporarily block the light from the display screen. Area illumination state representation is a data structure that logically divides the display screen into multiple independent display areas and associates the current light state information, such as brightness, color temperature, and contrast, with each area. Display parameter adjustment refers to optimizing display parameters such as brightness, contrast, color temperature, and gamma value for each display area based on its light state information to improve visual effects.

[0028] The touch screen image enhancement method provided in this embodiment will be described in further detail below:

[0029] First, it is necessary to acquire local light intensity data from multiple light sensing units arranged on or near the display screen surface, as well as interaction event information from user interaction sensing units.

[0030] One implementation approach is to use a distributed array of photosensors. For example, an array of multiple photoresistors or photodiodes can be integrated inside the bezel of the display or below the display area. Each photosensor monitors the light intensity of a specific small area on the display. These sensors can be arranged in a grid pattern or non-uniformly depending on the size of the display and the desired resolution. When ambient light shines on the display, each sensor outputs an analog or digital signal representing the local light intensity at its location. These signals are then acquired and digitized to form local light intensity data.

[0031] As another approach, light sensing functionality can be integrated into the pixel layer of the display. For example, some advanced display technologies allow certain pixels to function as light sensors when not displaying images. Through time-division multiplexing or spatial multiplexing techniques, these pixels can periodically switch to light-sensing mode to obtain more refined local light intensity data on the display surface.

[0032] For the user interaction sensing unit, a capacitive touchscreen controller can be used. The capacitive touchscreen itself provides precise touch location information, which can be considered as interaction event information. When a user touches the screen, the touch controller reports data such as the coordinates of the touch point, the touch area, and the pressure applied (if supported). This data can be used to determine whether the user is interacting with the screen and the specific location of the interaction.

[0033] In addition, near-field infrared sensors can be used. Multiple small infrared emitters and receivers are integrated into the display bezel. When a user's hand approaches or blocks the screen, the infrared light is reflected or blocked, thus being detected by the receiver. By analyzing the signal changes from different sensors, the approximate position and movement trajectory of the user's hand can be inferred as interaction event information.

[0034] Secondly, based on local light intensity data and interactive event information, areas of persistent lighting difference and local occlusion on the display screen are identified.

[0035] One approach is to use statistical analysis to identify regions of persistent illumination difference. For example, the system can continuously collect local light intensity data from each light sensing unit and calculate its average and standard deviation over a period of time. If the average light intensity of a certain area consistently differs significantly from the average of the surrounding area (e.g., exceeding a preset threshold), it can be identified as a region of persistent illumination difference. This method can effectively detect fixed or slowly changing illumination unevenness caused by external light sources (such as sunlight or vehicle interior lights).

[0036] As an alternative approach, image processing-based methods can be used to identify partially occluded areas. For example, the system can use a miniature camera (if present) above the display to capture images of the display surface. By analyzing these images in real time, such as using background subtraction, motion detection, or object recognition algorithms, it can detect whether an object (such as a hand) is approaching or obstructing the screen surface. When an object is detected obstructing the view, combined with the instantaneous decrease in light intensity in that area by the light sensing unit, the area can be identified as partially occluded.

[0037] Furthermore, based on the recognition results, a regional illumination state representation of the display screen is constructed. The regional illumination state representation divides the display screen into multiple display areas and includes the light state information of each display area.

[0038] One implementation method is grid partitioning. The entire display screen is logically divided into a regular grid, such as an 8x8 or 16x16 area. Each grid area corresponds to one or a group of light sensing units. When areas with persistent illumination differences or local occlusion are identified, these areas are mapped onto the corresponding grid units. Then, based on the light intensity data of these areas, each grid unit is filled with its light state information, such as average brightness, contrast ratio, and color temperature. For grid units not identified as special illumination areas, their light state information can be estimated based on their surrounding area or global ambient light.

[0039] As another implementation method, dynamic clustering can be used. Instead of pre-setting a fixed grid, the system dynamically divides the display screen into different display areas based on the shape, size, and location of continuously differentiated lighting areas and locally occluded areas identified in real time. For example, if an irregularly shaped area under direct sunlight is detected, the system will treat it as an independent display area. Then, for each dynamically divided display area, the system calculates the light sensing unit data within it and generates the light state information for that area. This method can more flexibly adapt to complex and changing lighting environments.

[0040] Finally, based on the regional illumination status, the corresponding display parameters are adjusted for each display area of ​​the screen to optimize the display effect of each display area.

[0041] One implementation method is the lookup table (LUT) mapping approach. A lookup table containing various lighting states and corresponding display parameter adjustment schemes is pre-built. Once the regional lighting state representation of the display screen is constructed, the system finds the best-matching entry in the lookup table based on the lighting state information of each display area and applies its corresponding display parameter adjustment scheme. For example, for an area identified as high-brightness direct light, the system might lower its brightness, increase contrast, and adjust the color temperature to reduce glare; while for an area identified as having localized shadows, the system might increase its brightness, decrease contrast, and adjust the gamma value to enhance shadow details.

[0042] As another implementation approach, a model-based adaptive adjustment method can be employed. The system establishes a mathematical model that can calculate the optimal display parameter adjustments in real time based on the input lighting state information. For example, this could be a neural network-based model that learns the optimal display effect under different lighting conditions through extensive training data. After obtaining the regional lighting state representation, the lighting state information of each display area is input into the model, which then outputs adjustment suggestions for parameters such as brightness, contrast, and color temperature for that area. These adjustment suggestions are then sent to the display driver to achieve regional display optimization.

[0043] This application, by employing multiple light sensing units arranged on or near the display screen surface, can acquire local light intensity data covering different areas of the display screen, thereby overcoming the limitation of a single sensor being unable to detect local illumination differences. Simultaneously, the interaction event information provided by the user interaction sensing unit enables the system to identify momentary local occlusion caused by user operations.

[0044] After acquiring this multi-source data, the system can intelligently identify areas with persistent lighting differences (such as areas under direct sunlight or fixed shadows) and areas with partial occlusion (such as the shadow of a user's hand) on the display screen. This identification process is one of the key innovations of this application, as it distinguishes between static uneven external lighting and dynamic user interaction occlusion, laying the foundation for subsequent fine-tuning.

[0045] Based on these recognition results, the system constructs a regional illumination state representation of the display screen. This means that the display screen is no longer considered a single unit, but is logically divided into multiple independent display areas, each with its own unique illumination state information. This regionalized representation allows the system to customize processing for the specific illumination conditions of different areas.

[0046] Ultimately, based on the illumination status of this area, the system can adjust the corresponding display parameters for each display area of ​​the screen. For example, for areas under direct strong light, the brightness can be reduced and the contrast increased to reduce glare; for partially obscured areas, the brightness of that area can be appropriately increased and the gamma value adjusted to enhance details in dark areas, thus ensuring that information in the obscured areas remains clearly visible. This regional, adaptive adjustment of display parameters allows the screen to provide optimal visual effects under any complex lighting conditions, significantly improving user experience and information readability.

[0047] The core of this application lies in the introduction of multiple light sensing units and user interaction sensing units, enabling refined perception of the local lighting environment of the display screen and distinguishing between areas of continuous illumination difference and areas of partial occlusion. This capability of multi-source data fusion and intelligent regional recognition is not found in existing technologies. By constructing a regional illumination state representation, this application logically divides the display screen into multiple independent display areas and provides customized display parameter adjustment schemes for each area. For example, when sunlight shines directly on a corner of the screen, this application can optimize the brightness, contrast, and color temperature only in that area without affecting the display effect of other areas of the screen under normal illumination. Similarly, when a user's hand covers a part of the screen, the system can quickly identify and only perform local brightness enhancement and detail enhancement on the obscured area, ensuring that the user can still clearly see the screen content during operation, without the overall brightness fluctuation or unreadable information caused by hand obstruction in traditional solutions.

[0048] This refined, regionalized image enhancement method enables the present application to respond more accurately and promptly to various lighting changes, especially in dynamic user interaction scenarios, effectively solving the problems of poor display effects and poor information readability in traditional methods. Therefore, the present application demonstrates significant technological advancements and practical value in improving user visual comfort and ensuring effective information delivery.

[0049] In some embodiments, the step of identifying areas of persistent illumination difference and areas of partial occlusion on the display screen based on local light intensity data and interaction event information includes:

[0050] When the local light intensity data of multiple light sensing units continuously differ above a preset threshold, the region of continuous illumination difference is identified.

[0051] When local light intensity data shows a momentary drop in a specific area, that area is marked as a potential occlusion area;

[0052] For potential occlusion areas, the light intensity changes of light sensing units in these areas at multiple specific wavelengths are analyzed to obtain the spectral variation patterns of potential occlusion areas.

[0053] The similarity between the spectral variation pattern and the pre-stored typical spectral feature patterns representing hand occlusion is calculated.

[0054] When the similarity reaches a preset threshold, the potential occlusion area is identified as a local occlusion area.

[0055] Specifically, multiple light sensing units can be arranged on or near the display surface to acquire local light intensity data of different areas of the display in real time. When the local light intensity data acquired by these light sensing units consistently shows a difference higher than a preset threshold over a period of time—for example, the left side of the display is consistently brighter than the right side—a region with a persistent lighting difference can be identified. This difference is usually caused by a fixed light source or shadow in the environment.

[0056] When the local light intensity data of a specific area suddenly and briefly drops, that area is initially marked as a potential occlusion area. This instantaneous drop can be caused by a variety of factors, including user hand occlusion, occlusion by other objects, or rapid changes in ambient light.

[0057] To accurately distinguish hand occlusion from other types of occlusion, areas marked as potential occlusion are further analyzed for changes in light intensity at multiple specific wavelengths within the light-sensing units. For example, light intensity data at different wavelengths such as red, green, and blue can be acquired to construct a spectral variation pattern for the area. Human hand skin has specific absorption and reflection characteristics for different wavelengths of light, thus its occlusion produces a unique spectral fingerprint.

[0058] The obtained spectral variation patterns are then compared with pre-stored typical spectral feature patterns representing hand occlusion for similarity calculation. These typical spectral feature patterns can be pre-established using experimental data, machine learning training, or physical models and stored in the system. Similarity calculation can employ various algorithms, such as cosine similarity, Euclidean distance, or neural network-based matching.

[0059] When the calculated similarity reaches a preset threshold, it indicates that the spectral characteristics of the potential occlusion area highly match the typical characteristics of hand occlusion. Only then will the potential occlusion area be ultimately identified as a local occlusion area. Conversely, if the similarity does not reach the threshold, the instantaneous drop in light in the area may be judged as non-hand occlusion, such as being caused by changes in external light sources or non-hand objects.

[0060] This application effectively solves the problem of potential misjudgment in identifying local occlusion areas solely based on a momentary drop in light intensity by introducing spectral analysis and pattern matching mechanisms. Specifically, when local light intensity data momentarily drops in a specific area, the system no longer directly identifies it as local occlusion but first marks it as a potential occlusion area. Subsequently, by conducting in-depth analysis of the light intensity changes of the light sensing unit in this potential occlusion area at multiple specific wavelengths, the unique spectral change pattern of that area can be obtained. Because the human hand has specific absorption and reflection characteristics for different wavelengths of light, the spectral change pattern caused by its occlusion differs significantly from the spectral patterns caused by non-hand occlusion (such as shadows, other objects, or changes in ambient light). By calculating the similarity between the real-time acquired spectral change pattern and pre-stored typical spectral feature patterns representing hand occlusion, and setting a reasonable similarity threshold, the system can accurately distinguish local occlusion caused by the user's hand, thereby avoiding misjudgment caused by light drops due to other factors.

[0061] In some embodiments, when the local light intensity data experiences a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes:

[0062] Acquire image frame sequences of regions where local light intensity data momentarily decreases;

[0063] Motion analysis is performed on the image frame sequence to identify whether there are non-hand movement sources in the image frame sequence. Non-hand movement sources include rapidly changing external light sources or shadow sources that do not conform to the hand movement trajectory.

[0064] Monitor the backlight drive signal of the display screen to determine if there are any instantaneous fluctuations in the backlight drive signal;

[0065] When motion analysis results show the presence of non-hand movement sources or instantaneous fluctuations in the backlight drive signal, areas where local light intensity data momentarily decreases are not marked as potential occlusion areas.

[0066] When motion analysis results show the presence of a non-hand movement source and there are no instantaneous fluctuations in the backlight drive signal, the area where the local light intensity data momentarily decreases is marked as a potential occlusion area.

[0067] Specifically, acquiring a sequence of image frames depicting a region where local light intensity momentarily decreases involves continuously capturing visual information of that specific region using a light-sensing unit integrated on or near the display screen surface, or an additional image sensor, and organizing it into a continuous sequence of image frames. This sequence of image frames is used for subsequent motion analysis to identify the object or phenomenon causing the decrease in light intensity. Motion analysis involves performing image processing and computer vision algorithms on the acquired image frame sequence, such as background subtraction, optical flow, and feature point tracking, to detect movement patterns of pixels or regions in the image. Motion analysis can identify whether there are non-hand movement sources in the image frame sequence. Non-hand movement sources can be understood as any moving object or change in light and shadow that is not caused by the user's hand and results in a decrease in local light intensity. Examples include rapidly changing external light sources (such as flashing headlights or changes in the angle of direct sunlight) or shadows cast by external objects (such as birds or leaves). These non-hand movement sources typically exhibit rapidly changing characteristics that do not correspond to the trajectory of hand movement. Meanwhile, monitoring the backlight drive signal of the display screen refers to obtaining real-time changes in backlight brightness by reading the signals from the backlight control circuit inside the display screen. Instantaneous fluctuations can be understood as unexpected and rapid changes in backlight brightness that occur within a short period of time, uncontrolled by the user, such as flickering caused by power instability or drive circuit failure.

[0068] This application, by introducing motion analysis of image frame sequences and monitoring of backlight drive signals, can effectively distinguish between local light intensity decreases caused by hand occlusion and decreases caused by other factors. Specifically, motion analysis can identify rapid movements of external light sources or shadows. These non-hand movement sources typically have different motion trajectories or change patterns than hands, thus avoiding misjudgment as hand occlusion. Simultaneously, monitoring of the backlight drive signal can eliminate local brightness changes caused by fluctuations in the display's own backlight, ensuring that the possibility of hand occlusion is only considered when the backlight is stable and there is no external interference. Therefore, the solution of this application can more accurately determine the true cause of local light intensity decreases, significantly improving the accuracy of potential occlusion area identification.

[0069] In some embodiments, the step of applying corresponding display parameter adjustments to each display area of ​​the display screen based on the regional illumination state representation to optimize the display effect of each display area includes:

[0070] Acquire spectral data from multiple multi-channel spectral sensing units arranged on the bezel or edge of the touch display screen;

[0071] Parallel preprocessing of spectral data is performed, including noise filtering and spatial gradient feature extraction.

[0072] Based on the parallel preprocessed spectral data, a coarse illumination distribution representation and a fine illumination distribution representation are generated in parallel. The coarse illumination distribution representation is updated at a high frequency, while the fine illumination distribution representation is updated at a slightly lower frequency.

[0073] The fine illumination distribution representation is generated through an adaptive spatial reconstruction method and a dynamic statistical clustering method. The adaptive spatial reconstruction method preserves the boundaries and local differences in the light distribution, while the dynamic statistical clustering method identifies regions with similar light characteristics and spatial proximity based on changes in real-time light data and spatial patterns.

[0074] Based on the coarse and fine illumination distribution representations, corresponding display parameters are adjusted for each display area of ​​the screen. The adjustments include: when a rapid change in the illumination environment is detected, global or large-area display parameter adjustments are applied based on the coarse illumination distribution representation; for areas with fragmented illumination distribution and gentle brightness gradients, customized display parameter adjustments are applied based on the fine illumination distribution representation.

[0075] Among them, the multi-channel spectral sensing unit refers to a sensor array capable of simultaneously sensing light intensity at multiple different wavelengths or bands. These sensing units are arranged on the bezel or edge of the touch display screen, with the aim of obtaining more comprehensive ambient light spectral information from the surrounding environment of the display screen, rather than just single brightness information. By acquiring spectral data, the color composition, color temperature, and characteristics of different light sources of ambient light can be identified more accurately, providing richer and more precise input for subsequent adjustment of display parameters.

[0076] Parallel preprocessing of the spectral data includes noise filtering and spatial gradient feature extraction. Noise filtering aims to remove potential random interference or sensor errors from the raw spectral data, ensuring data purity and reliability. Spatial gradient feature extraction identifies edges, transition regions, and local variations in the illumination distribution, which is crucial for subsequent accurate division of display areas and customized application adjustments. The introduction of parallel processing aims to improve data processing efficiency and ensure the system can respond to changes in ambient light in real time.

[0077] Based on the parallel preprocessed spectral data, coarse and fine illumination distribution representations are generated in parallel. The coarse illumination distribution representation is updated at a high frequency to quickly capture overall or large changes in ambient illumination, suitable for scenarios requiring rapid response. The fine illumination distribution representation is updated at a slightly lower frequency to provide more detailed and accurate illumination distribution details, paying particular attention to boundaries and local differences in the illumination distribution. This combination of coarse and fine representations allows the system to balance response speed and adjustment accuracy.

[0078] The refined illumination distribution representation is generated through an adaptive spatial reconstruction method and a dynamic statistical clustering method. The adaptive spatial reconstruction method aims to reconstruct the illumination distribution of the entire display area based on limited sensor data, with particular emphasis on preserving the boundaries and local differences in the light distribution to avoid excessive smoothing that could lead to loss of detail. The dynamic statistical clustering method, based on changes in real-time light data and spatial patterns, identifies regions with similar light characteristics and spatial proximity, thereby dividing the display screen into display areas with uniform illumination characteristics, providing a basis for subsequent regional adjustments.

[0079] Based on both coarse and fine illumination distribution representations, the system adjusts the display parameters for each display area. This adjustment strategy is layered and adaptive: when a rapid change in the lighting environment is detected, such as when a user moves from indoors to outdoors, the system quickly applies global or large-area adjustments based on the frequently updated coarse illumination distribution representation to rapidly adapt to the new lighting conditions. For areas with fine illumination distribution and gentle brightness gradients, such as under partial indoor shadows or soft lighting, the system applies customized adjustments based on a more detailed fine illumination distribution representation to ensure optimal display performance in these areas and avoid overly bright or dark local areas.

[0080] This application introduces a multi-channel spectral sensing unit to acquire richer and more accurate ambient lighting information than traditional brightness sensors. Parallel preprocessing effectively improves data quality and processing efficiency. More importantly, by generating both coarse and fine-grained representations of the lighting distribution in parallel and combining adaptive spatial reconstruction and dynamic statistical clustering methods, the system can simultaneously address both rapid changes in the lighting environment and precise capture of lighting distribution details. This hierarchical, adaptive adjustment strategy enables the display to provide optimal display performance under various complex and changing lighting conditions, thus overcoming potential accuracy deficiencies and response lag issues when handling complex lighting environments.

[0081] The following is a specific example to illustrate this.

[0082] Imagine a car's central touchscreen display experiencing various lighting conditions while in motion, such as direct sunlight, tunnels, shaded areas, and obstructions from the driver's or passenger's hands. Traditional image enhancement methods, which may only offer simple brightness or contrast adjustments, struggle to handle these complex and dynamically changing scenarios.

[0083] In this application, multiple multi-channel spectral sensing units are arranged around the bezel of the touchscreen display to continuously acquire spectral data of the in-vehicle environment. When the car enters a tunnel, the ambient light intensity drops rapidly, and the spectral sensing units detect this rapid change. At this time, the parallel preprocessing module performs noise filtering and spatial gradient feature extraction on the spectral data, and the coarse illumination distribution representation is updated at a high frequency to quickly reflect the sharp drop in overall illumination. Based on this coarse illumination distribution representation, the display adjustment module immediately applies a global brightness boost adjustment to the entire display screen to ensure that the driver can still clearly see the displayed content after entering the tunnel.

[0084] When a car is driving on a sunny road, and the driver's hand occasionally obscures a portion of the display screen, the spectral sensing unit detects a sudden drop in light intensity in that area, with spectral characteristics similar to those of hand occlusion. Simultaneously, sunlight may enter through side windows, creating fragmented areas of light on the display with gentle brightness gradients. In this situation, the fine illumination distribution representation, through adaptive spatial reconstruction and dynamic statistical clustering methods, accurately identifies the area obscured by the hand, as well as other areas with fragmented illumination distribution. For example, the adaptive spatial reconstruction method accurately delineates the boundaries of the hand's shadow, while the dynamic statistical clustering method categorizes small areas with similar light characteristics. The display adjustment module then uses the fine illumination distribution representation to perform local brightness compensation on the hand-obscured area, while applying customized contrast or gamma adjustments to other areas with fragmented illumination distribution to ensure optimal image detail in these areas and avoid visual impact caused by uneven local illumination. By employing this coarse-to-fine adjustment strategy, the display can quickly respond to changes in overall lighting while accurately compensating for local and subtle differences in lighting, thereby providing superior display performance in various driving scenarios.

[0085] In some embodiments, when the local light intensity data experiences a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes:

[0086] Continuously acquire local light intensity data;

[0087] Determine the current ambient light intensity in a specific area based on local light intensity data;

[0088] Based on the current ambient light intensity, the threshold used to identify a decrease in light intensity is dynamically adjusted. The dynamic adjustment includes: increasing the threshold when the current ambient light intensity is high; and decreasing the threshold when the current ambient light intensity is low.

[0089] When the current local light intensity of a specific area decreases by more than a dynamically adjusted threshold relative to its baseline value, the area is marked as a potential occlusion area.

[0090] Specifically, continuously acquiring local light intensity data refers to continuously collecting light intensity values ​​detected by multiple light sensing units deployed on or near the display screen at a preset frequency. This data is transmitted in real-time to a processing unit for subsequent analysis. Determining the current ambient light intensity of a specific area based on the local light intensity data can be understood as averaging, weighted averaging, or filtering the local light intensity data of that specific area and its surrounding areas to obtain a comprehensive value that represents the current ambient light level. For example, the average intensity of all light sensing units in the area over a period of time can be calculated, or it can be corrected by combining the overall brightness information of the display screen. In practical applications, the threshold for identifying light intensity decreases is dynamically adjusted based on the current ambient light intensity. The purpose is to match the sensitivity of occlusion detection with the ambient light conditions. Specifically, when the current ambient light intensity is high, the light energy in the environment is sufficient, and even a small occlusion can lead to a relatively significant decrease in light intensity. Therefore, to avoid missed detections, the threshold is increased accordingly to ensure that only a sufficiently significant decrease in light intensity is considered a potential occlusion. Conversely, when the ambient light intensity is low, the ambient light energy is weak, and even small fluctuations in ambient light or noise can cause relative changes in light intensity. In this case, to avoid false identification, the threshold is lowered accordingly, so that only significant decreases that are still detectable in low-light conditions are marked. Therefore, when the current local light intensity of a specific area decreases more than the dynamically adjusted threshold relative to its baseline value, that area is reliably marked as a potential occlusion area. The baseline value here can be the average light intensity under unobstructed conditions, or the average light intensity of the area over a period of time.

[0091] This application effectively addresses the limitations of using a fixed threshold to identify potential occlusion areas under varying ambient light conditions by introducing a dynamic threshold adjustment mechanism. The system intelligently adapts to changes in lighting by continuously acquiring and analyzing local light intensity data and dynamically adjusting the identification threshold based on the current ambient light intensity. When ambient light is strong, increasing the threshold effectively filters out relatively small, non-occlusion-related light fluctuations caused by sufficient ambient light, thus reducing false alarms. Conversely, when ambient light is weak, decreasing the threshold increases sensitivity to actual occlusion, preventing inconspicuous occlusion signals due to insufficient ambient light energy, thereby reducing missed alarms. This adaptive threshold adjustment mechanism ensures more accurate and robust judgment of instantaneous drops in local light intensity under various lighting conditions, providing a more reliable foundation for subsequent identification of partially occluded areas.

[0092] In some embodiments, when the local light intensity data experiences a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes:

[0093] Continuously acquire image frame sequences of regions where local light intensity data momentarily decreases;

[0094] Perform image content analysis on image frame sequences to identify whether there are attachment features with specific shapes, textures, or transparency in the image frame sequences;

[0095] Dynamic background subtraction is performed on the image frame sequence to distinguish between stationary objects and moving hands;

[0096] When the image content analysis results show the presence of attachment features, or the dynamic background difference results show that the attachments are stationary, areas where the local light intensity data shows a sudden drop are not marked as potential occlusion areas.

[0097] When the image content analysis results show no attachment features and the dynamic background subtraction results show a moving hand, the area where the local light intensity data momentarily decreases is marked as a potential occlusion area.

[0098] Specifically, continuously acquiring image frame sequences of areas where local light intensity data momentarily decreases refers to continuously capturing visual information of this specific area using a camera or other visual sensor integrated into or near the display screen, forming a series of consecutive image frames. These image frames provide detailed visual context of the light intensity decrease area, providing foundational data for subsequent analysis. Image content analysis of the image frame sequence, identifying the presence of attachments with specific shapes, textures, or transparency, can be understood as using image processing and pattern recognition techniques to analyze the acquired image frames. For example, methods such as edge detection, feature point extraction, texture analysis, or color distribution analysis can be used to identify non-hand objects that may be attached to the display surface, such as stickers, labels, bubbles or stains on screen protectors. These attachments typically have relatively fixed shapes, specific texture patterns, or semi-transparent properties, significantly different from the visual characteristics of hand occlusion. The purpose is to initially screen for light decreases caused by non-hand attachments. In practical applications, dynamic background subtraction is performed on the image frame sequence to distinguish between stationary attachments and moving hands; specifically, this involves detecting movement within the region by comparing pixel changes between consecutive image frames. Dynamic background subtraction effectively separates the foreground (moving objects) from the background (stationary objects or the screen itself) in an image. For example, if an object in a light-dropping area remains stationary across consecutive frames, it is likely an attached object or a stationary external object; if the object exhibits a clear movement trajectory and shape change, it is more likely the user's hand. The aim is to further and more accurately distinguish between static occlusions and dynamic hand occlusions. Furthermore, when image content analysis indicates an attached object feature, or dynamic background subtraction indicates a stationary object, the system determines that the light drop is not caused by hand occlusion and therefore does not mark the area as a potential occlusion region. This avoids misjudgments caused by non-hand factors. Conversely, when image content analysis indicates a non-attached object feature, and dynamic background subtraction indicates a moving object, it strongly suggests that the light drop is caused by the user's hand occlusion, and only then will the area be accurately marked as a potential occlusion region.

[0099] This application employs multi-dimensional verification of regions experiencing a sudden drop in local light intensity by introducing continuous acquisition of image frame sequences, image content analysis, and dynamic background subtraction. Specifically, when the light sensing unit detects a sudden drop in local light intensity, the system no longer simply marks it as a potential occlusion area. Instead, it first acquires an image frame sequence of that area. By analyzing the image content of these frames, it can identify whether there are attachment features with specific shapes, textures, or transparency, thereby eliminating misjudgments caused by screen attachments. Simultaneously, dynamic background subtraction effectively distinguishes between stationary attachments and moving hands, ensuring that the area is only identified as a potential occlusion area when a moving object is detected and that object does not possess attachment features. It is precisely this multi-verification mechanism that enables the system to more accurately determine the true cause of the light drop and avoid misidentifying non-hand occlusion as a potential occlusion area.

[0100] In some embodiments, when the local light intensity data experiences a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes:

[0101] Continuously acquire image frame sequences of regions where local light intensity data momentarily decreases;

[0102] Image content analysis is performed on image frame sequences to identify whether wearable device features with reflective or semi-transparent properties exist in the image frame sequences;

[0103] Based on the wearable device features identified in the image frame sequence, the light reflection or transmission pattern of the image frame sequence at a specific angle is analyzed to obtain the optical influence pattern of the wearable device.

[0104] The similarity between the optical effect pattern and a pre-stored typical pattern representing the optical effect of smart wearable devices is calculated.

[0105] When the similarity reaches a preset threshold, areas where the local light intensity data suddenly decreases will not be marked as potential occlusion areas;

[0106] When the similarity does not reach the preset threshold, the area where the local light intensity data drops instantaneously is marked as a potential occlusion area.

[0107] Specifically, continuously acquiring image frame sequences of areas where local light intensity data momentarily decreases means that when the light sensing unit detects a momentary drop in light intensity in a specific area, the system initiates or continuously captures a continuous stream of image data for that area. These image frame sequences can be acquired by cameras or other visual sensors integrated into the display bezel, top, or bottom to provide visual information about that area.

[0108] Image content analysis of image frame sequences to identify the presence of wearable device features with reflective or semi-transparent properties can be understood as using image processing and computer vision techniques to analyze acquired image frame sequences in order to detect and identify wearable devices such as smartwatches, smart glasses, and wristbands that possess specific optical properties (such as reflectivity, semi-transparency, specific shapes, or textures). This can be achieved using pre-trained deep learning models (such as convolutional neural networks) trained to recognize the visual features of various wearable devices.

[0109] Based on the wearable device features identified in the image frame sequence, light reflection or transmission patterns at specific angles are analyzed to obtain the wearable device's optical influence pattern. This refers to further analyzing how the wearable device affects ambient light and the light emitted from the display screen after identification. For example, for reflective devices, the intensity, direction, and spectral composition of reflected light can be analyzed; for semi-transparent devices, the attenuation and spectral changes of transmitted light can be analyzed. This can be achieved through techniques such as multi-angle light source illumination, polarized light analysis, or spectral analysis to construct the unique optical influence pattern of the wearable device.

[0110] The purpose of similarity calculation is to quantify the degree of match between the currently detected optical effect pattern and a pre-stored set of typical patterns representing the optical effects of smart wearable devices. These pre-stored typical patterns can be obtained through testing and modeling various common smart wearable devices in a laboratory environment, encompassing typical reflection, transmission, or occlusion characteristics of these devices under different lighting conditions. Similarity calculation can employ various distance metrics, such as Euclidean distance, cosine similarity, or feature vector-based matching algorithms.

[0111] When the similarity reaches a preset threshold, areas where local light intensity data momentarily decreases are not marked as potential occlusion areas. This means that if the currently detected decrease in light intensity highly matches the optical influence pattern of the wearable device, the decrease is considered to be caused by the wearable device rather than hand occlusion, and therefore not considered a potential occlusion requiring display adjustment. Conversely, when the similarity does not reach the preset threshold, areas where local light intensity data momentarily decreases are marked as potential occlusion areas. This indicates that the decrease in light intensity is more likely caused by hand occlusion or other non-wearable device factors, and only then are they marked as potential occlusion areas for further processing and display adjustment.

[0112] This application, by introducing visual analysis of image frame sequences, can effectively distinguish between a sudden drop in light intensity caused by hand occlusion and a wearable device with reflective or semi-transparent properties. Specifically, when the light sensing unit detects a localized sudden drop in light intensity, the system no longer simply marks it as a potential occlusion area, but further acquires an image frame sequence of that area. By performing image content analysis on these image frame sequences, the presence of visual features of a wearable device can be identified. Once the wearable device is identified, the system further analyzes its influence pattern on light, such as reflection or transmission characteristics. Subsequently, this optical influence pattern is compared with pre-stored typical wearable device optical influence patterns. If the match is high, it indicates that the drop in light intensity is caused by the wearable device, thus avoiding misidentification as hand occlusion and preventing display parameter adjustments due to misjudgment. Conversely, if the match is low, the drop in light intensity is more likely due to hand occlusion, and only then is it marked as a potential occlusion area, ensuring the accuracy of subsequent processing.

[0113] The following is a specific example to illustrate this.

[0114] Suppose a user is using an in-vehicle touchscreen display and wearing a smartwatch. When the user's arm or the smartwatch moves above the display, the light-sensing units on the display surface may detect a momentary drop in localized light intensity. At this time, the system activates the camera to capture a sequence of image frames of the affected area. The image content analysis module processes these image frames, using a pre-trained neural network model to identify specific shape and texture features of the smartwatch in the image, and that the smartwatch surface has reflective properties. Subsequently, the system analyzes the light reflection pattern of the smartwatch under the current lighting conditions, such as the intensity distribution and spectral characteristics of its reflected light, to obtain the smartwatch's optical effect pattern. This optical effect pattern is then fed into a similarity calculation module and compared with pre-stored typical optical effect patterns of various smartwatches. If the calculated similarity (e.g., cosine similarity) exceeds a preset threshold (e.g., 0.8), the system determines that the localized light intensity drop is caused by the smartwatch. In this case, the area will not be marked as a potential occlusion area, thus avoiding unnecessary brightness or contrast adjustments to the display. Conversely, if the similarity does not reach the threshold, the system will consider the decrease to be more likely caused by the user's hand occlusion and mark it as a potential occlusion area for subsequent display optimization.

[0115] In some embodiments, the step of marking the region where local light intensity data momentarily decreases as a potential occlusion region includes:

[0116] Continuously acquire continuous light intensity values ​​in areas where local light intensity data momentarily decreases;

[0117] Calculate the rate of decrease and recovery of the light intensity value based on the continuous light intensity value;

[0118] When the rate of decrease exceeds a preset rate of decrease threshold and the rate of recovery is lower than a preset rate of recovery threshold, the area where the local light intensity data momentarily decreases is marked as a potential occlusion area.

[0119] When the rate of decrease does not exceed the preset rate of decrease threshold, or the rate of recovery exceeds the preset rate of recovery threshold, the area where the local light intensity data momentarily decreases will not be marked as a potential occlusion area.

[0120] Specifically, continuously acquiring continuous light intensity values ​​in areas where a sudden drop in local light intensity occurs means that after detecting a decrease in light intensity, the system does not immediately make a judgment, but continues to collect a series of continuous light intensity readings within the same area at a certain sampling frequency. These continuous light intensity values ​​can form a time series for subsequent dynamic analysis. The purpose is to capture the complete process of light intensity change, not just a single instantaneous drop point. Calculating the rate of decrease and recovery rate of light intensity values ​​can be understood as performing time derivative or difference operations on the acquired continuous light intensity value series. The rate of decrease refers to the speed at which light intensity changes from its normal level to its lowest point, and can be obtained by calculating the average decrease or maximum slope of decrease per unit time. The rate of recovery refers to the speed at which light intensity recovers from its lowest point to its normal level, and can also be obtained by calculating the average increase or maximum slope of increase per unit time. These rate calculations aim to quantify the dynamic characteristics of light intensity change. In practical applications, preset thresholds for the rate of decrease and the recovery rate are parameters pre-set based on empirical data or experimental results. The drop rate threshold determines whether the light intensity decreases quickly enough to match the typical characteristics of hand occlusion; the recovery rate threshold determines whether the light intensity recovers rapidly after the drop. If the recovery is too fast, it may not be hand occlusion. For example, hand occlusion typically causes a rapid drop in light intensity, which recovers relatively slowly after the hand is removed, or remains at a low level during the occlusion. These thresholds can be calibrated and optimized according to different display types, ambient light conditions, and user habits.

[0121] This application, by introducing analysis of the light intensity decrease and recovery rate, enables more precise identification of potential occlusion areas, effectively solving the problem of misjudgment that may arise from judging solely based on instantaneous light intensity decrease. When a hand covers the display screen, it typically causes a rapid decrease in local light intensity, and the light intensity remains at a low level while the hand remains covering the screen, or recovers relatively slowly as the hand is slowly removed. Conversely, if the light intensity decreases too slowly or recovers rapidly after decreasing, it is more likely caused by non-occlusion events such as brief shadows, ambient light flicker, or a non-hand object quickly passing by. By comparing the actually measured decrease and recovery rates with preset thresholds, the system can distinguish these different types of light change patterns. Specifically, only when the light intensity decreases sufficiently rapidly and recovers insufficiently is it identified as a potential occlusion area, thus avoiding misidentification of non-hand occlusion events.

[0122] The following is a specific example to illustrate this.

[0123] Assuming a touchscreen display is operating normally, light sensing units on its surface continuously monitor ambient light intensity. When a light sensing unit in a specific area detects that the light intensity begins to decrease, the system initiates a continuous sampling process, acquiring continuous light intensity values ​​for that area at a frequency of 100 times per second for, for example, 2 seconds. The system analyzes the light intensity data within these 2 seconds. For example, if the light intensity drops from 500 lux to 50 lux in 0.1 seconds, the rate of decrease is 4500 lux / second. A preset rate of decrease threshold might be set to 3000 lux / second. If, after reaching its lowest point, the light intensity only recovers to 100 lux in the following 1.9 seconds, the recovery rate is 26 lux / second. A preset recovery rate threshold might be set to 100 lux / second. In this scenario, because the descent rate (4500 lux / s) exceeds the descent rate threshold (3000 lux / s) and the recovery rate (26 lux / s) is below the recovery rate threshold (100 lux / s), the system marks the area as a potential occlusion. This indicates a rapid decrease in light intensity followed by a slow recovery, consistent with typical hand occlusion characteristics. Conversely, if the light intensity recovers rapidly to 400 lux within 0.2 seconds after the initial decrease, with a recovery rate of 1750 lux / s, this recovery rate (1750 lux / s) exceeds the recovery rate threshold (100 lux / s). Even if the descent rate meets the criteria, the system will not mark the area as a potential occlusion, as it is more likely a transient shadow or light flicker. In this way, the system effectively filters out transient light changes not caused by hand occlusion, ensuring that only genuine user hand occlusion is further processed.

[0124] In some embodiments, the step of marking the region where the local light intensity data momentarily decreases as a potential occlusion region may specifically include:

[0125] Continuously acquire image frame sequences of regions where local light intensity data momentarily decreases;

[0126] Human body part recognition is performed on image frame sequences to identify whether hand features exist in the image frame sequences;

[0127] When hand features are identified, the system determines whether the hand belongs to the driver or the co-driver based on the spatial position of the hand features on the display screen and the movement trajectory of the hand features.

[0128] When it is determined that the hand is obstructing the light, the area where the local light intensity data drops instantaneously is marked as the potential obstruction area;

[0129] When it is determined that the hand is obstructed by a non-driver's hand, areas where the local light intensity data shows a momentary drop are not marked as potential obstruction areas.

[0130] Specifically, continuously acquiring image frame sequences of areas where local light intensity momentarily decreases refers to continuously capturing image or video data of a specific area experiencing a drop in light intensity using a camera or other image sensor integrated near or inside the display screen. These image frame sequences are used for subsequent visual analysis to identify and determine the specific nature of any obstructions.

[0131] Human body part recognition in image frame sequences, specifically identifying the presence of hand features, can be understood as using computer vision techniques and machine learning models to analyze the acquired image frame sequences. This analysis aims to detect whether features in the image match a pre-defined hand model, such as the contours, textures, and movement patterns of the palm and fingers. Its purpose is to preliminarily confirm whether the occluded object is a human hand.

[0132] When hand features are detected, the system determines whether the hand belongs to the driver or the passenger based on its spatial position on the display screen and its movement trajectory. This determination can be made by combining information about the vehicle's interior layout, pre-defined driver and passenger areas, and typical hand operation areas and movement paths on the display screen. For example, the left side of the display screen can be pre-defined as primarily for driver operations, and the right side as primarily for passenger operations. Simultaneously, by tracking the hand's movement trajectory, the system determines whether it matches the driver's habitual actions while driving or represents a specific interaction by the passenger. The purpose is to differentiate hand occlusion for different user roles, allowing for differentiated processing.

[0133] When the system determines that the hand is obstructing the view due to the driver's hand, areas where the local light intensity data momentarily drops are marked as potential obstruction areas. This means the system considers this obstruction to have a direct impact on the driver's display experience, requiring further evaluation and potentially triggering display parameter adjustments. Conversely, when the system determines that the hand is obstructing the view due to a non-driver's hand, areas where the local light intensity data momentarily drops are not marked as potential obstruction areas. The purpose is to avoid unnecessary display adjustments for temporary obstructions by non-critical users (such as the front passenger), thereby improving the system's intelligence and the accuracy of the user experience.

[0134] This application addresses the potential over-response issue in multi-user environments by introducing user role recognition for occluded hands. Specifically, when the system detects a sudden drop in localized light intensity on the display screen, it no longer simply marks the area as a potential occlusion region and proceeds with further processing. Instead, it first identifies human body parts through image frame sequences to confirm whether the occlusion is a hand. Once confirmed, the system further analyzes the hand's spatial position and movement trajectory, combining this with pre-defined user area information to determine whether the hand belongs to the driver or the co-driver. Only when the occlusion is determined to be the driver's hand is the area marked as a potential occlusion region, thus initiating the subsequent image enhancement process. For occlusions by non-driver hands, they are not marked as potential occlusion regions, avoiding unnecessary adjustments to display parameters. This mechanism ensures that display enhancement is only activated when it actually impacts the visual experience of the primary user (driver), thereby improving the system's decision-making efficiency and the relevance of the user experience.

[0135] In a vehicle-mounted touchscreen display system, the screen is divided into a driver's operating area and a passenger's operating area. While the vehicle is in motion, the passenger's hand accidentally obscures an area on the right side of the screen, causing the light sensing unit in that area to detect a momentary drop in light intensity. The system continuously acquires image frame sequences of that area and identifies the presence of a hand feature using a human body recognition module. Further, based on the hand's spatial position on the right side of the screen and its movement trajectory (e.g., a trajectory that does not conform to the driver's typical operating pattern), the system determines that the hand belongs to the passenger. Since this obstruction is determined to be non-driver's hand obstruction, the system will not mark this area as a potential obstruction area, and therefore will not trigger subsequent display parameter adjustments. Simultaneously, if the driver's hand obscures the navigation information area on the left side of the screen, the system will also recognize the hand feature, but based on its spatial position in the driver's operating area and its movement trajectory that conforms to the driver's operating habits, it will determine that it is the driver's hand obstructing the area. In this case, the area will be marked as a potential obstruction area, and subsequent image enhancement processes will be initiated to optimize the driver's visual experience.

[0136] like Figure 2 As shown in the embodiments, this application also discloses a touch screen image enhancement system, including:

[0137] The data acquisition module 1 is used to acquire local light intensity data from multiple light sensing units arranged on or near the display screen surface, as well as interaction event information from the user interaction sensing unit.

[0138] The area recognition module 2 is used to identify areas of continuous illumination difference and local occlusion on the display screen based on local light intensity data and interactive event information.

[0139] The state representation construction module 3 is used to construct the regional illumination state representation of the display screen based on the recognition results. The regional illumination state representation divides the display screen into multiple display areas and includes the light state information of each display area.

[0140] The display adjustment module 4 is used to adjust the corresponding display parameters of each display area of ​​the display screen according to the regional illumination status, so as to optimize the display effect of each display area.

[0141] The system provided in this application is capable of intelligently enhancing images on touch screens.

[0142] The above are merely embodiments of this application and are 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 touch display screen, characterized in that, include: Acquire local light intensity data from multiple light sensing units arranged on or near the display screen surface, as well as interaction event information from user interaction sensing units; Based on local light intensity data and interactive event information, identify areas of persistent illumination difference and local occlusion on the display screen; Based on the recognition results, a regional illumination state representation of the display screen is constructed. The regional illumination state representation divides the display screen into multiple display areas and includes the light state information of each display area. Based on the regional illumination status, the corresponding display parameters are adjusted for each display area of ​​the screen to optimize the display effect of each display area.

2. The touch screen image enhancement method according to claim 1, characterized in that, The step of identifying areas of persistent illumination difference and areas of partial occlusion on the display screen based on local light intensity data and interactive event information includes: When the local light intensity data of multiple light sensing units continuously differ above a preset threshold, the region of continuous illumination difference is identified. When local light intensity data shows a momentary drop in a specific area, that area is marked as a potential occlusion area; For potential occlusion areas, the light intensity changes of light sensing units in these areas at multiple specific wavelengths are analyzed to obtain the spectral variation patterns of potential occlusion areas. The similarity between the spectral variation pattern and the pre-stored typical spectral feature patterns representing hand occlusion is calculated. When the similarity reaches a preset threshold, the potential occlusion area is identified as a local occlusion area.

3. The touch screen image enhancement method according to claim 2, characterized in that, When the local light intensity data shows a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes: Acquire image frame sequences of regions where local light intensity data momentarily decreases; Motion analysis is performed on the image frame sequence to identify whether there are non-hand movement sources in the image frame sequence. Non-hand movement sources include rapidly changing external light sources or shadow sources that do not conform to the hand movement trajectory. Monitor the backlight drive signal of the display screen to determine if there are any instantaneous fluctuations in the backlight drive signal; When motion analysis results show the presence of non-hand movement sources or instantaneous fluctuations in the backlight drive signal, areas where local light intensity data momentarily decreases are not marked as potential occlusion areas. When motion analysis results show the presence of a non-hand movement source and there are no instantaneous fluctuations in the backlight drive signal, the area where the local light intensity data momentarily decreases is marked as a potential occlusion area.

4. The touch screen image enhancement method according to claim 1, characterized in that, The step of adjusting the corresponding display parameters for each display area of ​​the screen based on the regional illumination state to optimize the display effect of each display area includes: Acquire spectral data from multiple multi-channel spectral sensing units arranged on the bezel or edge of the touch display screen; Parallel preprocessing of spectral data is performed, including noise filtering and spatial gradient feature extraction. Based on the parallel preprocessed spectral data, coarse and fine illumination distribution representations are generated in parallel. Based on the coarse and fine illumination distribution representations, corresponding display parameters are adjusted for each display area of ​​the screen. The adjustments include: applying global or large-area display parameter adjustments based on the coarse illumination distribution representation; and applying customized display parameter adjustments based on the fine illumination distribution representation for areas with fragmented illumination distribution and gentle brightness gradients.

5. The touch screen image enhancement method according to claim 2, characterized in that, When the local light intensity data shows a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes: Continuously acquire local light intensity data; Determine the current ambient light intensity in a specific area based on local light intensity data; The threshold for identifying a decrease in light intensity is dynamically adjusted based on the current ambient light intensity. When the current local light intensity of a specific area decreases by more than a dynamically adjusted threshold relative to its baseline value, the area is marked as a potential occlusion area.

6. The touch display screen image enhancement method according to claim 2, characterized in that, When the local light intensity data shows a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes: Continuously acquire image frame sequences of regions where local light intensity data momentarily decreases; Perform image content analysis on image frame sequences to identify whether there are attachment features with specific shapes, textures, or transparency in the image frame sequences; Dynamic background subtraction is performed on the image frame sequence to distinguish between stationary objects and moving hands; When the image content analysis results show the presence of attachment features, or the dynamic background difference results show that the attachments are stationary, areas where the local light intensity data shows a sudden drop are not marked as potential occlusion areas. When the image content analysis results show no attachment features and the dynamic background subtraction results show a moving hand, the area where the local light intensity data momentarily decreases is marked as a potential occlusion area.

7. The touch display screen image enhancement method according to claim 2, characterized in that, When the local light intensity data shows a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes: Continuously acquire image frame sequences of regions where local light intensity data momentarily decreases; Image content analysis is performed on image frame sequences to identify whether wearable device features with reflective or semi-transparent properties exist in the image frame sequences; Based on the wearable device features identified in the image frame sequence, the light reflection or transmission pattern of the image frame sequence at a specific angle is analyzed to obtain the optical influence pattern of the wearable device. The similarity between the optical effect pattern and a pre-stored typical pattern representing the optical effect of smart wearable devices is calculated. When the similarity reaches a preset threshold, areas where the local light intensity data suddenly decreases will not be marked as potential occlusion areas; When the similarity does not reach the preset threshold, the area where the local light intensity data drops instantaneously is marked as a potential occlusion area.

8. The touch screen image enhancement method according to claim 2, characterized in that, When the local light intensity data shows a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes: Continuously acquire continuous light intensity values ​​in areas where local light intensity data momentarily decreases; Calculate the rate of decrease and recovery of the light intensity value based on the continuous light intensity value; When the rate of decrease exceeds a preset rate of decrease threshold and the rate of recovery is lower than a preset rate of recovery threshold, the area where the local light intensity data momentarily decreases is marked as a potential occlusion area. When the rate of decrease does not exceed the preset rate of decrease threshold, or the rate of recovery exceeds the preset rate of recovery threshold, the area where the local light intensity data momentarily decreases will not be marked as a potential occlusion area.

9. A touch screen image enhancement method according to claim 2, characterized in that, When the local light intensity data shows a momentary drop in a specific area, the step of marking that area as a potential occlusion area includes: Continuously acquire image frame sequences of regions where local light intensity data momentarily decreases; Human body part recognition is performed on image frame sequences to identify whether hand features exist in the image frame sequences; When hand features are identified, the system determines whether the hand belongs to the driver or the co-driver based on the spatial position of the hand features on the display screen and the movement trajectory of the hand features. When it is determined that the hand is obstructing the light, the area where the local light intensity data drops instantaneously is marked as the potential obstruction area; When it is determined that the hand is obstructed by a non-driver's hand, areas where the local light intensity data shows a momentary drop are not marked as potential obstruction areas.

10. A touch screen image enhancement system, characterized in that, include: The data acquisition module is used to acquire local light intensity data from multiple light sensing units arranged on or near the display screen surface, as well as interaction event information from the user interaction sensing unit; The area recognition module is used to identify areas of persistent illumination difference and partial occlusion on the display screen based on local light intensity data and interactive event information. The state representation construction module is used to construct the regional illumination state representation of the display screen based on the recognition results. The regional illumination state representation divides the display screen into multiple display areas and includes the light state information of each display area. The display adjustment module is used to adjust the corresponding display parameters of each display area of ​​the screen according to the regional lighting status, so as to optimize the display effect of each display area.

Images