Image segmentation method and system for liquid crystal screen display

Abstract

The invention relates to the technical field of machine vision, in particular to an image segmentation method and system for liquid crystal screen display. The method comprises the following steps: acquiring basic data of the liquid crystal screen; performing screen response characteristic analysis according to the basic data of the liquid crystal screen to obtain screen response characteristic data; performing multi-scale image acquisition according to the basic data of the liquid crystal screen to generate multi-scale image data; performing characteristic classification on the liquid crystal display according to the screen response characteristic data, and performing edge compensation analysis on the multi-scale image data to obtain edge characteristic compensation data; acquiring real-time data of the touch screen; and performing touch interference identification analysis on the real-time data of the touch screen according to the edge feature compensation data to generate touch interference mask data. By analyzing the screen response characteristics and performing edge compensation analysis, the phenomenon of edge blurring caused by liquid crystal response time limitation is reduced, and the accuracy of edge detection is improved.

Description

Technical Field

[0001] The present invention relates to the field of machine vision technology, and in particular to an image segmentation method and system for liquid crystal display. Background Art

[0002] With the improvement of the performance of LCD display devices, especially the popularity of high-resolution, large-size screens and high-refresh-rate screens, accurately and efficiently processing and segmenting screen display content has become an important technical challenge. Image segmentation of LCD display refers to the process of analyzing and processing images on the LCD screen to extract the target area in the image from the background or other irrelevant information. This technology plays a vital role in computer vision, augmented reality, touch screen devices, and smart device displays. However, traditional image segmentation methods for LCD display often have the following problems: when displaying fast-moving objects, due to the response time limit of the liquid crystal, the edges may be blurred, affecting the effect of the segmentation method based on edge detection; it is necessary to consider the physical properties of the LCD screen and develop a segmentation algorithm that can handle dynamic scenes. On touch screen devices, the user's touch operation may interfere with the image segmentation process. Especially for the method based on region growing, it is necessary to design a segmentation algorithm that can distinguish and filter out touch interference. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide an image segmentation method and system for liquid crystal display to solve at least one of the above technical problems.

[0004] To achieve the above object, an image segmentation method for liquid crystal display comprises the following steps:

[0005] Step S1: acquiring basic data of the LCD screen; performing screen response characteristic analysis according to the basic data of the LCD screen to obtain screen response characteristic data; performing multi-scale image acquisition according to the basic data of the LCD screen to generate multi-scale image data; classifying the characteristics of the LCD screen according to the screen response characteristic data, and performing edge compensation analysis on the multi-scale image data respectively to obtain edge feature compensation data;

[0006] Step S2: acquiring real-time data of the touch screen; performing touch interference recognition analysis on the real-time data of the touch screen according to the edge feature compensation data, and generating touch interference mask data, wherein the touch interference includes pressure deformation interference, electric field interference, refresh rate interference and timing interference;

[0007] Step S3: performing edge feature correction processing on the edge feature compensation data according to the touch interference mask data, and performing adaptive threshold segmentation to obtain preliminary image segmentation data; performing temporal consistency constraint processing on the preliminary image segmentation data according to the screen response characteristic data, thereby generating segmentation strategy data;

[0008] Step S4: real-time image data acquisition is performed on the LCD screen, and moving target detection is performed based on the real-time image data to obtain moving target data; region growing processing is performed based on the segmentation strategy data and the moving target data to obtain preliminary segmentation region data;

[0009] Step S5: Perform edge refinement processing on the preliminary segmented area data, evaluate the segmentation quality, and optimize the segmentation results to obtain the final segmentation result data.

[0010] The present invention obtains the basic data of the liquid crystal screen to perform response characteristic analysis, and generates data in combination with multi-scale image acquisition. Through the response characteristic analysis, the dynamic behavior of the screen can be deeply understood, including its refresh rate, response time, etc., which lays the foundation for subsequent edge feature compensation. Multi-scale image acquisition can capture the performance of the liquid crystal screen at different resolutions, which helps to analyze and deal with edge blur issues more comprehensively, especially when dealing with complex image structures. The acquisition and interference identification of real-time data of the touch screen are key steps to ensure the quality of image segmentation; by generating a touch interference mask, common interferences such as pressure deformation and electric field interference can be effectively identified and eliminated, which greatly improves the accuracy of image segmentation. Especially in the scenario of high refresh rate and complex touch operation, the influence of timing interference on image segmentation may be great, so this step helps to shield external interference and ensure the authenticity and consistency of data. Based on the touch interference mask data, the edge features are corrected and adaptive threshold segmentation is performed, which can make image segmentation more accurate and reduce the influence of edge blur. Through the temporal consistency constraint, the segmentation strategy can not only consider the spatial features, but also combine the time response behavior of the LCD screen to ensure the consistency of image segmentation in continuous frames and avoid image drift or edge distortion caused by screen response time problems. By real-time acquisition of LCD screen image data and combined with moving target detection, static and dynamic objects can be effectively distinguished. The addition of moving target data makes image segmentation more dynamically adaptable, especially suitable for scenes where moving pictures are displayed on LCD screens. The region growing process makes the segmented area more natural and smooth, reducing the segmentation discontinuity caused by screen characteristics or touch interference. The final edge refinement process is the key link to ensure the high quality of the segmentation results. By refining the preliminary segmented area, not only the clarity of the segmentation edge is improved, but also artifacts and noise can be effectively reduced. Through quality evaluation and optimization, the final segmentation result is more accurate, especially when dealing with complex scenes and high dynamic pictures, the optimized segmentation result is more in line with the actual application requirements. Through multi-level analysis and processing, this method combines the LCD screen response characteristics with touch interference recognition and image segmentation, effectively improving the accuracy and stability of segmentation. It is particularly suitable for LCD screen image segmentation in complex scenes and dynamic pictures. This not only reduces the segmentation deviation caused by the LCD screen characteristics, but also improves the segmentation robustness under touch operations.

[0011] The present invention also provides an image segmentation system for liquid crystal display, which is used to execute the image segmentation method for liquid crystal display as described above. The image segmentation system for liquid crystal display includes:

[0012] The screen characteristic analysis module is used to obtain basic data of the LCD screen; perform screen response characteristic analysis based on the basic data of the LCD screen to obtain screen response characteristic data; perform multi-scale image acquisition based on the basic data of the LCD screen to generate multi-scale image data; perform characteristic classification of the LCD screen based on the screen response characteristic data, and perform edge compensation analysis on the multi-scale image data respectively to obtain edge feature compensation data;

[0013] A touch interference recognition module is used to obtain real-time data of the touch screen; perform touch interference recognition analysis on the real-time data of the touch screen according to the edge feature compensation data, and generate touch interference mask data, wherein the touch interference includes pressure deformation interference, electric field interference, refresh rate interference and timing interference;

[0014] A preliminary segmentation module is used to perform edge feature correction processing on edge feature compensation data according to touch interference mask data, and perform adaptive threshold segmentation to obtain preliminary image segmentation data; perform temporal consistency constraint processing on the preliminary image segmentation data according to screen response characteristic data, thereby generating segmentation strategy data;

[0015] The moving target detection module is used to collect real-time image data of the LCD screen, and perform moving target detection based on the real-time image data to obtain moving target data; perform region growing processing based on the segmentation strategy data and the moving target data to obtain preliminary segmentation region data;

[0016] The edge refinement module is used to perform edge refinement processing on the preliminary segmentation area data, evaluate the segmentation quality, and optimize the segmentation results to obtain the final segmentation result data.

[0017] The present invention obtains basic data of liquid crystal screen, and performs screen response characteristic analysis, multi-scale image acquisition, characteristic classification and edge compensation analysis. The acquisition of basic data provides the basis for subsequent analysis and ensures the accuracy of analysis; by analyzing the screen response characteristics, the performance indicators of the liquid crystal screen, such as response time and contrast, can be evaluated, which is helpful to identify the pros and cons of the display effect; image data of different scales are generated, which is convenient for subsequent detailed analysis and processing, especially for processing targets of different sizes and detail levels; by classifying the screen characteristics, a suitable compensation method can be selected to improve the edge processing effect, improve the final image quality, and reduce the distortion caused by the difference in screen characteristics. Real-time data of the touch screen is obtained, touch interference recognition analysis is performed, and touch interference mask data is generated. The touch screen state is monitored in time, and any potential interference source can be effectively identified; multiple interferences such as pressure deformation, electric field, refresh rate and timing influence can be identified, and the interference type can be accurately located and classified, which is helpful for subsequent compensation and adjustment; touch interference mask data is generated to provide a basis for subsequent processing, so that the touch control system can filter interference more accurately and improve the touch control experience. The edge feature compensation data is corrected, adaptive threshold segmentation is performed, and segmentation strategy data is generated. By correcting the edges, the accuracy of the segmented area is improved and the segmentation errors are reduced; the threshold is dynamically adjusted according to the image content to achieve more accurate segmentation processing, so that the segmentation effect between different backgrounds and foregrounds is better; the temporal consistency of the segmentation results is ensured, especially when dealing with dynamic scenes, the errors are reduced and the segmentation stability is improved. Real-time image data acquisition and moving target detection, and regional growth processing are performed. Through real-time image acquisition, moving targets can be quickly identified and tracked to adapt to rapidly changing scenes; the ability to accurately detect moving targets enables the system to respond quickly and is suitable for a variety of application scenarios, such as security monitoring and automated detection; using regional growth technology, the target area can be effectively extracted to ensure the integrity and accuracy of the moving target and optimize the segmentation results. The edge refinement and segmentation quality evaluation of the preliminary segmented area are performed to optimize the segmentation results. Improve the clarity of the edge of the segmented area, ensure the accuracy of the target shape, and reduce blur and irregularity; by evaluating the segmentation quality, unqualified areas can be quickly identified, guiding subsequent optimization, and improving the overall segmentation quality; the optimized segmentation results are more in line with actual application needs and improve the reliability and accuracy of the system. These modules work together to improve the performance of LCD and touch screen systems through efficient data acquisition, interference identification, feature analysis and dynamic processing. The system design ensures that each link can respond to changes accurately and promptly, ultimately achieving high-quality image segmentation and accurate moving target detection, suitable for a variety of practical application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0019] Figure 1 A schematic diagram of the steps of the image segmentation method for liquid crystal display display of the present invention;

[0020] Figure 2 for Figure 1 Detailed step flow diagram of step S1;

[0021] Figure 3 for Figure 1 Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0022] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0023] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0024] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0025] To achieve this, please refer to Figures 1 to 3 The present invention provides an image segmentation method for liquid crystal display, the method comprising the following steps:

[0026] Step S1: acquiring basic data of the LCD screen; performing screen response characteristic analysis according to the basic data of the LCD screen to obtain screen response characteristic data; performing multi-scale image acquisition according to the basic data of the LCD screen to generate multi-scale image data; classifying the characteristics of the LCD screen according to the screen response characteristic data, and performing edge compensation analysis on the multi-scale image data respectively to obtain edge feature compensation data;

[0027] Step S2: acquiring real-time data of the touch screen; performing touch interference recognition analysis on the real-time data of the touch screen according to the edge feature compensation data, and generating touch interference mask data, wherein the touch interference includes pressure deformation interference, electric field interference, refresh rate interference and timing interference;

[0028] Step S3: performing edge feature correction processing on the edge feature compensation data according to the touch interference mask data, and performing adaptive threshold segmentation to obtain preliminary image segmentation data; performing temporal consistency constraint processing on the preliminary image segmentation data according to the screen response characteristic data, thereby generating segmentation strategy data;

[0029] Step S4: real-time image data acquisition is performed on the LCD screen, and moving target detection is performed based on the real-time image data to obtain moving target data; region growing processing is performed based on the segmentation strategy data and the moving target data to obtain preliminary segmentation region data;

[0030] Step S5: Perform edge refinement processing on the preliminary segmented area data, evaluate the segmentation quality, and optimize the segmentation results to obtain the final segmentation result data.

[0031] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of an image segmentation method for liquid crystal display in the present invention. In this example, the image segmentation method for liquid crystal display includes the following steps:

[0032] Step S1: acquiring basic data of the LCD screen; performing screen response characteristic analysis according to the basic data of the LCD screen to obtain screen response characteristic data; performing multi-scale image acquisition according to the basic data of the LCD screen to generate multi-scale image data; classifying the characteristics of the LCD screen according to the screen response characteristic data, and performing edge compensation analysis on the multi-scale image data respectively to obtain edge feature compensation data;

[0033] The embodiment of the present invention collects basic data, including screen size (e.g., 65 inches), resolution (e.g., 1920x1080), and pixel density (e.g., 100PPI), through a sensor module connected to the LCD screen. These data are standardized through a preprocessing algorithm to ensure consistency. Next, based on the specifications of the LCD screen, the system runs a screen response characteristic analysis algorithm, and uses historical data and experimental results to generate response characteristic data such as response time (e.g., 5ms) and dynamic contrast (e.g., 1000:1). At the same time, the system sets a multi-scale image acquisition scheme, and by adjusting the lens focal length and exposure time, captures images of different resolutions (e.g., 480p, 720p, and 1080p) in layers, and stores them as multi-scale image data. Finally, the system uses the response characteristic data to classify the characteristics of the LCD screen, uses a decision tree algorithm to analyze each classification feature, and uses edge detection technology (e.g., Sobel algorithm) to perform edge compensation on the multi-scale image data, and finally outputs edge feature compensation data.

[0034] Step S2: acquiring real-time data of the touch screen; performing touch interference recognition analysis on the real-time data of the touch screen according to the edge feature compensation data, and generating touch interference mask data, wherein the touch interference includes pressure deformation interference, electric field interference, refresh rate interference and timing interference;

[0035] In the embodiment of the present invention, the touch screen sensor collects the touch coordinates, pressure value (e.g., 5N) and touch area (e.g., 10cm 2 ) to form real-time data of the touch screen. Then, the system uses edge feature compensation data to implement a touch interference recognition algorithm. The specific operation includes comparing the real-time data with the preset interference model to identify pressure deformation (such as false touches caused by screen bending), electric field interference (such as static electricity), refresh rate effects (such as blur caused by low refresh rates) and timing effects (such as delays). After recognition, the system generates touch interference mask data, marks the interference area and outputs it for subsequent processing.

[0036] Step S3: performing edge feature correction processing on the edge feature compensation data according to the touch interference mask data, and performing adaptive threshold segmentation to obtain preliminary image segmentation data; performing temporal consistency constraint processing on the preliminary image segmentation data according to the screen response characteristic data, thereby generating segmentation strategy data;

[0037] In an embodiment of the present invention, the system performs edge feature correction processing on the edge feature compensation data based on the touch interference mask data. Specifically, the system uses a region-based edge adjustment algorithm to perform local correction on the disturbed area, and uses an interpolation method to fill in the missing edge data. After the correction is completed, the system applies an adaptive threshold segmentation algorithm to automatically determine the segmentation threshold based on the local brightness and contrast of the image to obtain preliminary image segmentation data. Next, the system uses temporal consistency constraint processing and a Kalman filter to smooth the preliminary image segmentation data to ensure segmentation consistency in dynamic scenes, and finally generates segmentation strategy data.

[0038] Step S4: real-time image data acquisition is performed on the LCD screen, and moving target detection is performed based on the real-time image data to obtain moving target data; region growing processing is performed based on the segmentation strategy data and the moving target data to obtain preliminary segmentation region data;

[0039] The embodiment of the present invention performs real-time image data acquisition on the LCD screen, adopts high-frequency (e.g., 60fps) sampling to capture dynamic display content. Subsequently, the system applies a foreground extraction algorithm, and uses optical flow analysis and frame difference technology to identify and extract moving target data. The specific operation includes calculating the pixel changes between each frame, finding the moving area, and marking it. Then, based on the segmentation strategy data and the moving target data, the system implements regional growth processing, by setting a seed point (e.g., the center position of the moving target) and using a regional similarity metric, gradually expanding the identification area, and obtaining preliminary segmentation area data to ensure accurate capture of the target boundary.

[0040] Step S5: Perform edge refinement processing on the preliminary segmented area data, evaluate the segmentation quality, and optimize the segmentation results to obtain the final segmentation result data.

[0041] When the embodiment of the present invention performs edge refinement processing on the preliminary segmented area data, the system uses a high-precision edge detection algorithm, such as the Canny algorithm, to extract the edge information of the segmented area. In order to improve the accuracy, the system will perform sub-pixel refinement processing on the edge data and use quadratic interpolation to fine-tune the edge position. Next, the system analyzes the accuracy of the segmentation results through a segmentation quality assessment algorithm, and identifies areas whose quality is lower than a preset threshold (such as areas with blurred edges) and marks them as areas to be optimized. Finally, the system applies the optimization strategy of local threshold adjustment and regional merging and splitting operations to reprocess the area to be optimized, thereby obtaining the final segmentation result data to ensure the accuracy and completeness of the segmentation.

[0042] The present invention obtains the basic data of the liquid crystal screen to perform response characteristic analysis, and generates data in combination with multi-scale image acquisition. Through the response characteristic analysis, the dynamic behavior of the screen can be deeply understood, including its refresh rate, response time, etc., which lays the foundation for subsequent edge feature compensation. Multi-scale image acquisition can capture the performance of the liquid crystal screen at different resolutions, which helps to analyze and deal with edge blur issues more comprehensively, especially when dealing with complex image structures. The acquisition and interference identification of real-time data of the touch screen are key steps to ensure the quality of image segmentation; by generating a touch interference mask, common interferences such as pressure deformation and electric field interference can be effectively identified and eliminated, which greatly improves the accuracy of image segmentation. Especially in the scenario of high refresh rate and complex touch operation, the influence of timing interference on image segmentation may be great, so this step helps to shield external interference and ensure the authenticity and consistency of data. Based on the touch interference mask data, the edge features are corrected and adaptive threshold segmentation is performed, which can make image segmentation more accurate and reduce the influence of edge blur. Through the temporal consistency constraint, the segmentation strategy can not only consider the spatial features, but also combine the time response behavior of the LCD screen to ensure the consistency of image segmentation in continuous frames and avoid image drift or edge distortion caused by screen response time problems. By real-time acquisition of LCD screen image data and combined with moving target detection, static and dynamic objects can be effectively distinguished. The addition of moving target data makes image segmentation more dynamically adaptable, especially suitable for scenes where moving pictures are displayed on LCD screens. The region growing process makes the segmented area more natural and smooth, reducing the segmentation discontinuity caused by screen characteristics or touch interference. The final edge refinement process is the key link to ensure the high quality of the segmentation results. By refining the preliminary segmented area, not only the clarity of the segmentation edge is improved, but also artifacts and noise can be effectively reduced. Through quality evaluation and optimization, the final segmentation result is more accurate, especially when dealing with complex scenes and high dynamic pictures, the optimized segmentation result is more in line with the actual application requirements. Through multi-level analysis and processing, this method combines the LCD screen response characteristics with touch interference recognition and image segmentation, effectively improving the accuracy and stability of segmentation. It is particularly suitable for LCD screen image segmentation in complex scenes and dynamic pictures. This not only reduces the segmentation deviation caused by the LCD screen characteristics, but also improves the segmentation robustness under touch operations.

[0043] Preferably, step S1 comprises the following steps:

[0044] Step S11: collecting information on the size, resolution and pixel density of the LCD screen, thereby obtaining screen physical property data;

[0045] Step S12: acquiring electrical parameters of the liquid crystal display screen based on the driving voltage and the refresh rate, thereby obtaining screen electrical characteristic data;

[0046] Step S13: merging the screen electrical characteristic data and the screen physical characteristic data into LCD screen basic data;

[0047] Step S14: analyzing the screen response characteristics according to the basic data of the liquid crystal screen, thereby obtaining the screen response characteristic data;

[0048] Step S15: designing a multi-scale sampling scheme according to the screen physical characteristic data, and using the sampling parameter data to perform multi-scale image acquisition on the LCD display content to generate multi-scale image data;

[0049] Step S16: classifying the characteristics of the liquid crystal screen according to the screen response characteristic data, and performing dynamic edge compensation analysis on the multi-scale image data to obtain edge feature compensation data.

[0050] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in the embodiment of the present invention, step S1 includes the following steps:

[0051] Step S11: collecting information on the size, resolution and pixel density of the LCD screen, thereby obtaining screen physical property data;

[0052] In the embodiment of the present invention, the system collects the physical parameters of the LCD screen, including size (e.g., 65 inches), resolution (e.g., 1920x1080 pixels), and pixel density (e.g., 100PPI) through sensors and measuring devices connected to the LCD screen. These data are standardized through the data processing module to ensure format consistency and prepare for subsequent analysis. The system records this information as screen physical property data, forming the first part of the basic data set.

[0053] Step S12: acquiring electrical parameters of the liquid crystal display screen based on the driving voltage and the refresh rate, thereby obtaining screen electrical characteristic data;

[0054] The embodiment of the present invention obtains the electrical parameters of the LCD screen, especially the driving voltage (such as 5V) and the refresh rate (such as 60Hz) through electrical measuring instruments. Using an oscilloscope and a voltage probe, the system monitors the electrical signals of the LCD screen in real time and records the electrical characteristics under different operating conditions. Subsequently, the system analyzes these data and extracts the electrical characteristic data as another part of the basic data of the LCD screen.

[0055] Step S13: merging the screen electrical characteristic data and the screen physical characteristic data into LCD screen basic data;

[0056] The embodiment of the present invention combines the screen physical characteristic data and electrical characteristic data obtained in step S11 and step S12 to form a complete LCD screen basic data set. The specific operation is to use a data fusion algorithm to match the two sets of data by attributes and store them in a database, ensuring that all characteristics of each screen can be referenced in subsequent analysis, providing a comprehensive information basis for screen response characteristic analysis.

[0057] Step S14: analyzing the screen response characteristics according to the basic data of the liquid crystal screen, thereby obtaining the screen response characteristic data;

[0058] In the embodiment of the present invention, the system performs screen response characteristic analysis based on the merged basic data of the LCD screen. The data mining and statistical analysis methods are used to quantitatively evaluate the characteristics of the screen such as response time, dynamic contrast, and brightness uniformity. The specific operation includes establishing a mathematical model, comparing the historical experimental data with the current data, and finally generating screen response characteristic data to reflect the performance of the screen under different working conditions.

[0059] Step S15: designing a multi-scale sampling scheme according to the screen physical characteristic data, and using the sampling parameter data to perform multi-scale image acquisition on the LCD display content to generate multi-scale image data;

[0060] In the embodiment of the present invention, the system designs a multi-scale sampling scheme based on the screen physical property data. This scheme includes determining the sampling level (such as full HD and ultra HD) and the sampling frequency (such as 30 frames per second). The system automatically adjusts the camera parameters to adapt to different resolutions and collects the corresponding LCD display content through the control signal generation module. Finally, the multi-scale image data generated by the system will be stored for subsequent analysis.

[0061] Step S16: classifying the characteristics of the liquid crystal screen according to the screen response characteristic data, and performing dynamic edge compensation analysis on the multi-scale image data to obtain edge feature compensation data.

[0062] In the embodiment of the present invention, the system classifies the characteristics of the LCD screen according to the screen response characteristic data. The screen is trained and classified using a machine learning algorithm (such as a support vector machine) to identify different types of screens (such as high contrast screens, fast response screens, etc.). Then, for each type of screen, the system performs dynamic edge compensation analysis on the multi-scale image data, applies an edge detection algorithm (such as the Canny algorithm), and obtains corresponding edge feature compensation data to improve the accuracy of subsequent processing.

[0063] The present invention accurately characterizes the basic physical properties of the liquid crystal screen, and these data directly affect the display effect of the screen, especially in the application of high-precision display and large-size screen. These physical characteristics provide a basic reference for the design of the multi-scale sampling scheme in the subsequent steps, which helps to process the image more accurately. By deeply understanding the electrical characteristics of the screen, the response time, brightness change and current fluctuation of the liquid crystal screen can be better predicted and adjusted. These parameters are very critical for processing high refresh rate display and coping with response delay phenomena, which helps to ensure display stability. Combining physical and electrical characteristics, the overall characteristics of the liquid crystal screen can be described in an all-round way. By integrating these two types of data, the system can more accurately simulate the response behavior of the liquid crystal screen under different conditions, providing a more accurate basis for subsequent screen response characteristic analysis and image processing. Based on the comprehensive analysis of physical and electrical parameters, the characteristics of the screen under different operating conditions, such as response speed, refresh rate fluctuation and color performance, can be effectively predicted. These response characteristics are key factors affecting image display quality and edge clarity, and this analysis can better provide guidance for the compensation of subsequent edge features. Through multi-scale sampling, the system can obtain rich image data from different resolutions and different scales. Multi-scale images can help capture the details of the LCD display, while adapting to the image performance under different screen sizes and pixel densities, especially when processing complex images or large-size displays. Dynamic edge compensation analysis can adjust the image edge processing method according to the response characteristics of the LCD screen, reducing the edge blur caused by response delays, refresh rate fluctuations, etc. By processing the characteristics in each category separately, the edge compensation effect in different situations can be accurately optimized to ensure that the final display result is clearer and more stable. The entire process comprehensively captures the characteristics of the LCD screen by combining physical and electrical characteristics, and improves the flexibility and accuracy of image processing through multi-scale acquisition and dynamic edge compensation; physical characteristic data provides guidance for the design of image acquisition solutions, while electrical characteristic analysis helps to better understand the response behavior of the screen; this comprehensive processing method ensures high-quality segmentation and edge processing of LCD screen images, especially showing better performance under high refresh rates and complex scenes.

[0064] Preferably, step S14 comprises the following steps:

[0065] Step S141: performing a grayscale switching response time test according to basic data of the LCD screen, thereby obtaining grayscale response time data;

[0066] The embodiment of the present invention first prepares a standard LCD screen test device that can quickly switch grayscale values. Before the test begins, the system records the current LCD screen setting parameters, including brightness and contrast. Subsequently, the device switches between different grayscale values ​​(such as from black to white) at a preset frequency, and uses a high-precision photoelectric sensor to measure the response time required by the LCD screen for each switch. The data is recorded and stored in real time. After multiple tests, the system calculates the average response time and standard deviation, and finally generates grayscale response time data for subsequent analysis.

[0067] Step S142: performing dynamic contrast analysis based on the grayscale response time data to generate dynamic contrast characteristic data;

[0068] In the embodiment of the present invention, the system performs dynamic contrast analysis using the grayscale response time data obtained in step S141. During operation, the system creates a database containing different grayscale values ​​and corresponding response times. Then, a curve fitting algorithm is applied to analyze the relationship between response time and contrast through linear or polynomial regression to determine the changing trend of dynamic contrast. The generated dynamic contrast characteristic data is stored to evaluate the performance of the LCD screen in rapidly changing images.

[0069] Step S143: testing and analyzing the overshoot and undershoot phenomena of pixels according to the electrical characteristic data of the screen to obtain pixel stability data;

[0070] The system of the embodiment of the present invention uses the screen electrical characteristic data to test the overshoot and undershoot phenomena of each pixel. In specific operation, the system applies a fast voltage pulse to the LCD screen through a signal generator, and then uses an oscilloscope to monitor the response waveform of each pixel. By analyzing the captured waveform, the system identifies the overshoot and undershoot characteristics of the pixel and calculates the response stability index of each pixel. Finally, these data are sorted to generate pixel stability data to evaluate the stability of the LCD screen in dynamic image display.

[0071] Step S144: constructing a motion blur characteristic model according to the grayscale response time data and the dynamic contrast characteristic data, thereby generating motion blur characteristic data;

[0072] According to the grayscale response time data and dynamic contrast characteristic data, the system of the present invention constructs a motion blur characteristic model. During operation, the system uses multiple regression analysis, takes the grayscale response time and dynamic contrast as input variables, establishes a mathematical model, and predicts the degree of motion blur. By comparing the blur degree of the real image and the moving image, the system adjusts the model parameters to optimize the accuracy of the motion blur characteristic data, thereby ensuring that the model can effectively describe the blur effect at different motion speeds.

[0073] Step S145: performing inter-frame interpolation capability analysis according to the refresh rate information in the screen electrical characteristic data, thereby obtaining frame interpolation characteristic data;

[0074] The system of the embodiment of the present invention performs inter-frame interpolation capability analysis based on the refresh rate information in the screen electrical characteristic data. The specific operation includes collecting the display content of multiple consecutive frames and comparing it with the current refresh rate to evaluate its impact on image smoothness. Using an interpolation algorithm (such as bilinear interpolation or cubic interpolation), the system calculates the smooth transition between frames and generates frame interpolation characteristic data to characterize the smoothness of the LCD screen in dynamic scenes.

[0075] Step S146: performing data integration based on high-dimensional feature fusion on the pixel stability data, frame interpolation characteristic data, motion blur characteristic data and grayscale response time data to obtain screen response characteristic data.

[0076] The system of the embodiment of the present invention performs high-dimensional feature fusion on the pixel stability data, frame interpolation characteristic data, motion blur characteristic data and grayscale response time data obtained in steps S143, S144 and S145. In the specific operation, the principal component analysis (PCA) and other dimensionality reduction techniques are applied to integrate these multi-dimensional data into a comprehensive data set, extract key features, and form screen response characteristic data. This data set is then used to evaluate the overall performance of the LCD screen and provide data support for subsequent optimization.

[0077] The grayscale response time of the present invention is a key indicator for measuring the time required for the LCD screen to switch from one grayscale (brightness level) to another grayscale. By accurately measuring the grayscale switching time, the response speed of the LCD screen can be understood, especially the performance in high dynamic scenes. The grayscale switching speed directly affects the smoothness and image quality of the image display. A faster grayscale response can reduce smearing and blurring, and improve the visual experience. The dynamic contrast ratio represents the brightness ratio of the LCD screen under different brightness conditions. By analyzing the grayscale response time, the dynamic change of the contrast of the screen can be evaluated, and the performance of the LCD screen in scenes with drastic changes in light and dark can be better optimized to ensure that the dark details are as clear as the highlights, thereby improving the image quality. The improvement of dynamic contrast can effectively enhance the visual impact, especially in the application of high dynamic range (HDR) display technology. Overshoot and undershoot phenomena can cause inaccurate brightness performance in the image, affecting the overall image quality. By testing these phenomena, it can be ensured that each pixel can accurately reach the expected value when responding, thereby improving the stability and consistency of the display, reducing image distortion, and improving the viewing experience. By building a motion blur model, the system can predict and correct blur caused by changes in response time and contrast, which helps improve image clarity and sharpness, especially for high-dynamic scenes such as games and sports. Interframe interpolation is a technology that generates transition frames between two image frames to improve image smoothness and fluency. By analyzing the frame interpolation capability, it can effectively reduce the problem of screen freeze and jitter in low frame rate display or complex motion scenes, and improve the overall viewing experience. Especially in LCD screens with high refresh rates (such as 120Hz, 240Hz), good frame interpolation performance helps make motion pictures smoother and more natural. By fusing multi-dimensional data, a comprehensive screen response characteristic data model can be formed to fully characterize the performance of LCD screens under different conditions. This high-dimensional feature fusion can more accurately reflect the dynamic response capability of the LCD screen, especially in rapidly changing scenes. This integrated data can help optimize the overall display effect of the LCD screen. The entire process builds a high-precision LCD screen response characteristic model through comprehensive analysis and integration of grayscale response time, dynamic contrast, pixel stability, frame interpolation capability and motion blur characteristics. Based on the grayscale response time test, dynamic contrast analysis improves the detail performance of brightness changes, pixel overshoot and undershoot tests ensure the accuracy of display, motion blur model reduces blur in high-speed motion images, and frame interpolation analysis ensures the smoothness of the picture. Finally, through high-dimensional feature fusion, this systematic analysis ensures that the LCD screen can provide excellent display effects in various complex scenes.

[0078] Preferably, step S16 comprises the following steps:

[0079] Step S161: establishing a screen response characteristic classification model according to the grayscale response time data, dynamic contrast characteristic data and pixel stability data in the screen response characteristic data;

[0080] The embodiment of the present invention collects grayscale response time data, dynamic contrast characteristic data and pixel stability data of the LCD screen. These data are obtained through experimental measurement, and a classification model is established using machine learning algorithms such as linear regression, decision tree or support vector machine. The goal of the model is to classify the LCD screen according to different response characteristics. The specific operations include data preprocessing, feature selection and model training, and finally generate a model that can be used for screen response characteristic classification.

[0081] Step S162: Identify the type of the liquid crystal screen according to the screen response characteristic classification model, thereby generating screen type identification data, wherein the screen type identification data includes a high grayscale response time type, a high dynamic contrast type, an unstable pixel type, and a standard response type;

[0082] The screen response characteristic classification model established in the embodiment of the present invention identifies the type of liquid crystal screen. The input is the previously collected screen response characteristic data, and the model output is the screen identification of different types. The specific operation includes standardizing the input data, and using the classification model for prediction to generate screen type identification data. According to the model results, the identification data can be divided into high grayscale response time type, high dynamic contrast type, unstable pixel type and standard response type, which is convenient for subsequent processing.

[0083] Step S163: selecting an edge detection algorithm according to the screen type identification data, and extracting edge features from the multi-scale image data using the selected edge detection algorithm, thereby obtaining preliminary edge feature data, wherein the algorithm selection is specifically: for a screen whose screen type identification data is a high grayscale response time type, the selected algorithm is the Canny edge detection algorithm; for a screen whose screen type identification data is a high dynamic contrast type, the selected algorithm is the Sobel algorithm; for a screen whose pixel is unstable type, the selected algorithm is the Laplacian algorithm; and for a screen whose screen type is a standard response type, the selected algorithm is the Prewitt algorithm;

[0084] The embodiment of the present invention selects a suitable edge detection algorithm based on the screen type identification data generated in step S162. The specific operation is to first determine the type of screen, and then specify the corresponding edge detection algorithm for each type: the Canny algorithm is used for the high grayscale response time type, the Sobel algorithm is used for the high dynamic contrast type, the Laplacian algorithm is used for the pixel instability type, and the Prewitt algorithm is used for the standard response type. After the algorithm is selected, the multi-scale image data is processed using the selected edge detection algorithm to extract preliminary edge feature data to ensure that the extraction results can reflect the edge characteristics of different types of screens.

[0085] Step S164: construct a dynamic edge blur compensation model according to the screen response characteristic data, and use the dynamic edge blur compensation model to perform edge compensation analysis on the preliminary edge feature data, so as to obtain edge feature compensation data.

[0086] The embodiment of the present invention uses screen response characteristic data to construct a dynamic edge blur compensation model. First, the preliminary edge feature data is analyzed to determine the characteristics and influencing factors of the blur phenomenon. A model is designed based on these characteristics, and signal processing technology, such as Fourier transform or wavelet transform, is used to analyze the impact of edge blur on image quality. After the model is established, it is applied to the preliminary edge feature data for edge compensation analysis to generate edge feature compensation data, thereby improving the clarity of the image and the accuracy of the edge.

[0087] The present invention can accurately classify the key response characteristics of the liquid crystal screen according to the actual performance characteristics of different liquid crystal screens by classifying and modeling. This classification model provides a customized basis for subsequent image processing and algorithm selection, especially for different types of liquid crystal screens, it has a significant optimization effect when processing image features. For example, some liquid crystal screens may respond slowly in grayscale switching, while others show higher dynamic contrast, so classifying them helps to optimize the subsequent steps for different problems. The classification model identifies the specific types of liquid crystal screens, such as high grayscale response time, high dynamic contrast, unstable pixels, etc., and generates corresponding type identification data. This enables the system to perform targeted image processing and optimization based on the recognition results, improving the flexibility and adaptability of the overall processing flow. For different types of screens, the use of corresponding optimization strategies can effectively improve the image processing effect and avoid unnecessary losses caused by blindly adopting a unified processing method. The system automatically selects the most suitable edge detection algorithm according to different types of liquid crystal screens. For example, the Canny algorithm is selected for screens with high grayscale response time, which can more effectively deal with the edge blur problem caused by slow grayscale response; for screens with high dynamic contrast, the Sobel algorithm can better capture the edge changes caused by brightness differences; the Laplacian algorithm is used for screens with unstable pixels to better eliminate the problem of detail loss caused by overshoot or undershoot. The Prewitt algorithm is used for standard response type screens to achieve stable edge detection. This differentiated algorithm selection improves the accuracy of edge detection and ensures that the edge detection results can achieve the best effect under different screen characteristics. Through the dynamic edge blur compensation model, the system can dynamically optimize the edge detection results to compensate for the edge blur problem caused by factors such as LCD screen response time, contrast, and pixel instability. Especially when processing moving images or images at high refresh rates, this compensation model can effectively improve the clarity and stability of the edge, reduce blur and jagged phenomena, and improve the display effect. Through compensation strategies for different screen types, the high accuracy of edge features is ensured, and the quality of image segmentation is ultimately improved. The entire process forms an efficient edge detection and compensation mechanism by establishing a classification model for LCD screen response characteristics, selecting appropriate edge detection algorithms according to different screen types, and combining dynamic edge blur compensation models. The classification model enables the system to accurately select optimization solutions based on screen characteristics, the edge detection algorithm selection improves the processing effect, and the compensation model ensures the clarity of edge features. In this way, the entire system can adapt to different types of LCD screens and achieve accurate edge detection and compensation, improving the quality and robustness of image processing.

[0088] Preferably, step S2 comprises the following steps:

[0089] Step S21: collecting the touch coordinates, pressure value and touch area information of the touch screen in real time, so as to obtain real-time data of the touch screen;

[0090] Step S22: identifying and extracting the pressure deformation interference area according to the real-time data of the touch screen, thereby generating pressure deformation interference data;

[0091] Step S23: analyzing the electric field changes around the touch point according to the touch coordinate information in the real-time data of the touch screen, thereby obtaining electric field interference data;

[0092] Step S24: Analyze the timing relationship between touch operation and screen refresh using the refresh rate information in the basic data of the LCD screen, so as to generate refresh rate influence interference data;

[0093] Step S25: analyzing the response delay of the touch operation according to the real-time data of the touch screen and the edge feature compensation data, thereby obtaining response delay interference data;

[0094] Step S26: analyzing the afterimage effect caused by the touch operation on the touch screen real-time data and the screen response characteristic data, thereby generating afterimage effect interference data;

[0095] Step S27: performing spatial superposition and temporal correlation analysis on the pressure deformation interference data, the electric field interference data, the refresh rate interference data, the response delay interference data and the afterimage effect interference data, so as to obtain comprehensive touch interference data;

[0096] Step S28: establishing a touch interference space-time model according to the comprehensive touch interference data, and generating preliminary touch interference mask data;

[0097] Step S29: using the edge feature compensation data to perform edge alignment and refinement processing on the preliminary touch interference mask data, thereby obtaining the touch interference mask data.

[0098] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S2 in the embodiment of the present invention, step S2 includes the following steps:

[0099] Step S21: collecting the touch coordinates, pressure value and touch area information of the touch screen in real time, so as to obtain real-time data of the touch screen;

[0100] The embodiment of the present invention collects the touch coordinates, pressure value and touch area information of the touch screen in real time, adopts a high-frequency sampling method, and generally continuously monitors the touch events of the touch screen within a millisecond time interval. When the user performs a touch operation, the system obtains the coordinate position (X, Y coordinates) of the touch point, the pressure value applied to the screen surface (usually measured by pressure sensing technology), and the touch area (calculated by the contact area) through the touch control chip. These data can be updated in real time and used as the basic input data for subsequent interference identification to obtain real-time data of the touch screen.

[0101] Step S22: identifying and extracting the pressure deformation interference area according to the real-time data of the touch screen, thereby generating pressure deformation interference data;

[0102] The embodiment of the present invention analyzes the pressure deformation of the touch point and its surroundings based on the collected touch pressure values ​​and combined with the screen physical property data (such as material properties, thickness, etc.) by presetting the pressure value threshold. The deformation area around the touch point is simulated using the finite element analysis method to determine the range of pressure deformation. By comparing the actual pressure distribution with the model of normal touch operation, the deformation area caused by excessive or uneven pressure can be extracted and marked as the pressure deformation interference area, thereby generating pressure deformation interference data.

[0103] Step S23: analyzing the electric field changes around the touch point according to the touch coordinate information in the real-time data of the touch screen, thereby obtaining electric field interference data;

[0104] During the touch operation, the embodiment of the present invention uses electric field sensing technology to analyze the electric field changes around the touch point based on the touch coordinate information and the characteristics of the capacitive screen. By measuring the capacitance change around the touch point, the electric field disturbance caused by the user's finger or other objects on the screen surface is inferred. When the electric field change exceeds the preset threshold, it is marked as an electric field interference area. At the same time, the electric field interference range is adjusted in combination with the basic screen data (such as screen size, resolution, etc.), and finally the electric field interference data is obtained.

[0105] Step S24: Analyze the timing relationship between touch operation and screen refresh using the refresh rate information in the basic data of the LCD screen, so as to generate refresh rate influence interference data;

[0106] The embodiment of the present invention uses the refresh rate information of the liquid crystal screen to compare and analyze the timestamp of the touch operation with the screen refresh cycle to determine the time synchronization between the touch event and the screen refresh event. By calculating the time difference between the time point when the touch event occurs and the screen refresh moment, it is identified whether the touch event has interfered with the display of the screen. If the touch event occurs at a critical moment of screen refresh (such as frame switching or refresh), it may cause display instability. At this time, the area is marked as a refresh rate interference area, thereby generating refresh rate interference data.

[0107] Step S25: analyzing the response delay of the touch operation according to the real-time data of the touch screen and the edge feature compensation data, thereby obtaining response delay interference data;

[0108] The embodiment of the present invention analyzes the delay between the touch operation and the screen response based on the touch coordinates and the screen response characteristic data in the real-time data of the touch screen. By calculating the difference between the time when the touch event occurs and the actual response time displayed on the screen, it is possible to identify whether there is a large response delay phenomenon. When the delay exceeds a preset threshold (for example, the response time is greater than a certain number of milliseconds), the area is marked as a response delay interference area, and response delay interference data is formed for subsequent edge feature correction.

[0109] Step S26: analyzing the afterimage effect caused by the touch operation on the touch screen real-time data and the screen response characteristic data, thereby generating afterimage effect interference data;

[0110] After the touch operation is completed, the embodiment of the present invention uses the response characteristic data of the LCD screen to analyze whether the afterimage effect is generated on the screen. By comparing the screen display images before and after the touch, combined with the response time and grayscale conversion characteristics of the LCD screen, it is detected whether there is residual image information on the screen that has not completely disappeared. Through the grayscale difference analysis method, the range and intensity of the afterimage effect can be determined, and the affected area can be marked to form the afterimage effect interference data.

[0111] Step S27: performing spatial superposition and temporal correlation analysis on the pressure deformation interference data, the electric field interference data, the refresh rate interference data, the response delay interference data and the afterimage effect interference data, so as to obtain comprehensive touch interference data;

[0112] The embodiment of the present invention performs spatial superposition analysis on the pressure deformation interference data, electric field interference data, refresh rate interference data, response delay interference data, and afterimage effect interference data obtained above, and finds out the comprehensive interference area by calculating the overlapping parts of each interference area. At the same time, a time series correlation analysis is performed to arrange each interference data in chronological order, identify which interferences occur in the same time period, and analyze the interactions between them. In this way, more accurate comprehensive touch interference data can be obtained for building a comprehensive interference model.

[0113] Step S28: establishing a touch interference space-time model according to the comprehensive touch interference data, and generating preliminary touch interference mask data;

[0114] The embodiment of the present invention uses a space-time model to model touch interference based on comprehensive touch interference data. The model combines the two dimensions of spatial position and time change to analyze the impact of touch events in different times and regions. Through spatial interpolation and temporal interpolation methods, a more refined touch interference area boundary can be generated, and the interference area can be expanded or contracted through morphological operations. Finally, the model outputs preliminary touch interference mask data to mark the image area that may be affected by touch.

[0115] Step S29: using the edge feature compensation data to perform edge alignment and refinement processing on the preliminary touch interference mask data, thereby obtaining the touch interference mask data.

[0116] The embodiment of the present invention uses edge feature compensation data to perform edge alignment and refinement processing on the preliminary touch interference mask data. First, according to the compensated edge features, the boundaries of the touch interference area are aligned to match the real edges of the image. Then, through morphological filtering, refinement algorithm and other technologies, the noise or burrs in the mask are removed to ensure smooth and accurate edges. Finally, high-precision touch interference mask data is generated for interference elimination in the subsequent image segmentation process.

[0117] The present invention can obtain the physical information of touch operation in real time, capture key data such as touch position, force and contact area, and through this information, the system can accurately identify the user's touch behavior, ensure the accuracy of touch interference analysis in subsequent steps, especially in the multi-touch scene, real-time data collection lays the foundation for accurate touch interference identification. By analyzing the pressure value, the screen deformation interference caused by excessive pressure is identified. Pressure deformation may cause image distortion or operation error. By identifying these interference areas, the interference between touch operation and image processing can be avoided, and the accuracy of image display and the accuracy of touch response can be improved. The working principle of the touch screen depends on electric field induction. The touch operation may cause electric field changes, which in turn affects image display or other touch operations. By analyzing the electric field changes, potential electric field interference can be identified, the touch recognition accuracy can be optimized, and the situation of false touch or touch failure can be reduced. The refresh rate of the LCD screen will affect the timing response of the touch, especially on a screen with a high refresh rate, the incoordination between the touch operation and the screen refresh may cause the picture to lag or skip frames. By analyzing this timing relationship, the interference caused by the mismatch between touch and refresh rate can be reduced, the touch operation can be ensured to be synchronized with the image refresh, and the response speed and fluency of the touch operation can be improved. Identify the delay phenomenon in touch operation. Touch delay will affect the user experience, especially in application scenarios with frequent interactions, where delay may cause lag or error in operation response. By analyzing the delay, delay interference can be identified and reduced, the response time of the system can be optimized, and more timely feedback can be provided. By analyzing the residual image effect, the residual image interference caused by touch operation can be identified and processed, the image blur can be reduced, and the image clarity can be ensured, especially in high-speed operation scenarios. Through cross-comparison in space and time, a global interference recognition model is formed. Comprehensive analysis helps to capture the superposition of multiple interferences in complex scenarios, so as to more comprehensively identify touch interference and improve the accuracy and detail of interference judgment. Different types and dimensions of touch interference data are integrated into a unified model to generate a touch interference mask. This mask can shield the interference part in the touch operation and avoid the influence of the interference area on the image processing or display effect, thereby improving the overall stability and display accuracy of the system. By combining with edge feature compensation data, the touch interference mask is accurately aligned and compensated to ensure the consistency between the edge of the mask area and the actual interference area. Refined processing can improve the accuracy of the mask and avoid misidentification or omission of interference areas, thereby ensuring the final touch interference shielding effect and improving the robustness of image processing. The entire process identifies interference such as pressure deformation, electric field changes, refresh rate effects, response delays, and afterimage effects through multi-dimensional analysis of real-time touch screen data, and finally generates touch interference mask data. The effect of these steps is reflected in improving the response accuracy of touch operations, reducing the impact of interference on image processing, and ensuring that the system can still maintain efficient and accurate image processing capabilities in complex touch operation scenarios.Through comprehensive analysis of various interference factors and refined edge processing, effective shielding and optimization of touch interference were finally achieved.

[0118] Preferably, step S3 comprises the following steps:

[0119] Step S31: performing mask processing on the edge feature compensation data according to the touch interference mask data, so as to obtain preliminary edge feature correction data;

[0120] The embodiment of the present invention performs mask processing on the edge feature compensation data according to the touch interference mask data. First, the mask processing is performed through Boolean operations at the pixel level: the marked area in the touch interference mask data is bitwise operated with the edge feature compensation data. For the area marked as interference in the mask, the corresponding edge feature compensation data is set to zero or the intensity is reduced proportionally (usually weight attenuation) to eliminate the influence of touch interference. Thereafter, for areas with broken or distorted edges, the edge data of the interfered area is smoothed using interpolation algorithms such as bilinear interpolation or Gaussian smoothing, so that the edge feature transition is natural and continuity is maintained, thereby obtaining preliminary edge feature correction data.

[0121] Step S32: using the grayscale response time data in the screen response characteristic data to perform time domain compensation on the preliminary edge feature correction data, thereby obtaining time compensated edge feature data;

[0122] The embodiment of the present invention uses the grayscale response time data in the screen response characteristic data to perform time domain compensation on the preliminary edge feature correction data. First, according to the grayscale response time in the screen response characteristics, the response characteristics of each pixel are analyzed to identify the pixel area with a longer response time. Then, in the time domain, the edge data of these areas are accumulated, and the time series weighted average method is used to smooth the edges of the delayed response. By analyzing the edge feature changes of the previous and next frames, the influence of the tailing effect is smoothed in time. The pixel area with a slower response is filtered through time weighting to reduce the edge jitter or blur, and generate time-compensated edge feature data.

[0123] Step S33: performing contrast enhancement processing on the time-compensated edge feature data, and performing noise suppression processing according to the pixel stability data, so as to obtain final edge feature correction data;

[0124] The embodiment of the present invention performs contrast enhancement processing on time-compensated edge feature data. First, the grayscale distribution of the edge feature data in the current image is analyzed using a histogram stretching technique to enhance the contrast of the grayscale values. The specific operation is to stretch the low-contrast grayscale values ​​in the image to a wider range, thereby making the subtle edge features more prominent. Next, based on the pixel stability data, an adaptive median filtering method is used to suppress the noise area. During the processing, filtering is applied to pixel areas with poor stability to eliminate random noise, and the edge data is retained and enhanced, ultimately generating clearer edge feature data to form final edge feature correction data.

[0125] Step S34: constructing an adaptive threshold model based on local image features and global brightness distribution according to the final edge feature correction data, and performing segmentation processing on the final edge feature correction data according to the adaptive threshold model, thereby obtaining preliminary image segmentation data;

[0126] The embodiment of the present invention constructs an adaptive threshold model based on the final edge feature correction data. First, the image is divided into several local areas, and the grayscale histogram of each area is calculated, and the local grayscale distribution characteristics are analyzed to generate a local adaptive threshold. Then, combined with the global brightness distribution, the global threshold of the image is calculated to ensure that the edge segmentation effect of different brightness areas is consistent. The specific threshold generation uses a weighted average algorithm to combine the local threshold with the global threshold. Using the adaptive threshold model, a threshold judgment is performed on each pixel, and pixels with grayscale values ​​exceeding the threshold are marked as edge pixels, thereby segmenting the final edge feature correction data to generate preliminary image segmentation data.

[0127] Step S35: performing time consistency constraint processing on the preliminary image segmentation data according to the screen response characteristic data, thereby generating segmentation strategy data.

[0128] The embodiment of the present invention performs temporal consistency constraint processing on the preliminary image segmentation data according to the screen response characteristic data. First, a temporal consistency constraint model is constructed using the time response data in the screen response characteristics. In this process, the edge change trend of the moving target is analyzed by comparing the segmentation results of multiple consecutive frames of images. A matching analysis is performed on similar edge regions segmented in adjacent frames, and a temporal filtering algorithm is used to eliminate discontinuous short-term noise or false edges to ensure the temporal continuity of the moving target. At the same time, a timing analysis technique such as Kalman filtering is used to predict and correct rapidly changing edge regions to ensure the stability of edge segmentation in dynamic scenes. Finally, segmentation strategy data is generated through temporal consistency constraints to guide subsequent segmentation processing.

[0129] The present invention shields the interference area caused by the touch operation through the touch interference mask to prevent these interferences from affecting the accuracy of edge feature analysis; the mask processing can exclude irrelevant areas or interference areas, so that the edge correction is more accurate, ensuring that the subsequent edge feature correction steps focus on the real image features. The grayscale response time data reflects the difference in response speed of the LCD screen under different grayscale switching, which affects the dynamic change of the edge feature. Through time domain compensation, the edge blur or delay caused by inconsistent response time is adjusted to improve the time accuracy of the edge feature, especially to reduce the blur effect in high-speed scenes. Through contrast enhancement, the edge feature is made more obvious in the image, especially in complex image scenes. Enhance the edge clarity. At the same time, the noise suppression processing can remove random noise in the image through pixel stability data to ensure the purity and consistency of the edge feature. Finally, the edge feature correction data is more accurate and clear, laying a solid foundation for subsequent image segmentation. By constructing an adaptive threshold model, the system can dynamically adjust the segmentation threshold according to the local details and global brightness distribution of the image. The adaptive threshold model can make image segmentation more flexible, adapt to different lighting conditions and changes in image details, ensure that the segmentation process can distinguish between edge areas and background areas, and improve segmentation accuracy. By constraining the screen response characteristic data, the segmentation results are prevented from jumping or being inconsistent in the time series, ensuring that the segmentation results of each frame of the image are consistent with the previous and next frames. Doing so can significantly improve the stability of image segmentation, making the segmentation strategy smoother and more reliable throughout the operation. The effect of the entire chain of steps is that the edge feature data is more accurate and clear through multiple processing methods such as mask processing, time domain compensation, contrast enhancement and noise suppression. Subsequently, the segmentation processing based on the adaptive threshold model adapts to the local and global characteristics of the image, while the temporal consistency constraint ensures the stability of the segmentation results in the time dimension; these steps ultimately improve the accuracy, stability and consistency of image segmentation, and are particularly suitable for LCD screen image processing in complex scenes.

[0130] Preferably, step S35 includes the following steps:

[0131] Step S351: establishing a temporal consistency constraint model according to the dynamic contrast characteristic data and the motion blur characteristic data in the screen response characteristic data;

[0132] The embodiment of the present invention establishes a timing consistency constraint model based on the dynamic contrast characteristic data and motion blur characteristic data in the screen response characteristic data. First, the dynamic contrast change data and motion blur characteristic data are extracted from the screen response characteristic data, and the contrast adjustment and motion blur characteristics of the screen in different scenes are analyzed. These data are used to build a timing consistency constraint model, which ensures that the boundaries segmented in the motion scene have timing consistency by comparing the dynamic contrast changes and motion blur effects of consecutive frames. The model comprehensively considers the response time and contrast changes, so that the edges can maintain a consistent visual effect between different frames, forming a basic timing constraint model.

[0133] Step S352: using the temporal consistency constraint model to perform consistency analysis on the cross-frame segmentation results of the preliminary image segmentation data, thereby obtaining temporal correction segmentation data;

[0134] The embodiment of the present invention utilizes a temporal consistency constraint model to perform consistency analysis on the segmentation results across frames of preliminary image segmentation data. In specific operation, the preliminary image segmentation data of multiple consecutive frames are input into the temporal consistency constraint model, and the model detects the edge changes across frames by comparing and matching the segmentation results between adjacent frames. For the erroneous segmentation areas caused by timing inconsistencies, timing correction is performed through interpolation or filtering operations. The edge stability in consecutive frames is analyzed, and short-term edge jitters or erroneous edges are corrected to stable edge contours to ensure that the edges of moving objects remain consistent in each frame, thereby obtaining timing-corrected segmentation data.

[0135] Step S353: performing refinement processing on the segmentation boundary according to the timing correction segmentation data and the screen response characteristic data, thereby obtaining optimized segmentation data;

[0136] The embodiment of the present invention performs fine processing of the segmentation boundary according to the timing correction segmentation data and the screen response characteristic data. First, the edge enhancement algorithm is used to refine the segmentation boundary by combining the edge area marked in the timing correction segmentation data with the contrast and blur characteristics in the screen response characteristic data. The specific operation is to perform secondary edge detection on the fuzzy boundary using the Sobel operator or the Canny algorithm to enhance the edge clarity. Then, for complex edge details, sub-pixel level interpolation processing technology is used to ensure the accuracy and continuity of the boundary, eliminate possible blur or damage, and finally generate optimized segmentation data.

[0137] Step S354: constructing a segmentation region growing rule set according to the optimized segmentation data, wherein the segmentation region growing rule set includes a region similarity evaluation and a boundary extension strategy;

[0138] The embodiment of the present invention constructs a segmentation region growing rule set based on the optimized segmentation data. First, based on the segmentation results, the similarity between the boundary of each segmented region and its adjacent regions is evaluated, and the differences in color, texture and grayscale of each region are analyzed. The region growing rule set is constructed using these similarity data. In the rule set, the region expansion strategy is defined: according to the similarity score, the adjacent pixels are evaluated, and if the similarity reaches the threshold, the boundary expansion is allowed. At the same time, the smoothness of the boundary shape is considered to ensure that the boundary does not expand excessively or produce abnormal expansion during expansion, thereby forming a reasonable region growing rule set.

[0139] Step S355: Optimizing the segmentation strategy of the optimized segmentation data according to the segmentation region growing rule set, thereby generating segmentation strategy data.

[0140] The embodiment of the present invention optimizes the segmentation strategy of the optimized segmentation data according to the segmentation region growth rule set. In this process, each region in the optimized segmentation data is firstly expanded according to the growth rule. For each pixel, the range of the region is gradually expanded according to the similarity evaluation and boundary expansion strategy to ensure the consistency of the pixels in the region and prevent excessive overlap between regions. In the process of boundary expansion, the accuracy of the boundary expansion process is ensured by dynamically adjusting the expansion step size and the similarity threshold, and finally the optimized segmentation strategy data is generated, so that the segmented region is more consistent with the real structure of the image.

[0141] The dynamic contrast and motion blur of the present invention reflect the visual characteristics of the image when it is displayed. By establishing a temporal consistency constraint model through them, it can be ensured that the segmentation results of the image in continuous frames remain consistent, and there will be no discontinuous or jittery segmentation caused by rapid scene switching or motion blur. The temporal consistency constraint model provides a temporal continuity guarantee for the image segmentation results, which is particularly suitable for dynamic display and video application scenarios. Through cross-frame analysis, inconsistent segmentation phenomena caused by inter-frame differences can be identified and corrected. Timing correction can keep the segmentation results the same logic between different frames, avoiding segmentation errors caused by scene changes or noise interference. This is particularly critical for processing dynamically changing images, ensuring the continuity of segmentation across frames and improving the overall image segmentation quality. The segmentation data after timing correction is further refined by combining the screen response characteristics, which can enhance the accuracy of the segmentation results, especially when the boundaries are blurred or unclear. Through this processing step, the segmentation boundaries are clearer and sharper, especially in scenes with more severe motion blur, and this refinement greatly improves the accuracy of the boundaries. By defining regional similarity and boundary extension strategies, the growth process of the segmented region can be accurately controlled. The regional similarity evaluation can ensure that the visual characteristics of the new segmented region are consistent with the existing region, while the boundary extension strategy further optimizes the boundary processing of the segmentation result, so that the segmented region can be naturally expanded according to the image content, thereby ensuring the integrity and coherence of the segmented region. By optimizing the segmentation region growth rule, the final segmentation strategy can be made more flexible and adaptable. In the region growing process, not only the similarity evaluation is considered, but also the boundary extension is adjusted according to the local characteristics of the image to ensure the accurate division of the segmented region. The final segmentation strategy can provide more accurate and complete segmentation regions in complex image scenes and can adapt to dynamically changing image content. These steps further improve the accuracy, stability and continuity of image segmentation by establishing a temporal consistency constraint model, cross-frame consistency analysis, boundary refinement processing and segmentation region growth optimization. Especially in dynamic scenes, these processing steps enable image segmentation to effectively deal with the blur and jump problems caused by the screen response characteristics, ensuring the integrity and consistency of the segmentation results.

[0142] Preferably, step S4 comprises the following steps:

[0143] Step S41: using the image acquisition device of the LCD screen to perform high-frequency sampling on the current display content, thereby obtaining real-time image data;

[0144] The embodiment of the present invention utilizes the image acquisition device of the LCD screen to perform high-frequency sampling of the current display content. In specific implementation, the display content on the LCD screen can be continuously sampled multiple times per second through the built-in image acquisition sensor or external camera device. During the sampling process, it is necessary to ensure that the sampling frequency of the device is high enough to capture the slight changes in the screen content, especially the details of the fast-moving picture. Each sampled image data will be pre-processed, such as noise reduction and color space conversion, to ensure that the captured image has a high quality, and finally generate continuous real-time image data, providing a basis for subsequent processing.

[0145] Step S42: using a foreground extraction algorithm to perform preliminary separation of static background and dynamic foreground on the real-time image data, and extracting motion features of the separated dynamic foreground based on optical flow analysis and inter-frame difference, thereby obtaining moving target data;

[0146] The embodiment of the present invention uses a foreground extraction algorithm to perform a preliminary separation of static background and dynamic foreground on real-time image data. First, the static part of the displayed content is modeled by a static background modeling method (such as a Gaussian mixture model), and then the image of the current frame is compared with the background model to preliminarily separate the dynamic foreground. Next, the separated dynamic foreground is subjected to motion analysis using an optical flow method, and the motion trajectory of the object is extracted by the change in the optical flow field between frames. At the same time, the inter-frame difference algorithm is used to compare consecutive frames to further confirm the motion characteristics of the dynamic foreground, and finally the motion target data is obtained to provide motion information for subsequent clustering and boundary processing.

[0147] Step S43: clustering the moving targets according to the moving target data and the regional similarity evaluation rules in the segmentation strategy data, thereby obtaining preliminary moving target candidate regions;

[0148] The embodiment of the present invention clusters moving targets according to the moving target data and the regional similarity evaluation rules in the segmentation strategy data. First, the extracted moving target data is analyzed, and the moving targets are classified according to the similarity features such as the color, texture and shape of the target using K-means clustering or hierarchical clustering algorithm. In this process, according to the similarity evaluation rules defined in the segmentation strategy data, it is ensured that the clustered regions have high internal consistency. By scoring the similarity of the moving targets, similar targets are classified into one category to form a preliminary candidate region of the moving target, which is used to calibrate the possible target objects.

[0149] Step S44: performing a moving target boundary refinement process based on shape analysis and contour extraction algorithm on the preliminary moving target candidate region to generate accurate moving target contour data;

[0150] The embodiment of the present invention performs a refinement process on the boundary of the preliminary moving target candidate area based on shape analysis and contour extraction algorithm. The preliminary boundary of the moving target is extracted by a contour detection algorithm (such as Canny edge detection), and then a shape analysis method, such as Hough transform or shape context algorithm, is used to analyze the detected boundary to remove possible noise and pseudo contours. In order to further enhance the accuracy, the target contour is adjusted using a shape constraint model to ensure that the final contour is consistent with the shape of the actual moving object, and finally generate accurate moving target contour data.

[0151] Step S45: Determine the seed point and growth direction of region growth according to the precise moving target contour data and the segmentation strategy data, and perform adaptive region growing processing according to the region growing rule in the segmentation strategy data, so as to obtain preliminary segmentation region data.

[0152] The embodiment of the present invention determines the seed point and growth direction of regional growth based on the precise moving target contour data and the segmentation strategy data. First, by analyzing the precise moving target contour data, the center point or other representative pixels in the target contour are selected as the seed point of regional growth. Then, according to the regional growth rules in the segmentation strategy data, combined with the boundary extension algorithm, the growth area is gradually expanded outward. During the regional growth process, the growth direction and rate are dynamically adjusted according to the similarity and boundary conditions to ensure that the target integrity is maintained in the grown area, and finally the preliminary segmentation area data is obtained for subsequent further processing.

[0153] The present invention can accurately capture the changes in the displayed content on the LCD screen, especially the subtle movement of the moving target, through high-frequency sampling; this sampling method can update the image data in time, ensure that the subsequent analysis is based on the latest displayed content, thereby ensuring the real-time and accuracy of the moving target detection. The foreground extraction algorithm can effectively separate the static background and the dynamic foreground, so that the moving target features can be extracted independently. Optical flow analysis and inter-frame difference can identify the actual moving part in the foreground, and extract feature information such as the motion trajectory and direction, so as to generate accurate moving target data. This separation and feature extraction method improves the recognition accuracy of the moving target and reduces background interference. By combining the moving target data and the regional similarity evaluation rules in the segmentation strategy, the target areas with similar motion characteristics can be clustered together. This clustering process helps to preliminarily determine the possible moving target area, reduces the risk of misjudgment, and provides a more directional candidate area for the subsequent boundary refinement process. The shape analysis and contour extraction algorithm can further refine the boundary of the preliminary moving target. Through these algorithms, the boundary of the moving target can be clearer, avoiding the problem of blurred or irregular boundaries. This refinement process can improve the accuracy of the moving target boundary, making the outline of the moving target closer to reality, which is suitable for subsequent region segmentation and growth processing. Through accurate seed point selection and determination of growth direction, region growth can be expanded in a targeted manner. The adaptive region growing rule can adaptively adjust the growth process according to the shape and characteristics of the moving target, ensuring that the segmented area can be reasonably extended to the complete moving target area to avoid omission or over-segmentation. This adaptive region growing method helps to accurately identify the complete boundary and internal area of ​​the moving target. These steps can accurately capture dynamic moving targets on the LCD screen and perform precise segmentation through high-frequency sampling, foreground extraction, moving target clustering, boundary refinement processing and region growing processing. In particular, the adaptive region growing processing ensures the complete segmentation of the moving target, making the final segmentation result have higher accuracy and consistency. This series of steps has significant advantages in target detection and segmentation applications in dynamic scenes.

[0154] Preferably, step S5 comprises the following steps:

[0155] Step S51: extracting contour information of the segmented area from the preliminary segmented area data, and performing sub-pixel level refinement processing according to the screen response characteristic data to generate high-precision edge contour data;

[0156] The embodiment of the present invention extracts contour information from the preliminary segmented area data, and uses a traditional edge detection algorithm such as the Sobel or Canny algorithm to first perform edge detection on the segmented area. The detected edge is extracted with complete contour information through a contour tracking algorithm (such as a chain code or a Freeman chain code). Thereafter, the contour is refined at the sub-pixel level according to the screen response characteristic data of the LCD screen (such as grayscale response time, dynamic contrast), and an interpolation technique or an edge refinement algorithm (such as sub-pixel edge detection) is used to improve the spatial resolution of the edge, thereby generating high-precision edge contour data to ensure that the contour details are more accurate and clear.

[0157] Step S52: constructing an edge smoothing model based on curvature and direction according to the high-precision edge contour data to locally optimize the edge and obtain smoothed edge data;

[0158] The embodiment of the present invention constructs an edge smoothing model based on curvature and direction according to high-precision edge contour data, and uses Gaussian smoothing or bilateral filters to perform local smoothing on the edge. The curvature is calculated by the geometric characteristics of the contour, such as calculating the radius of curvature on the contour to ensure that the edge features are retained in places with large curvatures, while strong smoothing is performed in places with small curvatures. This model flexibly adjusts the smoothing intensity according to the local curvature and directionality of the edge to reduce edge noise and jaggedness, and finally generates smoothed edge data, making the edge more natural and smooth.

[0159] Step S53: performing a regional morphological operation to eliminate holes and protrusions according to the smoothed edge data and the preliminary segmented region data to generate morphologically optimized data;

[0160] The embodiment of the present invention uses morphological operations to further optimize the segmented area based on the smoothed edge data. In the specific implementation, morphological operations such as dilation and erosion are used to eliminate small holes and protrusions. The small noise and burrs on the edge are first removed by the erosion operation, and then the normal edge shape is restored by the dilation operation. According to the morphological operation results, the segmented area is corrected in morphological structure to ensure the integrity and smoothness of the edge of the area, and finally the morphological optimization data is generated to eliminate irregular noise and unnecessary details in the area.

[0161] Step S54: performing texture consistency analysis based on the morphological optimization data and the real-time image data, and locally adjusting the segmentation boundary, thereby obtaining texture optimized segmentation data;

[0162] The embodiment of the present invention performs texture consistency analysis on the morphological optimization data, analyzes the texture difference between the segmented area and its surrounding background, combines the texture features extracted from the real-time image data, and uses texture analysis algorithms such as Gabor filter or wavelet transform to fine-tune the boundary of the segmented area, especially at the boundary where there is a texture mutation, to ensure smooth texture transition between the edge and the adjacent area. This process makes local adjustments based on the consistency of the texture, optimizes the boundary of the texture transition area, and finally generates texture optimized segmentation data, making the segmentation result more natural and accurate.

[0163] Step S55: performing a segmentation quality evaluation on the texture optimized segmentation data based on edge accuracy, regional consistency, and consistency with the moving target, and identifying areas with segmentation quality less than a preset value, marking them as areas to be optimized, thereby obtaining segmentation quality evaluation data;

[0164] The embodiment of the present invention performs segmentation quality assessment on texture optimized segmentation data, firstly calculates the accuracy of edges, the consistency of regions and the degree of fit with moving target data, and uses a scoring-based assessment model (such as Jaccard index, Dice coefficient, etc.) for quantitative evaluation. For regions whose segmentation quality is lower than the preset standard identified in the assessment, they are automatically marked as regions to be optimized. These regions may be inaccurately segmented due to complex textures, dynamic blur or segmentation errors. Segmentation quality assessment data is generated through these assessment data to provide a basis for subsequent optimization.

[0165] Step S56: Optimizing the segmentation results based on the local threshold adjustment and the region merging and splitting operations on the segmentation quality evaluation data, thereby obtaining the final segmentation result data.

[0166] According to the segmentation quality evaluation data, the embodiment of the present invention implements an optimization operation based on local threshold adjustment for the area to be optimized. First, the boundaries of these areas are recalculated, and the segmentation boundaries are optimized by adjusting the local threshold. For adjacent similar areas, the region merging algorithm is used to merge them into a unified segmentation area; for areas with significant differences inside, the region splitting algorithm is used to perform subdivision processing to ensure consistency within the area. Through these local optimization measures, the final segmentation result data after the segmentation result optimization is finally generated.

[0167] The present invention significantly improves the positioning accuracy of the segmentation edge through sub-pixel level refinement processing. The subtle changes that may exist in the response process of the liquid crystal screen, such as grayscale response and dynamic contrast, will affect the accuracy of the edge contour. Sub-pixel level processing can capture these subtle features and generate high-precision edge contour data, thereby reducing edge blur and discontinuity. The smoothing model based on curvature and direction can effectively process the jaggedness or irregularity of the edge, making the segmentation edge smoother and more natural. Especially when processing high dynamic contrast and motion blur, this smoothing process can reduce the influence of noise and irregular edges and enhance the visual consistency of the segmentation result. The small holes and irregular protrusions in the segmentation area are eliminated, and the coherence and integrity of the area are enhanced. These morphological operations help to improve the geometric integrity of the segmentation area, especially in complex moving targets or multi-layer structures, and can effectively remove unnecessary interference and enhance the stability of the target area. The segmentation boundary can be adjusted according to the texture features of the target area to ensure that the segmentation boundary is consistent with the actual image content. The effect of this step is that by analyzing the texture changes in the area, the position of the segmentation boundary can be further optimized, so that the segmentation result is more consistent with the natural boundary of the target area and avoids unnecessary boundary errors. An automated evaluation mechanism for segmentation quality is provided. By comprehensively evaluating edge accuracy, regional consistency, and target fit, the region that needs further optimization is identified. This not only improves the overall segmentation accuracy, but also avoids the impact of local region segmentation quality differences, providing a target region for final optimization. The shape and boundary of the segmented region are further optimized through local threshold adjustment and region merging / splitting operations, making the final segmentation result more consistent with the actual target characteristics. The adjustment of the local threshold can accurately control the segmentation of the boundary, while the region merging and splitting operations help to deal with irregular parts in the segmented region, ensuring that the structure and morphology of the segmented region are more reasonable. These steps can effectively improve the accuracy and consistency of the segmented region through fine edge extraction, optimization processing, and quality evaluation, and ultimately generate high-quality segmentation results. In particular, the combination of sub-pixel processing, morphological operations, texture analysis and other technologies ensures that the segmentation boundary of the moving target is accurate and coherent, which is suitable for high-demand visual inspection tasks.

[0168] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0169] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An image segmentation method for liquid crystal display, characterized in that: The following steps are involved: Step S1: acquiring basic data of the LCD screen; performing screen response characteristic analysis based on the basic data of the LCD screen to obtain screen response characteristic data; performing multi-scale image acquisition based on the basic data of the LCD screen to generate multi-scale image data; The characteristics of the LCD screen are classified according to the screen response characteristic data, and edge compensation analysis is performed on the multi-scale image data to obtain edge feature compensation data; Step S2: acquiring real-time data of the touch screen; performing touch interference recognition analysis on the real-time data of the touch screen according to the edge feature compensation data, and generating touch interference mask data, wherein the touch interference includes pressure deformation interference, electric field interference, refresh rate interference and timing interference; Step S3: performing edge feature correction processing on the edge feature compensation data according to the touch interference mask data, and performing adaptive threshold segmentation to obtain preliminary image segmentation data; Performing time-series consistency constraint processing on the preliminary image segmentation data according to the screen response characteristic data, thereby generating segmentation strategy data; Step S4: collecting real-time image data of the LCD screen, and detecting moving targets based on the real-time image data to obtain moving target data; Performing region growing processing according to the segmentation strategy data and the moving target data, thereby obtaining preliminary segmentation region data; Step S5: Perform edge refinement processing on the preliminary segmented area data, evaluate the segmentation quality, and optimize the segmentation results to obtain the final segmentation result data.

2. The image segmentation method for liquid crystal display according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: collecting information on the size, resolution and pixel density of the LCD screen, thereby obtaining screen physical property data; Step S12: acquiring electrical parameters of the liquid crystal display screen based on the driving voltage and the refresh rate, thereby obtaining screen electrical characteristic data; Step S13: merging the screen electrical characteristic data and the screen physical characteristic data into LCD screen basic data; Step S14: analyzing the screen response characteristics according to the basic data of the liquid crystal screen, thereby obtaining the screen response characteristic data; Step S15: designing a multi-scale sampling scheme according to the screen physical characteristic data, and using the sampling parameter data to perform multi-scale image acquisition on the LCD display content to generate multi-scale image data; Step S16: classifying the characteristics of the liquid crystal screen according to the screen response characteristic data, and performing dynamic edge compensation analysis on the multi-scale image data to obtain edge feature compensation data.

3. The image segmentation method for liquid crystal display according to claim 2, characterized in that: Step S14 includes the following steps: Step S141: performing a grayscale switching response time test according to basic data of the LCD screen, thereby obtaining grayscale response time data; Step S142: performing dynamic contrast analysis based on the grayscale response time data to generate dynamic contrast characteristic data; Step S143: testing and analyzing the overshoot and undershoot phenomena of pixels according to the electrical characteristic data of the screen to obtain pixel stability data; Step S144: constructing a motion blur characteristic model according to the grayscale response time data and the dynamic contrast characteristic data, thereby generating motion blur characteristic data; Step S145: performing inter-frame interpolation capability analysis according to the refresh rate information in the screen electrical characteristic data, thereby obtaining frame interpolation characteristic data; Step S146: performing data integration based on high-dimensional feature fusion on the pixel stability data, frame interpolation characteristic data, motion blur characteristic data and grayscale response time data to obtain screen response characteristic data.

4. The image segmentation method for liquid crystal display according to claim 3, characterized in that: Step S16 includes the following steps: Step S161: establishing a screen response characteristic classification model according to the grayscale response time data, dynamic contrast characteristic data and pixel stability data in the screen response characteristic data; Step S162: Identify the type of the liquid crystal screen according to the screen response characteristic classification model, thereby generating screen type identification data, wherein the screen type identification data includes a high grayscale response time type, a high dynamic contrast type, an unstable pixel type, and a standard response type; Step S163: selecting an edge detection algorithm according to the screen type identification data, and extracting edge features from the multi-scale image data using the selected edge detection algorithm, thereby obtaining preliminary edge feature data, wherein the algorithm selection is specifically: for a screen whose screen type identification data is a high grayscale response time type, the selected algorithm is the Canny edge detection algorithm; for a screen whose screen type identification data is a high dynamic contrast type, the selected algorithm is the Sobel algorithm; for a screen whose pixel is unstable type, the selected algorithm is the Laplacian algorithm; and for a screen whose screen type is a standard response type, the selected algorithm is the Prewitt algorithm; Step S164: construct a dynamic edge blur compensation model according to the screen response characteristic data, and use the dynamic edge blur compensation model to perform edge compensation analysis on the preliminary edge feature data, so as to obtain edge feature compensation data.

5. The image segmentation method for liquid crystal display according to claim 4, characterized in that: Step S2 includes the following steps: Step S21: collecting the touch coordinates, pressure value and touch area information of the touch screen in real time, so as to obtain real-time data of the touch screen; Step S22: identifying and extracting the pressure deformation interference area according to the real-time data of the touch screen, thereby generating pressure deformation interference data; Step S23: analyzing the electric field changes around the touch point according to the touch coordinate information in the real-time data of the touch screen, thereby obtaining electric field interference data; Step S24: Analyze the timing relationship between touch operation and screen refresh using the refresh rate information in the basic data of the LCD screen, so as to generate refresh rate influence interference data; Step S25: analyzing the response delay of the touch operation according to the real-time data of the touch screen and the edge feature compensation data, thereby obtaining response delay interference data; Step S26: analyzing the afterimage effect caused by the touch operation on the touch screen real-time data and the screen response characteristic data, thereby generating afterimage effect interference data; Step S27: performing spatial superposition and temporal correlation analysis on the pressure deformation interference data, the electric field interference data, the refresh rate interference data, the response delay interference data and the afterimage effect interference data, so as to obtain comprehensive touch interference data; Step S28: establishing a touch interference space-time model according to the comprehensive touch interference data, and generating preliminary touch interference mask data; Step S29: using the edge feature compensation data to perform edge alignment and refinement processing on the preliminary touch interference mask data, thereby obtaining the touch interference mask data.

6. The image segmentation method for liquid crystal display according to claim 5, characterized in that: Step S3 includes the following steps: Step S31: performing mask processing on the edge feature compensation data according to the touch interference mask data, so as to obtain preliminary edge feature correction data; Step S32: using the grayscale response time data in the screen response characteristic data to perform time domain compensation on the preliminary edge feature correction data, thereby obtaining time compensated edge feature data; Step S33: performing contrast enhancement processing on the time-compensated edge feature data, and performing noise suppression processing according to the pixel stability data, so as to obtain final edge feature correction data; Step S34: constructing an adaptive threshold model based on local image features and global brightness distribution according to the final edge feature correction data, and performing segmentation processing on the final edge feature correction data according to the adaptive threshold model, thereby obtaining preliminary image segmentation data; Step S35: performing time consistency constraint processing on the preliminary image segmentation data according to the screen response characteristic data, thereby generating segmentation strategy data.

7. The image segmentation method for liquid crystal display according to claim 6, characterized in that: Step S35 includes the following steps: Step S351: establishing a temporal consistency constraint model according to the dynamic contrast characteristic data and the motion blur characteristic data in the screen response characteristic data; Step S352: using the temporal consistency constraint model to perform consistency analysis on the cross-frame segmentation results of the preliminary image segmentation data, thereby obtaining temporal correction segmentation data; Step S353: performing refinement processing on the segmentation boundary according to the timing correction segmentation data and the screen response characteristic data, thereby obtaining optimized segmentation data; Step S354: constructing a segmentation region growing rule set according to the optimized segmentation data, wherein the segmentation region growing rule set includes a region similarity evaluation and a boundary extension strategy; Step S355: Optimizing the segmentation strategy of the optimized segmentation data according to the segmentation region growing rule set, thereby generating segmentation strategy data.

8. The image segmentation method for liquid crystal display according to claim 7, characterized in that: Step S4 includes the following steps: Step S41: using the image acquisition device of the LCD screen to perform high-frequency sampling on the current display content, thereby obtaining real-time image data; Step S42: using a foreground extraction algorithm to perform preliminary separation of static background and dynamic foreground on the real-time image data, and extracting motion features of the separated dynamic foreground based on optical flow analysis and inter-frame difference, thereby obtaining moving target data; Step S43: clustering the moving targets according to the moving target data and the regional similarity evaluation rules in the segmentation strategy data, thereby obtaining preliminary moving target candidate regions; Step S44: performing a moving target boundary refinement process based on shape analysis and contour extraction algorithm on the preliminary moving target candidate region to generate accurate moving target contour data; Step S45: Determine the seed point and growth direction of region growth according to the precise moving target contour data and the segmentation strategy data, and perform adaptive region growing processing according to the region growing rule in the segmentation strategy data, so as to obtain preliminary segmentation region data.

9. The image segmentation method for liquid crystal display according to claim 8, characterized in that: Step S5 includes the following steps: Step S51: extracting contour information of the segmented area from the preliminary segmented area data, and performing sub-pixel level refinement processing according to the screen response characteristic data to generate high-precision edge contour data; Step S52: constructing an edge smoothing model based on curvature and direction according to the high-precision edge contour data to locally optimize the edge and obtain smoothed edge data; Step S53: performing a regional morphological operation to eliminate holes and protrusions according to the smoothed edge data and the preliminary segmented region data to generate morphologically optimized data; Step S54: performing texture consistency analysis based on the morphological optimization data and the real-time image data, and locally adjusting the segmentation boundary, thereby obtaining texture optimized segmentation data; Step S55: performing a segmentation quality evaluation on the texture optimized segmentation data based on edge accuracy, regional consistency, and consistency with the moving target, and identifying areas with segmentation quality less than a preset value, marking them as areas to be optimized, thereby obtaining segmentation quality evaluation data; Step S56: Optimizing the segmentation results based on local threshold adjustment and region merging and splitting operations on the segmentation quality assessment data, thereby obtaining final segmentation result data.

10. An image segmentation system for liquid crystal display, characterized in that: Used to execute the image segmentation method for liquid crystal display according to claim 1, the image segmentation system for liquid crystal display comprises: The screen characteristic analysis module is used to obtain basic data of the LCD screen; perform screen response characteristic analysis based on the basic data of the LCD screen to obtain screen response characteristic data; perform multi-scale image acquisition based on the basic data of the LCD screen to generate multi-scale image data; perform characteristic classification of the LCD screen based on the screen response characteristic data, and perform edge compensation analysis on the multi-scale image data respectively to obtain edge feature compensation data; A touch interference recognition module is used to obtain real-time data of the touch screen; perform touch interference recognition analysis on the real-time data of the touch screen according to the edge feature compensation data, and generate touch interference mask data, wherein the touch interference includes pressure deformation interference, electric field interference, refresh rate interference and timing interference; A preliminary segmentation module is used to perform edge feature correction processing on edge feature compensation data according to touch interference mask data, and perform adaptive threshold segmentation to obtain preliminary image segmentation data; perform temporal consistency constraint processing on the preliminary image segmentation data according to screen response characteristic data, thereby generating segmentation strategy data; The moving target detection module is used to collect real-time image data of the LCD screen, and perform moving target detection based on the real-time image data to obtain moving target data; perform region growing processing based on the segmentation strategy data and the moving target data to obtain preliminary segmentation region data; The edge refinement module is used to perform edge refinement processing on the preliminary segmentation area data, evaluate the segmentation quality, and optimize the segmentation results to obtain the final segmentation result data.

Images