Liquid crystal screen abnormal image analysis method and system

Abstract

The application discloses a liquid crystal screen display abnormal image analysis method and system, and belongs to the technical field of image recognition, and comprises the following steps: S1, obtaining a liquid crystal screen optimized display picture; S2, setting a first specification sliding window and a second specification sliding window to respectively perform sliding detection on the liquid crystal screen optimized display picture, obtaining a pixel block abnormality judgment result, if the pixel block abnormality judgment result is qualified, counting the qualified pixel blocks and storing the qualified pixel blocks into a qualified image temporary storage area, if the pixel block abnormality judgment result is abnormal, counting the abnormal pixel blocks and storing the abnormal pixel blocks into an abnormal image temporary storage area, and if the pixel block abnormality judgment result is conflict, performing conflict adjustment; S3, obtaining each abnormal pixel clustering block and corresponding abnormal parameters of each abnormal pixel clustering block, thereby obtaining a fault type of the liquid crystal screen display abnormal image, avoiding collection interference caused by improper image collection equipment parameters, and solving the problem that defects are not easy to be successfully detected due to image noise interference in the prior art.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, and in particular to a method and system for analyzing abnormal images displayed on a liquid crystal display screen. Background Technology

[0002] Existing LCD screen abnormal display detection systems achieve screen fault detection through image analysis combined with modeling and prediction. First, candidate defect locations in the screen image are identified. Then, an autoregressive analysis model based on the candidate defect locations is used to calculate the expected distribution of these locations, and the expected distribution is used to determine whether it is a display fault point.

[0003] For example, the screen display fault detection method, device, equipment, and storage medium disclosed in Chinese invention patent with announcement number CN119228713A includes: determining the candidate defect locations of the screen image to be processed; determining the expected distribution of the candidate defect locations through a preset display fault detection model, wherein the preset display fault detection model is an autoregressive analysis model established with the candidate defect locations in the image as independent variables; and determining the display fault points of the screen displaying the screen image to be processed based on the expected distribution.

[0004] For example, Chinese invention patent CN113379752B discloses an image segmentation method for dual-layer LCD screen display, which includes: first, determining the physical resolution of the front and rear LCD screens and the resolution of the input image; then, depending on the situation, selecting whether to perform upsampling or downsampling processing on the input image; if the resolution of the input image is the same as the physical resolution of the front LCD screen, then no processing is performed on the input image; then, arranging the gray values ​​of all pixels in the input image into a 1-dimensional vector in order from top to bottom and from left to right, and constructing a matrix Sθ with a viewing angle θ to record the position of light passing through the front and rear LCD screens; then constructing a 1-dimensional vector X of the gray values ​​of all pixels in the front and rear LCD screens; initially assigning random values ​​to all variables in vector X; calculating the product of Sθ and X to obtain a 1-dimensional vector Cθ; using the LSQLIN operator to solve for each variable in vector X; and inputting the variables in X into the dual-layer LCD screens to display the final image.

[0005] However, in the process of implementing the inventive technical solution in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems:

[0006] Existing technologies ignore noise interference factors during image acquisition, such as high-frequency stripe interference and isolated bright spots that may be generated by image acquisition equipment during imaging. However, these interferences can mask the true defect features or misjudge non-defect areas as abnormal areas, seriously affecting the accuracy and stability of defect detection. Therefore, there is a problem that defects are not easy to be successfully detected due to image noise interference. Summary of the Invention

[0007] To address the technical problem in existing technologies where defects are difficult to detect due to image noise interference, embodiments of the present invention provide a method and system for analyzing abnormal images of liquid crystal displays. The technical solution is as follows:

[0008] On the one hand, a method for analyzing abnormal images of an LCD screen is provided. The method includes: S1, acquiring an image of the LCD screen display through an image acquisition device and optimizing the image to obtain an optimized display image of the LCD screen; S2, setting a first-size sliding window and a second-size sliding window to perform sliding detection on the optimized display image of the LCD screen respectively, and obtaining the pixel block abnormality judgment result. If the pixel block abnormality judgment result is qualified, the qualified pixel blocks are counted and stored in the qualified image temporary storage area. If the pixel block abnormality judgment result is abnormal, the abnormal pixel blocks are counted and stored in the abnormal image temporary storage area. If the pixel block abnormality judgment result is conflicting, conflict adjustment is performed; S3, acquiring the abnormal pixel blocks in the abnormal image temporary storage area, and processing them to obtain each abnormal pixel cluster block and the abnormal parameters corresponding to each abnormal pixel cluster block, thereby obtaining the fault type of the abnormal image of the LCD screen display.

[0009] On the other hand, a system for analyzing abnormal images of an LCD screen is provided. This system includes an image optimization and adjustment module, an anomaly determination module, and a fault type analysis module. The image optimization and adjustment module acquires images of the LCD screen through an image acquisition device and optimizes them to obtain an optimized display image. The anomaly determination module sets up a first-size sliding window and a second-size sliding window to perform sliding detection on the optimized display image, obtaining pixel block anomaly determination results. If the pixel block anomaly determination result is qualified, qualified pixel blocks are counted and stored in a qualified image temporary storage area; if the pixel block anomaly determination result is abnormal, abnormal pixel blocks are counted and stored in an abnormal image temporary storage area; if the pixel block anomaly determination result is conflicting, conflict adjustment is performed. The fault type analysis module acquires abnormal pixel blocks in the abnormal image temporary storage area and processes them to obtain each abnormal pixel cluster and its corresponding abnormal parameters, thereby determining the fault type of the abnormal LCD screen display image.

[0010] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0011] 1. The LCD screen display abnormality image analysis method provided by the present invention obtains and analyzes the image acquisition interference parameters, thereby dynamically adjusting the image gain value to obtain an optimized display image of the LCD screen, thus avoiding image acquisition interference caused by improper setting of image acquisition device parameters, and effectively solving the problem in the prior art that defects are not easy to be successfully detected due to image noise interference.

[0012] 2. This invention performs multi-scale sliding detection on the optimized display image of the LCD screen by setting two sliding windows of different sizes, thereby obtaining different types of abnormal pixel blocks, thus achieving accurate acquisition of the pixel block abnormality judgment result and avoiding the problem of abnormal omission or misjudgment caused by single window size detection.

[0013] 3. By merging connected regions of abnormal pixel blocks using a clustering algorithm based on spatial connectivity and obtaining abnormal parameters based on coordinate position analysis, accurate identification of abnormal pixel cluster blocks is achieved, effectively improving the accuracy of spatial localization and feature extraction of abnormal regions.

[0014] 4. This invention achieves intelligent identification of fault types by using a pre-set LCD screen display anomaly model based on a database, combined with abnormal pixel clusters and corresponding abnormal parameters as input variables, thereby improving the efficiency and accuracy of fault diagnosis of abnormal LCD screen images. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart of the LCD screen display abnormality image analysis method provided in the embodiments of this application;

[0017] Figure 2 Macroscopic determination flowchart of the LCD screen display abnormality image analysis method provided in the embodiments of this application;

[0018] Figure 3 A flowchart illustrating the conflict adjustment process of the LCD screen display abnormality image analysis method provided in this application embodiment;

[0019] Figure 4 A first example diagram of image acquisition for the LCD screen display anomaly image analysis system provided in this application embodiment;

[0020] Figure 5 A second example diagram of image acquisition for the LCD screen display anomaly image analysis system provided in this application embodiment;

[0021] Figure 6 This is a first example diagram of pixel defect results from the LCD screen display anomaly image analysis system provided in this application embodiment;

[0022] Figure 7 This is a magnified partial image of the first example of the pixel defect results of the LCD screen display anomaly image analysis system provided in this application embodiment;

[0023] Figure 8 This is a second example diagram of pixel defect results from the LCD screen display anomaly image analysis system provided in this application embodiment;

[0024] Figure 9 This is a partially enlarged image of the second example diagram of the pixel defect results of the LCD screen display anomaly image analysis system provided in this application embodiment;

[0025] Figure 10 This is a schematic diagram of the structure of the LCD screen display abnormality image analysis system provided in an embodiment of this application. Detailed Implementation

[0026] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0027] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0028] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0029] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0030] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0031] like Figure 1The diagram shows a flowchart of an abnormal LCD screen image analysis method provided in this application embodiment. The method includes the following steps: S1, acquiring an LCD screen display image through an image acquisition device and optimizing the image to obtain an optimized LCD screen display image; S2, setting a first-size sliding window and a second-size sliding window to perform sliding detection on the optimized LCD screen display image respectively, obtaining pixel block abnormality judgment results. If the pixel block abnormality judgment result is qualified, the qualified pixel blocks are counted and stored in the qualified image temporary storage area. If the pixel block abnormality judgment result is abnormal, the abnormal pixel blocks are counted and stored in the abnormal image temporary storage area. If the pixel block abnormality judgment result is conflicting, conflict adjustment is performed; S3, acquiring the abnormal pixel blocks in the abnormal image temporary storage area and processing them to obtain each abnormal pixel cluster block and the abnormal parameters corresponding to each abnormal pixel cluster block, thereby obtaining the fault type of the abnormal LCD screen display image.

[0032] In this embodiment, as Figure 1 As shown, Figure 1 A flowchart of a method for analyzing abnormal images of a liquid crystal display provided in an embodiment of this application. The process begins by acquiring images displayed on the LCD screen and optimizing them to obtain an optimized display image. This optimized image is then subjected to sliding detection analysis to identify any anomalies, resulting in pixel block anomaly determinations. These results are categorized as qualified, abnormal, or conflicting. Qualified pixel blocks are counted and stored in the qualified image temporary storage area, while abnormal pixel blocks are counted and stored in the abnormal image temporary storage area. If a pixel block anomaly determination result is a conflict, conflict adjustment is performed to obtain a conflict-adjusted pixel block anomaly determination result, including both abnormal and qualified pixel blocks. If a pixel block is determined to be abnormal, it is stored in the abnormal image temporary storage area; if it is deemed qualified, it is saved in the qualified image temporary storage area. The abnormal pixel blocks in the abnormal image temporary storage area are then counted and analyzed to obtain clusters of abnormal pixels and their corresponding anomaly parameters. These clusters and their corresponding anomaly parameters are then input into the LCD screen display anomaly model, outputting the specific fault type and completing the intelligent fault diagnosis and location of the LCD screen anomaly image.

[0033] This invention improves image quality consistency and pre-processing effectiveness by comparing image acquisition device parameters with an ideal execution set to obtain image acquisition interference values ​​and adjust image gain. It also enhances anomaly detection accuracy by using a dual-size sliding window for image scanning and combining abnormal feature values ​​with preset thresholds for pixel-level judgment. Furthermore, it merges connected regions of abnormal pixel blocks using a spatial connectivity clustering algorithm and extracts anomaly parameters, including length, width, and area, based on the cluster block's location coordinates, providing a basis for subsequent fault type identification. Finally, it uses the abnormal cluster blocks and their parameters as input, combined with a preset LCD screen anomaly model, to output specific fault types, achieving structured representation and automated classification of abnormal LCD screen images, thus improving detection accuracy, stability, and intelligence.

[0034] Furthermore, the optimized display image for the LCD screen is obtained through the following methods: Image acquisition interference parameters are acquired, including the focal length, exposure time, frame rate, and ISO sensitivity of the image acquisition device; a preset ideal execution set for image acquisition is obtained from the database, and its difference from the image acquisition interference parameters is analyzed to obtain the image acquisition difference analysis result; based on the image acquisition difference analysis result, corresponding weighting factors are introduced for coupling processing to obtain the image acquisition interference value; the ideal execution set for image acquisition includes ideal execution values ​​for focal length, exposure time, frame rate, and ISO sensitivity; the image acquisition interference value is matched with the database to obtain the image gain adjustment execution coefficient; based on the image gain adjustment execution coefficient, the image gain value of the image displayed on the LCD screen is reduced, thereby obtaining the optimized display image for the LCD screen.

[0035] In this embodiment, the focal length, exposure time, frame rate, and ISO sensitivity of the image acquisition device can be obtained through the LCD screen display detection backend system. After the device parameters of the image acquisition device are set, they need to be uploaded to the LCD screen display detection backend system for unified storage.

[0036] The specific method for obtaining image acquisition interference values ​​is as follows:

[0037] ;

[0038] In the formula, TG represents the image acquisition interference value, TJ represents the focal length of the image acquisition device, GJ represents the ideal focal length, TH represents the exposure time of the image acquisition device, GH represents the ideal exposure time, TP represents the frame rate of the image acquisition device, GP represents the ideal frame rate, TD represents the ISO sensitivity of the image acquisition device, GD represents the ideal ISO sensitivity, ω1 represents the focal length weighting factor, ω2 represents the exposure time weighting factor, ω3 represents the frame rate weighting factor, and ω4 represents the ISO sensitivity weighting factor.

[0039] Focal length weighting factors, exposure time weighting factors, frame rate weighting factors, and ISO sensitivity weighting factors can be obtained from a database. For example, the focal length weighting factor can be obtained by analyzing the historical focal length set stored in the database. The difference between each historical focal length in the set and the focal length of the image acquisition device is processed to obtain the historical focal length difference value. A preset focal length difference threshold range is obtained from the database and compared with each historical focal length difference value. If a historical focal length difference value falls within the threshold range, the corresponding historical focal length is obtained and marked as the historical reference focal length, thus obtaining each historical reference focal length. The extreme values ​​(maximum and minimum values) of each historical reference focal length are then removed, and the standard deviation is taken to obtain the standard deviation of the historical reference focal length. The system retrieves the preset historical reference focal length weighting factor, historical control focal length standard deviation benchmark, historical control standard deviation gradient, and single-level adjustment amount of the focal length weighting factor from the database. It then performs difference processing between the historical control focal length standard deviation and the historical control focal length standard deviation benchmark to obtain the mean of the historical control focal length standard deviation. Next, it performs a fold analysis (dividing the historical control focal length standard deviation mean by the historical control standard deviation gradient) to obtain the historical control gradient fold. Finally, it multiplies the historical control gradient fold by the single-level adjustment amount of the focal length weighting factor to obtain the comprehensive adjustment amount of the focal length weighting factor. Finally, it compares the historical control focal length standard deviation with the historical control focal length standard deviation benchmark. If the historical control focal length standard deviation is greater than the historical control focal length standard deviation benchmark, the system adds the historical reference focal length weighting factor to the comprehensive adjustment amount of the focal length weighting factor to obtain the focal length weighting factor; otherwise, it subtracts the historical reference focal length weighting factor from the comprehensive adjustment amount of the focal length weighting factor to obtain the focal length weighting factor. The other weighting factors are obtained in the same way as the focal length weighting factor. The exposure time weighting factor corresponds to the mean standard deviation of the exposure time in the historical comparison, the frame rate weighting factor corresponds to the mean standard deviation of the frame rate in the historical comparison, and the ISO sensitivity weighting factor corresponds to the mean standard deviation of the ISO sensitivity in the historical comparison.

[0040] This analysis examines parameters influencing image acquisition interference, including the focal length, exposure time, frame rate, and ISO sensitivity of the image acquisition device. This analysis considers the interrelationships between these parameters. For example, the focal length, exposure time, frame rate, and ISO sensitivity of the image acquisition device mutually constrain and synergistically affect image quality during the imaging process. Focal length determines the magnification and field of view, indirectly affecting the amount of light entering the camera. Exposure time controls the time each frame receives light; a longer exposure time allows more light in, but also increases the likelihood of motion blur. A higher frame rate means acquiring more frames per unit time, requiring a shorter exposure time per frame to reduce the amount of light entering the camera. ISO sensitivity compensates for insufficient light; a higher value indicates greater sensitivity to light and can amplify image noise.

[0041] By acquiring interference parameters such as focal length, exposure time, frame rate, and ISO sensitivity of the image acquisition device and analyzing their differences with preset ideal execution values ​​in the database, the interference impact caused by fluctuations in device parameters during image acquisition can be quantitatively evaluated. This effectively avoids problems such as brightness drift, color shift, detail blurring, and image noise caused by improper image acquisition device settings. In particular, phenomena such as stripe interference, overexposure or underexposure, and motion blur directly affect the accuracy and stability of subsequent defect detection. If interference factors of the image acquisition device are not evaluated and optimized, the quality of the acquired image will be uncertain, leading to increased errors in anomaly detection results and a higher false negative rate, affecting the accurate identification of LCD display defects. Therefore, by introducing image acquisition interference values ​​and obtaining image gain adjustment execution coefficients, dynamic adjustment of image gain can be achieved, improving the accuracy and correctness of LCD image anomaly identification.

[0042] Further, the pixel block anomaly determination results are obtained through the following method: A first-size sliding window and a second-size sliding window are set, with the first-size sliding window being larger than the second-size sliding window; based on the first-size sliding window, sliding detection is performed on the optimized display image of the LCD screen according to a first preset step size, and the pixel blocks obtained from each sliding detection are jointly marked as first-size units, thus obtaining each first-size unit; based on the second-size sliding window, sliding detection is performed on the optimized display image of the LCD screen according to a second preset step size, and the pixel blocks obtained from each sliding detection are jointly marked as second-size units, thus obtaining each second-size unit; abnormal feature parameters are obtained, and abnormal feature values ​​are analyzed, including the first abnormal feature value of each first-size unit and the second abnormal feature value of each second-size unit; a preset abnormal feature threshold is obtained from the database, and the first abnormal feature value of each first-size unit is compared with the abnormal feature threshold to obtain the anomaly determination result of each first-size unit; the second abnormal feature value of each second-size unit is compared with the abnormal feature threshold to obtain the anomaly determination result of each second-size unit; based on the anomaly determination results of each first-size unit and each second-size unit, the pixel block anomaly determination results are obtained through analysis.

[0043] In this embodiment, both the first-size sliding window and the second-size sliding window include a plurality of pixel blocks, wherein a pixel block refers to the smallest pixel unit. A first-size sliding window (e.g., a 16×16 pixel block) and a second-size sliding window (e.g., a 16×4 pixel block) are configured, wherein the width of the first-size sliding window is the same as the width of the second-size sliding window.

[0044] The abnormality determination result of each first specification unit is obtained by comparing the first abnormal feature value of each first specification unit with the abnormal feature threshold. The specific method is as follows: if the first abnormal feature value of a certain first specification unit is above the abnormal feature threshold, the abnormality determination result of the first specification unit is abnormal; otherwise, the abnormality determination result of the first specification unit is normal.

[0045] The abnormality determination result of each second specification unit is obtained by comparing the second abnormal feature value of each second specification unit with the abnormal feature threshold. The specific method is as follows: if the second abnormal feature value of a certain second specification unit is above the abnormal feature threshold, the abnormality determination result of the second specification unit is abnormal; otherwise, the abnormality determination result of the second specification unit is normal.

[0046] The method of using a first-size and a second-size sliding window for sliding detection of optimized LCD display images is primarily aimed at balancing the overall recognition of large-scale anomalies with the precise detection of small-scale detail anomalies during image analysis. The first-size sliding window, due to its larger size, can effectively cover a wider area, facilitating the identification of large-scale display anomalies such as bright bands, dark bands, overall brightness deviations, or color casts. The second-size sliding window, with its smaller size, possesses higher local resolution, making it more suitable for detecting fine-grained defects such as dead pixels, minute bright spots, and edge breaks. By setting two sliding windows of different sizes and combining their respective step sizes for sliding detection, rather than using a fixed window, the comprehensiveness and sensitivity of the detection are improved, reducing blind spots or omissions caused by a single window size, and ensuring accurate coverage of defects of different sizes and shapes.

[0047] Furthermore, the abnormal feature values ​​are obtained. The specific method is as follows: obtain abnormal feature parameters, which include the mean brightness, standard deviation of brightness, mean color channel deviation, and mean edge intensity; obtain a preset abnormal feature standard set in the database and perform a difference analysis with the abnormal feature parameters to obtain the abnormal feature difference analysis results; based on the abnormal feature difference analysis results, introduce corresponding weighting factors for coupling processing to obtain abnormal feature values; the abnormal feature standard set includes the standard value of brightness, the allowable value of standard deviation of brightness, the allowable value of color channel deviation, and the standard value of edge intensity.

[0048] In this embodiment, obtaining abnormal feature parameters and analyzing abnormal feature values ​​refers to obtaining the abnormal feature parameters of each first specification unit and each second specification unit respectively, and then analyzing them to obtain the first abnormal feature value of each first specification unit and the second abnormal feature value of each second specification unit. The abnormal feature parameters include the mean luminance, the standard deviation of luminance, the mean color channel deviation, and the mean edge intensity.

[0049] The mean luminance value refers to the average difference between the luminance of each pixel block within a specification unit (such as the first specification unit) and the standard luminance value. The standard deviation of luminance refers to the standard deviation of the luminance of the pixel blocks within a specification unit (such as the first specification unit). The mean color channel deviation value refers to the average difference between the color channel deviation and the allowable color channel deviation value for the pixel blocks within a specification unit (such as the first specification unit). The mean edge intensity value refers to the average pixel edge intensity of the pixel blocks within a specification unit (such as the first specification unit).

[0050] The specific method for obtaining abnormal feature values ​​is as follows:

[0051] ;

[0052] In the formula, YCi represents the abnormal feature value, i represents the specification unit number, i=1,2, where i=1 represents the first abnormal feature value of the first specification unit, i=2 represents the second abnormal feature value of the second specification unit, YLi represents the luminance mean of the i-th specification unit, CL represents the luminance standard value, YBi represents the luminance standard deviation of the i-th specification unit, CB represents the allowable value of the luminance standard deviation of the i-th specification unit, r represents a constant, YTi represents the color channel deviation mean of the i-th specification unit, CT represents the allowable value of the color channel deviation, YQi represents the edge intensity mean of the i-th specification unit, CQ represents the edge intensity standard value, φ1 represents the luminance mean weighting factor, φ2 represents the luminance standard deviation weighting factor, φ3 represents the color channel deviation mean weighting factor, and φ4 represents the edge intensity mean weighting factor.

[0053] It should be noted that r represents a very small constant to avoid computer calculation errors that would occur when the allowable value of the brightness standard deviation is zero.

[0054] The luminance mean weighting factor, luminance standard deviation weighting factor, color channel deviation mean weighting factor, and edge intensity mean weighting factor can be obtained from a database. For example, the luminance mean weighting factor can be obtained by analyzing the historical luminance mean set stored in the database. The differences between each historical luminance mean in the historical luminance mean set and the luminance mean of the image acquisition device are processed to obtain the historical luminance mean difference. A preset luminance mean difference threshold range in the database is obtained and compared with each historical luminance mean difference. If a historical luminance mean difference falls within the threshold range, the corresponding historical luminance mean is obtained and marked as the historical control luminance mean, thus obtaining each historical control luminance mean. The extreme values ​​(maximum and minimum values) of each historical control luminance mean are then removed, and the standard deviation is taken to obtain the standard deviation of the historical control luminance mean. Retrieve the preset historical reference luminance mean weighting factor, historical control luminance mean standard deviation benchmark, historical control standard deviation gradient, and single-level adjustment of the luminance mean weighting factor from the database. Difference is processed between the historical control luminance mean standard deviation and the historical control luminance mean standard deviation benchmark to obtain the mean of the historical control luminance mean standard deviation. Magnitude analysis is performed on the historical control luminance mean standard deviation and the historical control standard deviation gradient (dividing the historical control luminance mean standard deviation by the historical control standard deviation gradient) to obtain the historical control gradient magnitude. The historical control gradient magnitude is then compared with... The single-level adjustment amount of the luminance mean weighting factor is multiplied to obtain the comprehensive adjustment amount of the luminance mean weighting factor. Based on the comparison between the standard deviation of the historical control luminance mean and the benchmark value of the standard deviation of the historical control luminance mean, if the standard deviation of the historical control luminance mean is greater than the benchmark value, the historical reference luminance mean weighting factor is added to the comprehensive adjustment amount of the luminance mean weighting factor to obtain the luminance mean weighting factor; otherwise, the historical reference luminance mean weighting factor is subtracted from the comprehensive adjustment amount of the luminance mean weighting factor to obtain the luminance mean weighting factor.

[0055] By analyzing abnormal feature parameters including mean brightness, standard deviation of brightness, mean deviation of color channels, and mean edge intensity, the abnormal feature value is obtained considering the mutual influence relationship between these parameters. For example, the mean brightness determines the overall brightness level of the image, and the change in its value will directly affect the fluctuation degree of the standard deviation of brightness. For example, when the brightness is too low, the overall image is dark, the standard deviation of brightness tends to be small, showing uneven brightness distribution and a dull image; while when the brightness is too high, the standard deviation of brightness will increase, resulting in overexposure or local bright spots. The mean deviation of color channels reflects the difference in brightness of each RGB channel. If a certain channel is too strong, it will cause the mean brightness to shift, and at the same time cause local color deviation, further leading to edge information distortion, thus affecting the accuracy of the mean edge intensity; the mean edge intensity depends on the brightness gradient change and color contrast. When the standard deviation of brightness is too small or the color deviation is not obvious, the edge contour is blurred and the edge intensity is weakened; on the contrary, when the brightness and color contrast increase, the edge intensity rises, but if the contrast is too large, it will also amplify image noise, resulting in misjudgment.

[0056] Furthermore, the abnormal determination result of the pixel block is obtained, specifically including: obtaining the pixel blocks corresponding to each first specification unit. If there is a certain pixel block in different first specification units, and the multiple abnormal determination results corresponding to this pixel block show inconsistency, then the abnormal determination result of the first specification unit of this pixel block is the abnormal determination result obtained by the first sliding detection of this pixel block; obtaining the pixel blocks corresponding to each second specification unit. If there is a certain pixel block in different second specification units, and the multiple abnormal determination results corresponding to this pixel block show inconsistency, then the abnormal determination result of the second specification unit of this pixel block is the abnormal determination result obtained by the first sliding detection of this pixel block; counting the pixel blocks in the optimized display picture of the liquid crystal screen and the abnormal determination results of the first specification unit and the second specification unit corresponding to the pixel blocks. If there is a certain pixel block whose abnormal determination result of the first specification unit is normal and the abnormal determination result of the second specification unit is normal, then the abnormal determination result of this pixel block is qualified; if there is a certain pixel block whose abnormal determination result of the first specification unit is abnormal and the abnormal determination result of the second specification unit is abnormal, then the abnormal determination result of this pixel block is abnormal; if there is a certain pixel block whose abnormal determination results of the first specification unit and the second specification unit show inconsistency, then the abnormal determination result of this pixel block is in conflict.

[0057] In this embodiment, image detection is performed using a sliding window approach. During the detection process, due to the sliding characteristics, a pixel block is often repeatedly covered by sliding windows at multiple different positions, especially when the step size of the first or second specification window is small. A pixel block may belong to multiple adjacent specification units. This overlap can lead to different anomaly judgment results for the same pixel block in different specification units, resulting in inconsistencies in the abnormal state. To avoid instability or interference from subsequent detection results on the overall judgment, the judgment result obtained from the first sliding detection is taken as the standard, thereby ensuring the traceability and logical consistency of the judgment result and avoiding fluctuations introduced by changes in the detection order.

[0058] To ensure the accuracy of the final pixel block anomaly determination, this scheme introduces a dual-scale determination mechanism. This involves simultaneously analyzing the anomaly determination results for each pixel block in both the first and second scale units. By cross-validating the determination results at both scales, the scheme maximizes the coverage of both macroscopic and local detail anomalies, improving the overall robustness of the determination. Only when both scales show normal results can the pixel block be considered fully acceptable; only when both scales show anomalies can the pixel block be considered significantly abnormal and have sufficient confidence. When the determination results of the two scale windows are inconsistent, it indicates that the pixel block's anomaly features have issues with blurred edges or scale sensitivity. Directly drawing conclusions in this case may lead to misjudgments due to factors such as sliding window coverage edges and changes in image features. Therefore, this is set to a "conflict" state and enters a subsequent conflict adjustment process. Further analysis and adjustment of differential parameters can improve the determination accuracy and avoid judgment biases caused by differences in sliding window positions.

[0059] Furthermore, conflict adjustment is performed, specifically as follows: If the pixel block anomaly determination result is a conflict, the pixel block with the conflict determination result is marked as a conflicting pixel block, and the first specification unit anomaly determination result and the second specification unit anomaly determination result corresponding to the conflicting pixel block are obtained. If the first specification unit anomaly determination result is abnormal and the second specification unit anomaly determination result is normal, the conflicting pixel block is marked as a first specification problem pixel block; otherwise, it is marked as a second specification problem pixel block. A first adjustment coefficient is obtained based on the first abnormal feature value matching of the first specification problem pixel block. Based on the first adjustment coefficient, the gamma correction value and color temperature white balance adjustment of the first specification problem pixel block are reduced. The coefficient is used to obtain the first abnormal feature value for optimization adjustment. Based on the first abnormal feature value for optimization adjustment, it is compared with the abnormal feature threshold to obtain the pixel block abnormality judgment result. Based on the second abnormal feature value of the second specification problem pixel block, the second adjustment coefficient is obtained. Based on the second adjustment coefficient, the local contrast and blue channel gain value of the second specification problem pixel block are increased to obtain the second abnormal feature value for optimization adjustment. Based on the second abnormal feature value for optimization adjustment, it is compared with the abnormal feature threshold to obtain the pixel block abnormality judgment result. If the pixel block abnormality judgment result is that the pixel block is qualified, the pixel block is stored in the qualified image temporary storage area; if the pixel block abnormality judgment result is that the pixel block is abnormal, the pixel block is stored in the abnormal image temporary storage area.

[0060] In this embodiment, as Figure 3 As shown, Figure 3 This is a flowchart illustrating the conflict adjustment process of the LCD screen display abnormal image analysis method provided in this application embodiment. Upon receiving a conflict adjustment signal, the method acquires the first specification unit abnormality determination result and the second specification unit abnormality determination result corresponding to the conflicting pixel block. If the first specification unit abnormality determination result is abnormal and the second specification unit abnormality determination result is normal, the conflicting pixel block is marked as a first specification problem pixel block. If the first specification unit abnormality determination result is normal and the second specification unit abnormality determination result is abnormal, the conflicting pixel block is marked as a second specification problem pixel block. Based on the analysis of the first specification problem pixel block, a first adjustment coefficient is obtained. Based on the first adjustment coefficient, the gamma correction value and color temperature white balance adjustment coefficient of the first specification problem pixel block are reduced to obtain a pixel block abnormality determination result. Based on the analysis of the second specification problem pixel block, a second adjustment coefficient is obtained. Based on the second adjustment coefficient, the local contrast and blue channel gain value of the second specification problem pixel block are increased to obtain a pixel block abnormality determination result. If the pixel block abnormality determination result indicates that the pixel block is qualified, the pixel block is stored in the qualified image temporary storage area; if the pixel block abnormality determination result indicates that the pixel block is abnormal, the pixel block is stored in the abnormal image temporary storage area.

[0061] The gamma correction value and color temperature white balance adjustment coefficient of the first specification problem pixel block are reduced based on the first adjustment coefficient. The specific method is as follows: In the formula, γ1 represents the gamma correction value of the first specification problem pixel block after reduction and adjustment, γ0 represents the gamma correction value of the first specification problem pixel block, and r1 represents the first adjustment coefficient. In the formula, X1 represents the color temperature white balance adjustment coefficient of the first specification problem pixel block after reduction adjustment, X0 represents the color temperature white balance adjustment coefficient of the first specification problem pixel block, and r1 represents the first adjustment coefficient.

[0062] The first abnormal feature value for optimization and adjustment is obtained by taking the gamma correction value and color temperature white balance adjustment coefficient of the LCD screen optimized display image after reducing the gamma correction value and color temperature white balance adjustment coefficient of the first specification problem pixel block and then taking the abnormal feature value again.

[0063] The pixel block anomaly determination result is obtained by comparing the optimized and adjusted first abnormal feature value with the abnormal feature threshold. The specific method is as follows: the optimized and adjusted first abnormal feature value is compared with the abnormal feature threshold. If the optimized and adjusted first abnormal feature value is above the abnormal feature threshold, the pixel block anomaly determination result is that the pixel block is abnormal; otherwise, the pixel block anomaly determination result is that the pixel block is qualified.

[0064] The second adjustment coefficient is obtained by matching the second abnormal feature value of the second specification problem pixel block. The specific method is as follows: obtain the second abnormal feature value and the historical second adjustment coefficient of each sample pixel block in the database; calculate the second abnormal feature difference between the second specification problem pixel block and each sample pixel block; obtain the preset sample second abnormal feature difference range in the database and compare it. If the second abnormal feature difference between the second specification problem pixel block and a certain sample pixel block is within the sample second abnormal feature difference range, then obtain the sample pixel block as the reference sample pixel block; calculate the historical second adjustment coefficient of each reference sample pixel block and perform mean processing to obtain the mean of the historical second adjustment coefficient, and use it as the second adjustment coefficient.

[0065] The local contrast and blue channel gain of the second-specification problem pixel block are increased based on the second adjustment coefficient. The specific method is as follows: In the formula, W1 represents the local contrast of the second-size problematic pixel block after the adjustment, W0 represents the local contrast of the second-size problematic pixel block, and r2 represents the second adjustment coefficient. In the formula, L1 represents the blue channel gain value of the second-size problem pixel block after the adjustment, L0 represents the blue channel gain value of the second-size problem pixel block, and r2 represents the second adjustment coefficient.

[0066] The second abnormal feature value for optimization and adjustment is obtained by taking the optimized display image of the LCD screen after increasing and adjusting the local contrast and blue channel gain of the problematic pixel block of the second specification, and then taking the abnormal feature value again to obtain the second abnormal feature value for optimization and adjustment.

[0067] The pixel block anomaly determination result is obtained by comparing the optimized and adjusted first abnormal feature value with the abnormal feature threshold. The specific method is as follows: the optimized and adjusted first abnormal feature value is compared with the abnormal feature threshold. If the optimized and adjusted first abnormal feature value is above the abnormal feature threshold, the pixel block anomaly determination result is that the pixel block is abnormal; otherwise, the pixel block anomaly determination result is that the pixel block is qualified.

[0068] The reduction of gamma correction value for the first specification-related pixel block (large sliding block) is to reduce overly bright areas (such as bright spots, backlight spots, etc.) in the LCD screen display, weaken the excessive enhancement effect of bright areas on image edges, suppress image "flickering defects" caused by high brightness sensitivity, and also reduce interference such as stripes and noise caused by camera shooting, thereby reducing the overall risk of false anomalies in the sliding block. The reduction of the color temperature white balance adjustment coefficient is because the large window detection coverage is wide and easily affected by ambient cold light sources or display color cast. Lowering the white balance color temperature can effectively weaken blue and red casts under cold light dominance, reduce misjudgments of color spots caused by lighting or electromagnetic interference, and improve the accuracy of color consistency judgment.

[0069] The second specification addresses the issue of increased local contrast in small sliding pixel blocks (small sliding areas) because small windows focus on details, and low contrast can easily cause minor imperfections to be missed. Enhancing local contrast can highlight the characteristic textures of low-brightness areas in the image, improve detection sensitivity, and prevent pixel-level defects (such as micro-bright spots or dot-like dark spots) from being obscured or blurred by the background. Increasing the blue channel gain value is because small sliding windows tend to overlook cool-color defects (such as cold leakage or backlight edge defects) in low-light or cool-toned environments. Appropriately enhancing the blue channel response helps restore detail, compensates for local color distortion caused by white balance deviations, and improves the ability to identify anomalies against cool-toned backgrounds.

[0070] When analyzing pixel block anomaly detection results, there are three possible outcomes: qualified, abnormal, and conflicting. After conflict adjustment, only two outcomes remain: qualified and abnormal. This is because the adjusted image features more closely resemble the true image, and with the original interference factors effectively weakened or eliminated, a more stable and accurate classification can be achieved by comparing the result with the anomaly feature threshold. Therefore, the adjusted results only have two clear states: "qualified" and "abnormal," because the interference factors have been effectively mitigated.

[0071] like Figure 4 , Figure 5 As shown, Figure 4 This is a first example diagram of image acquisition for the LCD screen display anomaly image analysis system provided in this application embodiment. Figure 5 The second example image acquisition diagram of the LCD screen display anomaly image analysis system provided in this application embodiment shows the device connection status parameters and the LCD screen display images acquired under different test conditions.

[0072] Furthermore, the first adjustment coefficient is obtained by: acquiring the first abnormal feature value and the historical first adjustment coefficient of each sample pixel block in the database; calculating the first abnormal feature difference between the first specification problem pixel block and each sample pixel block; acquiring and comparing the preset sample first abnormal feature difference range in the database; if the first abnormal feature difference between the first specification problem pixel block and a certain sample pixel block is within the sample first abnormal feature difference range, then acquiring that sample pixel block as a reference sample pixel block; calculating the historical first adjustment coefficient of each reference sample pixel block and averaging it to obtain the average of the historical first adjustment coefficients, and using it as the first adjustment coefficient.

[0073] In this embodiment, by comparing the first abnormal feature value of the current first-specification problematic pixel block with the feature values ​​of historically accumulated sample pixel blocks in the database, and filtering out reference samples with similar features within a preset difference range, the mean value is calculated based on the historical first adjustment coefficients of these reference samples. This effectively achieves the experience transfer and dynamic adaptation of the adjustment strategy. By focusing on the actual adjustment experience of similar historical samples, reusable adjustment rules are extracted, thereby achieving more robust, refined, and better-matching image parameter optimization processing. Especially when dealing with complex pseudo-defect scenarios caused by camera interference, ambient light changes, screen brightness fluctuations, etc., it can improve the stability and response speed of the system in dealing with nonlinear interference features and local feature blurring problems.

[0074] Furthermore, the abnormal pixel clusters and their corresponding abnormal parameters are obtained. Specifically, the abnormal pixel blocks within the temporary storage area of ​​the abnormal image are acquired, and a spatial connectivity-based clustering algorithm is used to merge the connected regions of the abnormal pixel blocks to obtain each abnormal pixel cluster. A coordinate system is established with the preset position of the LCD screen optimized display image as the origin and the plane containing the LCD screen optimized display image as the coordinate plane, to obtain the coordinate position of each abnormal pixel cluster. Based on the coordinate position of each abnormal pixel cluster, the abnormal parameters corresponding to each abnormal pixel cluster are analyzed. The abnormal parameters corresponding to each abnormal pixel cluster include the maximum abnormal length, the maximum abnormal width, and the abnormal area.

[0075] In this embodiment, a clustering algorithm based on spatial connectivity is used to merge connected regions of abnormal pixel blocks. Specifically, a clustering algorithm based on spatial connectivity (such as a connected component labeling algorithm) is used to group abnormal pixels that are connected or close in position into a "cluster block" according to certain adjacency rules (e.g., 4-adjacency). This yields the individual abnormal pixel cluster blocks. Specifically, the analysis of each abnormal pixel cluster block involves merging "discrete abnormal pixel blocks" into "continuous abnormal regions" based on their relative positions in the image space.

[0076] A coordinate system is established based on the preset position of the LCD screen optimized display image as the origin and the plane containing the LCD screen optimized display image as the coordinate plane. The coordinate positions of each abnormal pixel cluster block are obtained. Specifically, the LCD screen optimized display image is defined as a two-dimensional plane (e.g., the upper left corner as the origin). Since the LCD screen optimized display image is rectangular, the lines containing its length and width can be used as the horizontal and vertical coordinates of the coordinate system, respectively. Thus, the coordinates of each pixel can be represented as (x, y). After merging connected regions, the abnormal parameters of each abnormal pixel cluster block can be obtained. The maximum abnormal length refers to the maximum horizontal length of the abnormal pixel cluster block, and the maximum abnormal width refers to the maximum vertical length of the abnormal pixel cluster block.

[0077] By introducing a clustering algorithm based on spatial connectivity to merge connected regions of anomalous pixel blocks, anomalous regions with spatial continuity and correlation can be effectively classified into unified clusters, thereby improving the accuracy and overall effectiveness of anomaly localization. Simultaneously, establishing a two-dimensional coordinate system based on the LCD screen plane provides each anomalous cluster with a clear location identifier, offering a geometric basis for subsequent fault type identification and localization analysis. By extracting parameters such as maximum anomaly length, maximum anomaly width, and anomaly area, the spatial extent and morphological characteristics of each cluster can be quantified, aiding in the identification of specific fault types such as bright spots, dark lines, and stripe interference.

[0078] like Figure 6 , Figure 7 , Figure 8 , Figure 9 As shown, Figure 6 , Figure 7 This is an example diagram of pixel defect results from the LCD screen display anomaly image analysis system provided in this application embodiment, wherein... Figure 6 The image shown is a dark spot defect, which is one of the pixel defect types. Figure 7 This is a magnified view of the dark spot defect. Figure 8 This is to demonstrate the dark line defect, which is a type of pixel defect. Figure 9 This is a magnified view of the dark line defect. The image shows the approximate location of the LCD screen fault and its cause.

[0079] Furthermore, the fault type of the abnormal image displayed on the LCD screen is obtained by: acquiring the preset abnormal LCD screen display model in the database; inputting each abnormal pixel cluster block and the abnormal parameters corresponding to each abnormal pixel cluster block as input variables into the abnormal LCD screen display model, thereby outputting the fault type of the abnormal LCD screen display image.

[0080] In this embodiment, the LCD screen display anomaly model refers to a model constructed using machine learning modeling to identify the types of LCD screen display anomalies. Its modeling process includes: first, collecting a large number of LCD screen display anomaly image samples with known fault types, and labeling each image with a specific anomaly type (such as bright spots, dark lines, color cast, etc.); then, standardizing the image data, including preprocessing steps such as size scaling, color normalization, and noise removal; selecting a CNN structure suitable for image classification tasks (such as ResNet, VGG, etc.), and inputting the images and their labeled data into the model for supervised training; during training, continuously adjusting the model parameters by optimizing the cross-entropy loss function to enable it to identify different types of LCD screen display anomalies; finally, evaluating and optimizing the model performance on the validation set to obtain a stably deployable LCD screen display anomaly model, achieving automatic identification of anomaly types in actual detected images. The core purpose of the LCD screen display anomaly model is to automatically determine the specific fault type of the anomaly based on the input anomaly region features (such as size, shape, location, etc.), achieving a closed-loop analysis from image anomaly to fault attribution.

[0081] By calling the preset LCD screen display anomaly model in the database and combining each abnormal pixel cluster block and its corresponding abnormal parameters as input, the system can accurately identify the fault type of abnormal LCD screen display images. It can effectively integrate spatial clustering features and quantitative abnormal parameters, improve the automation and accuracy of fault diagnosis, and avoid errors caused by human subjective judgment.

[0082] like Figure 10The diagram shows a schematic of the structure of the LCD screen display anomaly image analysis system provided in this application embodiment. The LCD screen display anomaly image analysis system provided in this application embodiment includes: an image optimization and adjustment module, an anomaly determination module, and a fault type analysis module. The image optimization and adjustment module is used to acquire LCD screen display images through an image acquisition device and perform image optimization and adjustment to obtain an optimized LCD screen display image. The anomaly determination module is used to set a first-size sliding window and a second-size sliding window to perform sliding detection on the optimized LCD screen display image, obtaining pixel block anomaly determination results. If the pixel block anomaly determination result is qualified, qualified pixel blocks are counted and stored in the qualified image temporary storage area; if the pixel block anomaly determination result is abnormal, abnormal pixel blocks are counted and stored in the abnormal image temporary storage area; if the pixel block anomaly determination result is conflicting, conflict adjustment is performed. The fault type analysis module is used to acquire abnormal pixel blocks in the abnormal image temporary storage area and process them to obtain each abnormal pixel cluster block and the corresponding abnormal parameters, thereby obtaining the fault type of the LCD screen display anomaly image.

[0083] In summary, this embodiment obtains and analyzes the parameters affecting image acquisition interference, thereby dynamically adjusting the image gain value to obtain an optimized image displayed on the LCD screen. This avoids image acquisition interference caused by improper parameter settings of the image acquisition device, effectively solving the problem in the prior art where defects are difficult to detect successfully due to image noise interference.

[0084] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0089] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for analyzing abnormal images displayed on a liquid crystal display screen, characterized in that, Includes the following steps: S1. Acquire images displayed on the LCD screen using an image acquisition device, and optimize and adjust the images to obtain optimized display images for the LCD screen; S2. Set the first specification sliding window and the second specification sliding window to perform sliding detection on the LCD screen optimized display image respectively, and obtain the pixel block abnormality judgment result. If the pixel block abnormality judgment result is qualified, the qualified pixel blocks are counted and stored in the qualified image temporary storage area. If the pixel block abnormality judgment result is abnormal, the abnormal pixel blocks are counted and stored in the abnormal image temporary storage area. If the pixel block abnormality judgment result is conflict, conflict adjustment is performed. S3. Obtain the abnormal pixel blocks in the abnormal image temporary storage area, and process them to obtain each abnormal pixel cluster block and the abnormal parameters corresponding to each abnormal pixel cluster block, thereby obtaining the fault type of the abnormal image displayed on the LCD screen. The specific methods for conflict mediation are as follows: If the pixel block anomaly determination result is a conflict, then the pixel block with the pixel block anomaly determination result is marked as a conflicting pixel block, and the first specification unit anomaly determination result and the second specification unit anomaly determination result corresponding to the conflicting pixel block are obtained. If the first specification unit anomaly determination result is abnormal and the second specification unit anomaly determination result is normal, then the conflicting pixel block is marked as a first specification problem pixel block; otherwise, it is marked as a second specification problem pixel block. Based on the first abnormal feature value matching of the first specification problem pixel block, a first adjustment coefficient is obtained. Based on the first adjustment coefficient, the gamma correction value and color temperature white balance adjustment coefficient of the first specification problem pixel block are reduced to obtain the optimized adjustment of the first abnormal feature value. Based on the optimized adjustment of the first abnormal feature value, the abnormal feature threshold is compared to obtain the pixel block abnormality judgment result. The second adjustment coefficient is obtained by matching the second abnormal feature value of the second specification problem pixel block. The local contrast and blue channel gain value of the second specification problem pixel block are increased based on the second adjustment coefficient to obtain the optimized adjustment of the second abnormal feature value. The optimized adjustment of the second abnormal feature value is compared with the abnormal feature threshold to obtain the pixel block abnormality judgment result. If the pixel block is deemed acceptable based on the anomaly determination result, the pixel block will be stored in the acceptable image temporary storage area. If the pixel block is determined to be abnormal, the pixel block will be stored in the abnormal image temporary storage area.

2. The method for analyzing abnormal images displayed on a liquid crystal screen as described in claim 1, characterized in that: The specific method for obtaining the optimized display image on the LCD screen is as follows: Obtain image acquisition interference parameters, which include the focal length, exposure time, frame rate, and ISO sensitivity of the image acquisition device; Obtain the preset ideal execution set for image acquisition from the database, and perform a difference analysis with the image acquisition interference parameters to obtain the image acquisition difference analysis results. Based on the image acquisition difference analysis results, introduce the corresponding weighting factor for coupling processing to obtain the image acquisition interference value. The ideal execution set for image acquisition includes ideal execution values ​​for focal length, exposure time, frame rate, and ISO sensitivity. By matching the image acquisition interference value with the database, the image gain adjustment execution coefficient is obtained. Based on the image gain adjustment execution coefficient, the image gain value of the image displayed on the LCD screen is reduced, thereby obtaining the optimized display image of the LCD screen.

3. The method for analyzing abnormal images displayed on a liquid crystal screen as described in claim 1, characterized in that: The specific method for obtaining the pixel block anomaly determination result is as follows: Set a first-size sliding window and a second-size sliding window, wherein the first-size sliding window is larger than the second-size sliding window; Based on the first specification sliding window, the LCD screen optimized display image is slide-detected according to the first preset step size, and the pixel blocks obtained from each slide detection are jointly marked as the first specification unit, thereby obtaining each first specification unit. Based on the second specification sliding window, the LCD screen optimized display image is slide-detected according to the second preset step size, and the pixel blocks obtained from each slide detection are jointly marked as the second specification unit, thereby obtaining each second specification unit. Obtain abnormal feature parameters and analyze them to obtain abnormal feature values, wherein the abnormal feature values ​​include the first abnormal feature value of each first specification unit and the second abnormal feature value of each second specification unit; Obtain the preset abnormal feature threshold in the database, compare the first abnormal feature value of each first specification unit with the abnormal feature threshold to obtain the abnormal judgment result of each first specification unit, and compare the second abnormal feature value of each second specification unit with the abnormal feature threshold to obtain the abnormal judgment result of each second specification unit. The pixel block anomaly determination results are obtained by analyzing the anomaly determination results of each first specification unit and each second specification unit.

4. The method for analyzing abnormal images displayed on a liquid crystal screen as described in claim 3, characterized in that: The specific method for obtaining the abnormal feature values ​​is as follows: Obtain abnormal feature parameters, including mean brightness, standard deviation of brightness, mean color channel deviation, and mean edge intensity; Obtain the preset abnormal feature standard set in the database, and perform a difference analysis with the abnormal feature parameters to obtain the abnormal feature difference analysis results. Based on the abnormal feature difference analysis results, introduce the corresponding weighting factor for coupling processing to obtain the abnormal feature value. The abnormal feature standard set includes luminance standard value, luminance standard deviation allowable value, color channel deviation allowable value, and edge intensity standard value.

5. The method for analyzing abnormal images displayed on a liquid crystal screen as described in claim 1, characterized in that: The obtained pixel block anomaly determination result specifically includes: Obtain the pixel block corresponding to each first specification unit. If there is a pixel block in a different first specification unit and the anomaly judgment results of the first specification units corresponding to the pixel block are inconsistent, then the first specification unit anomaly judgment result of the pixel block is the anomaly judgment result obtained by the first sliding detection of the pixel block. Obtain the pixel blocks corresponding to each second specification unit. If there is a pixel block in different second specification units and the anomaly judgment results of each second specification unit corresponding to the pixel block are inconsistent, then the anomaly judgment result of the second specification unit of the pixel block is the anomaly judgment result obtained by the first sliding detection of the pixel block. The system analyzes the pixel blocks in the optimized display image on the LCD screen and the anomaly judgment results of the first specification unit and the second specification unit corresponding to the pixel blocks. If there is a pixel block whose first specification unit anomaly judgment result is normal and the second specification unit anomaly judgment result is normal, then the pixel block anomaly judgment result of that pixel block is qualified. If the first specification unit of a pixel block is determined to be abnormal and the second specification unit is determined to be abnormal, then the pixel block abnormality determination result of that pixel block is abnormal. If the error determination results of the first specification unit and the second specification unit of a certain pixel block are inconsistent, then the pixel block error determination result of that pixel block is conflicting.

6. The method for analyzing abnormal images displayed on a liquid crystal screen as described in claim 1, characterized in that: The method for obtaining the first adjustment coefficient is as follows: Obtain the first abnormal feature value and the historical first adjustment coefficient of each sample pixel block in the database; Calculate the difference in the first outlier feature between the first-specification problematic pixel block and each sample pixel block; Obtain the preset sample first abnormal feature difference range in the database and compare it. If the first abnormal feature difference between the first specification problem pixel block and a certain sample pixel block is within the sample first abnormal feature difference range, then obtain the sample pixel block as the reference sample pixel block. The historical first adjustment coefficients of each reference sample pixel block are averaged to obtain the average historical first adjustment coefficient, which is then used as the first adjustment coefficient.

7. The method for analyzing abnormal images displayed on a liquid crystal screen as described in claim 1, characterized in that: The specific method for obtaining each abnormal pixel cluster block and the corresponding abnormal parameters of each abnormal pixel cluster block is as follows: Abnormal pixel blocks within the temporary storage area of ​​abnormal images are obtained, and a clustering algorithm based on spatial connectivity is used to merge the connected regions of the abnormal pixel blocks to obtain each abnormal pixel cluster block. A coordinate system is established based on the preset position of the image optimized for display on the LCD screen as the origin and the plane where the image optimized for display on the LCD screen is located as the coordinate plane, so as to obtain the coordinate position of each abnormal pixel cluster block. Based on the coordinate position analysis of each abnormal pixel cluster block, the abnormal parameters corresponding to each abnormal pixel cluster block are obtained. The abnormal parameters corresponding to each abnormal pixel cluster block include the maximum abnormal length, the maximum abnormal width, and the abnormal area.

8. The method for analyzing abnormal images displayed on a liquid crystal screen as described in claim 1, characterized in that: The specific method for obtaining the fault type of the abnormal image displayed on the LCD screen is as follows: Retrieve the preset LCD screen display anomaly model from the database; The abnormal pixel clusters and their corresponding abnormal parameters are input variables into the LCD screen display abnormality model, thereby outputting the fault type of the abnormal image displayed on the LCD screen.

9. A liquid crystal display abnormality image analysis system, employing the liquid crystal display abnormality image analysis method as described in any one of claims 1-8, characterized in that, include: Image optimization and adjustment module, anomaly detection module, and fault type analysis module; The image optimization and adjustment module is used to acquire images displayed on the LCD screen through an image acquisition device, and to optimize and adjust the images to obtain optimized images for display on the LCD screen. The anomaly determination module is used to set a first-size sliding window and a second-size sliding window to perform sliding detection on the optimized display image of the LCD screen, and obtain the pixel block anomaly determination result. If the pixel block anomaly determination result is qualified, the qualified pixel blocks are counted and stored in the qualified image temporary storage area. If the pixel block anomaly determination result is abnormal, the abnormal pixel blocks are counted and stored in the abnormal image temporary storage area. If the pixel block anomaly determination result is conflict, conflict adjustment is performed. The fault type analysis module is used to obtain abnormal pixel blocks in the abnormal image temporary storage area, and process them to obtain each abnormal pixel cluster block and the abnormal parameters corresponding to each abnormal pixel cluster block, thereby obtaining the fault type of the abnormal image displayed on the LCD screen.

Images