LCD screen surface state detection method, system, device and readable storage medium

Abstract

This application discloses a method, system, device, and readable storage medium for detecting the surface condition of an LCD screen. It employs three imaging methods—dark-field scattering imaging, polarization differential imaging, and transmitted light imaging—to acquire feature images, extract corresponding surface and internal features, and construct a multimodal feature set by spatial coordinate registration of the three types of images. Then, based on region matching relationships, it separates and determines surface contamination from internal defects to ascertain the screen's surface condition. This application utilizes the synergistic effect of three imaging methods to accurately capture the morphology and texture of fine surface contamination and reflective defects, simultaneously acquiring internal defect and contamination projection information. Coordinate registration eliminates image misalignment interference, constructing a complete and accurate multimodal feature set, enabling precise differentiation between surface contamination and internal defects. This application significantly improves the accuracy and efficiency of LCD screen surface condition detection, reduces the rate of missed and false detections, adapts to different screen specifications, and meets the quality inspection needs of high-end electronic equipment.

Description

Technical Field

[0001] This application relates to the field of image detection technology, and more specifically, to a method, system, device, and readable storage medium for detecting the surface state of an LCD screen. Background Technology

[0002] As a core display component of various electronic devices, LCD screens are widely used in mobile phones, computers, smart terminals, automotive displays, and other fields. Their surface condition directly affects the display effect, product lifespan, and user experience. Therefore, accurate detection of the surface condition of LCD screens is a crucial step in the production, quality inspection, and maintenance process and is extremely necessary.

[0003] In existing technologies, LCD screen surface inspection methods employ several approaches. One approach can only capture the scattering characteristics of specific types of surface contaminants, but it cannot effectively distinguish these characteristics from the projection signals of internal defects, easily leading to internal defects being misidentified as surface dirt. Another approach can capture localized surface reflection differences to identify specific types of surface scratches, but it lacks sensitivity in identifying minute dirt with inconspicuous reflection characteristics and cannot penetrate the screen to detect internal defects. A third imaging method can capture signals of internal defects using a test image, but surface dirt creates projection interference, with its brightness and color characteristics overlapping with internal defects, making them difficult to distinguish effectively. Furthermore, existing technologies suffer from high rates of missed and false detections and insufficient detection accuracy, failing to meet the testing requirements of high-end LCD screens.

[0004] Therefore, there is an urgent need for a new method for detecting the surface condition of LCD screens to overcome the shortcomings of existing technologies, achieve accurate separation of surface dirt and internal defects of LCD screens, improve detection accuracy, and meet the detection requirements of high-end LCD screens. Summary of the Invention

[0005] This application provides a method, system, device and readable storage medium for detecting the surface condition of an LCD screen. This application can significantly improve the accuracy and efficiency of LCD screen surface condition detection, reduce the rate of missed detection and false detection, adapt to different screen specifications, and meet the quality inspection needs of high-end electronic equipment.

[0006] A method for detecting the surface condition of an LCD screen, comprising:

[0007] Dark-field scattering imaging was used to acquire the first feature image of the LCD screen surface and extract the morphological and texture features of surface dirt.

[0008] A second feature image of the LCD screen surface is acquired using polarization differential imaging to extract surface reflection characteristics.

[0009] The third feature image is acquired by using transmitted light imaging while the test screen is displayed on the LCD screen, and the brightness, color and position features of the projection of internal defects and surface dirt are extracted.

[0010] The first feature image, the second feature image, and the third feature image are spatially registered to construct a multimodal feature set;

[0011] Based on the region matching relationship of each feature image in the multimodal feature set, surface dirt and internal defects are separated and determined to determine the surface condition of the LCD screen.

[0012] Optionally, the step of acquiring a first feature image of the LCD screen surface using dark-field scattering imaging includes:

[0013] Control the low-angle light source to illuminate the screen at a preset starting angle and acquire the first dark field image;

[0014] In the first dark field image, a defect-free area is selected to calculate the gray-scale mean. If the gray-scale mean exceeds a set threshold, the incident angle of the light source is gradually reduced, the dark field image is re-acquired, and the gray-scale mean of the corresponding area is calculated until the gray-scale mean is lower than the threshold, and the optimal dark field angle is determined.

[0015] A second dark field image is acquired at the optimal dark field angle. The second dark field image is then filtered, enhanced, and subjected to local adaptive threshold segmentation to extract candidate dirty regions and generate the first feature image.

[0016] Optionally, the step of acquiring a second feature image of the LCD screen surface using polarization differential imaging includes:

[0017] A polarizer fixed in front of the light source controls the analyzer set in front of the camera to rotate to the orthogonal polarization position, and acquires orthogonal polarization images;

[0018] The analyzer is rotated to the parallel polarization position to acquire a parallel polarization image;

[0019] Calculate the difference image between the parallel polarization image and the orthogonal polarization image, and normalize and enhance the contrast of the difference image to obtain the enhanced difference image.

[0020] Adaptive threshold segmentation and connected component analysis are used to extract candidate surface reflection regions from the enhanced difference image to generate the second feature image.

[0021] Optionally, the method of acquiring the third feature image using transmitted light imaging while the test image is displayed on the LCD screen includes:

[0022] Control the LCD screen to display multiple standard test images in sequence, and synchronously collect multiple frames of transmission images corresponding to each test image through hardware triggering;

[0023] Perform brightness normalization on each frame of the transmission image, and perform differential operations with the standard template images of the corresponding test images respectively to obtain the difference images corresponding to each image;

[0024] Fuse the difference images corresponding to each image to generate a transmission anomaly image, perform region extraction on the transmission anomaly image to generate the third feature image.

[0025] Optionally, the process of spatial coordinate registration includes:

[0026] Pre-jointly calibrate the three cameras corresponding to dark field scattering imaging, polarization difference imaging, and transmission light imaging, and establish the transformation matrix from each camera coordinate system to the target coordinate system;

[0027] Convert the region coordinates in the first feature image, the second feature image, and the third feature image to the target coordinate system;

[0028] For each candidate abnormal region in the third feature image, search for the candidate region with the closest spatial distance in the first feature image and the second feature image. If the distance is less than the preset matching threshold, establish a multi-modal feature correspondence relationship, and mark the candidate regions with corresponding relationships as associated regions.

[0029] Optionally, the separation determination of surface dirt and internal defects according to the region matching relationship of each feature image in the multi-modal feature set includes:

[0030] If both the first feature image and the second feature image are empty, and the third feature image is non-empty, it is determined as an internal defect;

[0031] If the first feature image or the second feature image is non-empty, and there is an abnormal region in the third feature image associated with the first feature image or the second feature image, it is determined as surface dirt;

[0032] If the first feature image or the second feature image is non-empty, and there is no abnormal region in the third feature image associated with the first feature image or the second feature image, it is determined as pure surface dirt;

[0033] If the first feature image, the second feature image, and the third feature image are all empty, it is determined as qualified.

[0034] An LCD screen surface state detection system includes:

[0035] The dark field scattering imaging module is used to illuminate the screen surface at an incident angle lower than the LCD screen surface, acquire the image of the scattered light from the screen surface, and obtain the first feature image characterizing the surface dirt after image processing.

[0036] The polarization difference imaging module includes a polarizer placed in front of the light source and an analyzer placed in front of the camera, used to acquire orthogonal polarization images and parallel polarization images respectively, and obtain a second feature image characterizing the surface reflection properties based on the difference between the two.

[0037] The transmitted light imaging module is used to acquire transmitted light images while the test screen is displayed on the LCD screen. After image processing, a third feature image containing projections of internal defects and surface dirt is obtained.

[0038] The feature fusion determination module is used to spatially register the first feature image, the second feature image, and the third feature image in a unified coordinate system, and to separate and determine the surface dirt and internal defects according to the regional matching relationship of each feature image, thereby determining the surface condition of the LCD screen.

[0039] Optionally, the dark field scattering imaging module includes a light source angle self-calibration unit, which is used to select a defect-free area in the acquired image to calculate the gray-scale mean, and use the gray-scale mean as feedback to control the incident angle of the light source until the gray-scale mean is lower than a preset threshold, to determine the optimal dark field angle, and control the dark field scattering imaging module to acquire an image at the optimal dark field angle to generate the first feature image.

[0040] The polarization differential imaging module includes a polarization orthogonal calibration unit, which is used to determine the analyzer angle that minimizes the average gray level of the entire image after each rotation by rotating the analyzer and calculating the average gray level of the entire image. The analyzer angle is then used as the orthogonal polarization acquisition position. The polarization differential imaging module is then controlled to acquire orthogonal polarization images at the orthogonal polarization acquisition position to generate the second feature image.

[0041] An LCD screen surface condition detection device includes a memory and a processor;

[0042] The memory is used to store programs;

[0043] The processor is used to execute the program to implement each step of the LCD screen surface state detection method as described in any of the above claims.

[0044] A readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the LCD screen surface state detection method as described in any of the preceding claims.

[0045] As can be seen from the above technical solutions, the LCD screen surface state detection method, system, device, and readable storage medium provided in this application employs dark field scattering imaging to acquire a first feature image to extract the morphological and texture features of surface dirt; employs polarization difference imaging to acquire a second feature image to extract surface reflection characteristics; employs transmitted light imaging to acquire a third feature image while the LCD screen displays a test image, extracting the brightness, chromaticity, and position features of internal defects and surface dirt projections; performs spatial coordinate registration on the three types of feature images to construct a multimodal feature set; and, based on the regional matching relationship of the feature images, separates and determines surface dirt and internal defects to determine the surface state of the LCD screen. This application utilizes the synergistic application of three imaging methods. Leveraging the characteristics of dark-field scattering imaging, it can accurately capture the morphology and texture details of minute surface contaminants, achieving efficient identification of various surface dirt such as dust and fingerprints. Through polarization differential imaging technology, it can clearly extract the reflective characteristics of the LCD screen surface, effectively identifying reflective defects such as surface scratches, compensating for the insufficient sensitivity in identifying minute contaminants. By combining transmitted light imaging with the acquisition of LCD screen test images, it can simultaneously acquire the brightness, color, and position information of internal defects and surface contaminant projections, providing comprehensive data support for subsequent feature differentiation. By performing spatial coordinate registration on the three types of feature images, it can eliminate image misalignment interference caused by different imaging methods, constructing a complete and accurate multimodal feature set. Furthermore, by analyzing the region matching relationships of each image in the multimodal feature set, it can achieve accurate separation of surface contaminants and internal defects, clearly distinguishing their characteristic differences. In summary, this application can significantly improve the detection accuracy and efficiency of LCD screen surface conditions, effectively reducing the rate of missed and false detections, adapting to different specifications of LCD screens, and meeting the quality inspection needs of high-end electronic devices. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0047] Figure 1 This is a flowchart of an LCD screen surface condition detection method disclosed in an embodiment of this application;

[0048] Figure 2 This is a schematic diagram of an LCD screen surface condition detection system disclosed in an embodiment of this application;

[0049] Figure 3 This is a hardware structure block diagram of an LCD screen surface condition detection device disclosed in an embodiment of this application. Detailed Implementation

[0050] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0051] This application can be used in a wide variety of general-purpose or special-purpose computing device environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor devices, distributed computing environments including any of the above devices, etc.

[0052] The following section introduces the solution proposed in this application. The technical solution is as follows, and details are provided below.

[0053] Figure 1 This is a flowchart of an LCD screen surface condition detection method disclosed in an embodiment of this application.

[0054] like Figure 1 As shown, the method may include:

[0055] Step S1: Using dark field scattering imaging, acquire the first feature image of the LCD screen surface and extract the morphological and texture features of the surface dirt.

[0056] Specifically, dark-field scattering imaging technology accurately captures the morphology and texture features of dirt on the LCD screen surface, generating a first-feature image that clearly reflects the surface dirt information. The core advantage of dark-field scattering imaging is its ability to highlight the scattering signals of minute surface dirt while suppressing specular reflection interference from the screen itself, thus achieving precise capture of fine dirt. This is suitable for detecting surface dust, fingerprints, and fine oil stains during LCD screen production and quality inspection. The following details each sub-step of this process:

[0057] ① Control the low-angle light source to illuminate the screen surface at a preset starting angle and acquire the first dark field image: First, start the low-angle light source in the detection device and adjust the illumination direction of the light source according to the preset starting angle to ensure that the light source shines obliquely on the LCD screen surface at a low angle, avoiding excessive specular reflection caused by direct light source, which would affect the capture of dirt features. At the same time, start the imaging device and synchronously acquire the image of the LCD screen surface at this time, which is the first dark field image. This image initially records the scattering signal of the screen surface, which contains the preliminary features of surface dirt, but there may be problems such as unsuitable imaging brightness and unclear dirt features, which need to be further optimized through subsequent sub-steps.

[0058] ② In the first dark-field image, select a defect-free area to calculate the average grayscale value. If the average grayscale value exceeds a set threshold, gradually reduce the incident angle of the light source, re-acquire the dark-field image, and calculate the average grayscale value of the corresponding area until the average grayscale value is below the threshold. Determine the optimal dark-field angle: Due to differences in surface materials and coating processes of LCD screens of different specifications and batches, their surface reflection characteristics also vary. A single incident angle of the light source cannot meet the imaging requirements of all screens. Therefore, in the acquired first dark-field image, a clean area without any dirt or defects should be selected as a reference area, and the average grayscale value of this area should be calculated. The average grayscale value is then used to determine whether the current imaging brightness is reasonable. If the average grayscale value exceeds the set threshold, it indicates that the current incident angle of the light source is too large, resulting in excessively high image brightness and strong specular reflection signal on the screen surface, which masks the scattering characteristics of the dirt and makes it impossible to clearly distinguish between the dirt and the background. At this time, it is necessary to gradually reduce the incident angle of the light source, re-acquire the dark field image after each adjustment, and repeatedly calculate the average grayscale value of the corresponding defect-free area until the average grayscale value is lower than the set threshold. At this time, the light source illumination angle can effectively suppress specular reflection and highlight the scattering characteristics of the dirt. This angle is the optimal dark field angle.

[0059] ③ Acquire a second dark-field image at the optimal dark-field angle. Perform filtering, enhancement, and local adaptive thresholding on the second dark-field image to extract candidate dirt regions and generate the first feature image: After determining the optimal dark-field angle, keep the light source angle unchanged and re-acquire a dark-field image of the LCD screen surface, which is the second dark-field image. Compared with the first dark-field image, this image has clearer dirt features and less background interference. To further improve image quality and eliminate noise interference in the image (such as noise from the imaging device itself, ambient light interference, etc.), filter the second dark-field image to filter out irrelevant noise signals; then perform image enhancement processing to improve the contrast between the dirt region and the background region, making the dirt features more prominent. Then, use local adaptive thresholding technology to set an adaptive segmentation threshold according to the gray-level distribution characteristics of different regions of the image to accurately segment the regions in the image that may contain dirt, i.e., candidate dirt regions. Finally, integrate the morphological and texture features of all candidate dirt regions to generate the first feature image, completing the extraction of surface dirt morphology and texture features in this step.

[0060] Step S2: Use polarization differential imaging to acquire the second feature image of the LCD screen surface and extract the surface reflection characteristics.

[0061] Specifically, polarization differential imaging technology is used to capture the differences in reflective characteristics of various defects on the LCD screen surface, generating a second feature image that clearly reflects the surface reflective features. This extracts surface reflective characteristics, compensating for the shortcomings of dark-field scattering imaging in identifying surface reflective defects, and providing reflective dimension features for subsequent separation of surface dirt from internal defects. Polarization differential imaging utilizes the polarization characteristics of light to effectively distinguish between reflected signals and background signals on the screen surface, making it particularly suitable for identifying defects with specific reflective characteristics, such as surface scratches and residual oil, thus improving the comprehensiveness and accuracy of detection. The following details each sub-step of this process:

[0062] ① A polarizer is fixedly installed in front of the light source, and the analyzer installed in front of the camera is rotated to an orthogonal polarization position to acquire orthogonal polarization images: First, the polarizer is fixedly installed in front of the light source so that the natural light emitted by the light source is converted into linearly polarized light after passing through the polarizer, ensuring that the light incident on the LCD screen surface has a uniform polarization direction. Then, the analyzer installed in front of the imaging camera is adjusted and rotated to a polarization position orthogonal to the polarizer (i.e., the polarization directions of the polarizer and the analyzer are perpendicular to each other). At this time, the imaging camera is started to acquire an image of the LCD screen surface, which is an orthogonal polarization image. In this polarization state, specular reflection light from the screen surface is significantly suppressed, and only reflection signals with specific polarization characteristics can be captured, initially screening out areas that may have reflection defects.

[0063] ② Rotate the analyzer to the parallel polarization position and acquire a parallel polarization image: Keeping the position and state of the polarizer unchanged, adjust the rotation angle of the analyzer to rotate it to a polarization position parallel to the polarizer (i.e., the polarization directions of the polarizer and the analyzer are completely consistent). At this time, restart the imaging camera to acquire an image of the LCD screen surface, which is the parallel polarization image. In this polarization state, the specular reflection light from the screen surface can pass smoothly through the analyzer and be captured by the camera, thereby obtaining the complete reflection signal of the screen surface. This signal is then compared with the orthogonal polarization image, laying the foundation for subsequent extraction of differences in reflection characteristics.

[0064] ③ Calculate the difference image between the parallel polarization image and the orthogonal polarization image, and perform normalization and contrast enhancement processing on the difference image to obtain the enhanced difference image: Since the polarization states of the parallel polarization image and the orthogonal polarization image are different, the reflected signals captured by the two are significantly different. The difference image can effectively highlight the reflection difference information under the two polarization states and suppress background interference and irrelevant signals. Specifically, by calculating the gray-level difference of corresponding pixels in the parallel polarization image and the orthogonal polarization image, a difference image is generated. In this image, the difference in the reflective characteristics of surface defects will be presented in the form of obvious gray-level changes. To eliminate the interference caused by the brightness difference in different areas and ensure the consistency of defect features, the difference image is normalized to uniformly adjust the image gray-level values ​​to a preset range; at the same time, contrast enhancement processing is performed to increase the gray-level difference between the defect area and the background area, making the surface reflective characteristics more clearly distinguishable, thus obtaining the enhanced difference image.

[0065] ④ Adaptive threshold segmentation and connected component analysis are used to extract candidate surface reflection regions from the enhanced difference image to generate a second feature image: For the enhanced difference image, adaptive threshold segmentation is applied. Based on the grayscale distribution characteristics of different regions of the image, the segmentation threshold is automatically set to accurately segment regions with significant differences in reflective properties, i.e., candidate surface reflection regions. These regions correspond to reflective defects such as scratches and residual oil stains that may exist on the screen surface. Subsequently, connected component analysis is used to filter and integrate the segmented candidate surface reflection regions, eliminating noisy regions that are too small or have abnormal shapes, and retaining regions with actual defect characteristics. Finally, the reflective characteristic features of all candidate surface reflection regions are integrated to generate the second feature image, completing the extraction of surface reflective characteristics in this step.

[0066] Step S3: Using transmitted light imaging, acquire the third feature image while the test screen is displaying the test image on the LCD screen, and extract the brightness, color and position features of the projection of internal defects and surface dirt.

[0067] Specifically, by using transmitted light imaging combined with the state of the test image displayed on the LCD screen, the feature information of internal defects and surface dirt projections is simultaneously captured, generating a third feature image. The brightness, chromaticity, and positional features of the internal defects and surface dirt projections are extracted, providing internal and projection-dimensional features for subsequent separation of surface dirt from internal defects. Transmitted light imaging can penetrate the LCD screen to capture signals of internal defects, and by combining it with a standard test image, the differences between internal defects and surface dirt projections can be further highlighted, solving the problem that a single imaging method cannot simultaneously capture internal defects and surface dirt projections. The following provides a detailed explanation of each sub-step in this process:

[0068] ① The system controls the LCD screen to display multiple standard test screens sequentially, and synchronously acquires multiple frames of transmission images corresponding to each test screen through hardware triggering: First, the control module sends instructions to control the LCD screen to display multiple preset standard test screens sequentially. These test screens are specially designed to cover pure color images or specific patterns with different brightness and chromaticity, which can effectively highlight the characteristics of internal defects in the screen (such as pixel damage, internal cracks, liquid crystal leakage, etc.), making the internal defects appear as obvious brightness and chromaticity abnormalities under transmitted light. At the same time, a hardware triggering method is used to synchronize the display signal of the LCD screen with the acquisition signal of the imaging device, ensuring that when the LCD screen switches to a certain test screen, the imaging device can timely and accurately acquire the transmission image under that screen state, avoiding problems such as image blurring and feature loss caused by asynchronous screen switching and acquisition. Finally, multiple frames of transmission images corresponding to each test screen are acquired, comprehensively capturing the projection information of internal defects and surface dirt under different screens.

[0069] ② The brightness of each frame of the transmission image is normalized, and a difference operation is performed between each frame and the standard template image of the corresponding test screen to obtain the difference map for each screen: Due to the differences in brightness and chromaticity between different test screens, and the influence of ambient light and the accuracy of the imaging equipment itself during the acquisition process, the brightness of each frame of the transmission image is uneven, affecting the accuracy of subsequent feature extraction. Therefore, the brightness of each frame of the transmission image is first normalized to adjust the grayscale values ​​of all transmission images to a preset standard range, eliminating the interference caused by brightness differences and ensuring the comparability of features of different frames. Then, the normalized transmission image of each frame is compared with the standard template image of the corresponding test screen to perform a difference operation, that is, to calculate the difference in brightness and chromaticity of corresponding pixels between the transmission image and the standard template image, generating the difference map for each test screen. In this difference map, the projection of internal defects and surface dirt will show obvious abnormal differences, while the difference in the normal area of ​​the screen is close to zero, thus effectively highlighting the features of abnormal areas.

[0070] ③ Merge the difference maps corresponding to each screen to generate a transmission anomaly map. Extract regions from the transmission anomaly map to generate a third feature image: The difference map corresponding to a single test screen may not be able to fully capture all the features of internal defects and surface dirt projections. Some defects may not be obvious in one test screen but are prominent in other test screens. Therefore, it is necessary to merge the difference maps corresponding to all test screens. Through image fusion algorithms, the abnormal region information in each difference map is integrated to make up for the deficiencies of a single difference map and generate a transmission anomaly map that can comprehensively reflect the abnormal projections of internal defects and surface dirt on the screen. Subsequently, region extraction is performed on the transmission anomaly map to select regions with obvious brightness and color anomalies. The brightness, color, and position features of these abnormal regions are integrated to generate a third feature image, completing the extraction of internal defect and surface dirt projection features in this step.

[0071] Step S4: Perform spatial coordinate registration on the first feature image, the second feature image, and the third feature image to construct a multimodal feature set.

[0072] Specifically, to eliminate spatial misalignment errors caused by different imaging methods in the first, second, and third feature images, spatial coordinate registration is used to unify the coordinate systems of the three types of images. This allows for the integration of feature information from the three types of images, constructing a multimodal feature set. This provides comprehensive and accurate multi-dimensional feature support for the subsequent separation and determination of surface contamination and internal defects. Because the first three steps use different imaging methods (dark-field scattering, polarization difference, and transmitted light imaging), there may be slight deviations in imaging angle, imaging range, and imaging accuracy. Directly fusing the features of the three types of images would lead to feature misalignment and matching errors, affecting the accuracy of subsequent determinations. Therefore, spatial coordinate registration is a crucial prerequisite for constructing a multimodal feature set.

[0073] First, registration reference feature points are selected. Fixed feature points are chosen from the surface of the LCD screen as the reference for coordinate registration. These feature points must meet the requirements of stable position, high recognizability, and clear identification in all three types of feature images. Examples include the corners of the LCD screen and fixed marker points, ensuring the stability and accuracy of the registration process. After selecting the reference feature points, the coordinate positions of these feature points are located in the first, second, and third feature images, respectively, and the coordinate information of each feature point in the three types of images is recorded.

[0074] Secondly, coordinate system merging and misalignment correction are performed. A feature point matching algorithm is employed, using the coordinate system of the first feature image as the reference coordinate system. The coordinates of the reference feature points in the second and third feature images are transformed to the reference coordinate system using a coordinate transformation algorithm, eliminating spatial misalignment errors caused by different imaging methods. During the coordinate transformation process, the scaling ratio, rotation angle, and translation distance of the three types of images are precisely calibrated to ensure that the spatial positions of corresponding regions in the three types of feature images are completely and accurately matched, avoiding feature matching errors caused by image misalignment.

[0075] Finally, a multimodal feature set is constructed. After spatial coordinate registration, the feature information of the three types of feature images is integrated: the morphology and texture features of surface dirt in the first feature image, the surface reflection characteristics in the second feature image, and the brightness, chromaticity, and position features of internal defects and surface dirt projections in the third feature image are extracted. These feature information from different dimensions are correlated and integrated to construct a multimodal feature set containing multidimensional and multi-type features. This feature set comprehensively covers various feature information of surface dirt, surface reflection defects, and internal defects of LCD screens, providing comprehensive and reliable feature support for the separation and judgment of surface dirt and internal defects in subsequent steps, ensuring the accuracy of the judgment results.

[0076] Step S5: Based on the region matching relationship of each feature image in the multimodal feature set, separate and determine the surface dirt and internal defects to determine the surface condition of the LCD screen.

[0077] Specifically, by utilizing the multimodal feature set constructed in step S4 and analyzing the region matching relationships of various feature images, accurate separation of surface dirt and internal defects is achieved. Ultimately, a comprehensive assessment and determination of the LCD screen's surface condition is made, providing a clear basis for quality inspection and maintenance. This step, by integrating the multi-dimensional features extracted in the previous three steps and combining them with region matching analysis, effectively solves the problem of easily confusing surface dirt with internal defects and the high rate of misjudgment, ensuring the accuracy and reliability of the detection results.

[0078] First, the region matching relationships of each feature image in the multimodal feature set are analyzed. The multimodal feature set constructed in step S4 is retrieved, and abnormal regions are extracted from the first, second, and third feature images respectively. The correspondence of each abnormal region in the three types of feature images is analyzed one by one to clarify the feature differences of different abnormal regions. Specifically, the location, shape, texture, reflectivity, brightness, and chromaticity of each abnormal region are compared and analyzed to establish region matching associations between the three types of feature images, laying the foundation for subsequent separation and determination.

[0079] Secondly, surface contamination and internal defects are separated and determined. Based on the region matching relationship and feature differences of the three types of feature images, surface contamination and internal defects are clearly distinguished: surface contamination has obvious morphological and texture features, thus presenting clear candidate contamination areas in the first feature image; simultaneously, surface contamination produces specific reflection signals, thus corresponding candidate surface reflection areas appear in the second feature image; furthermore, surface contamination forms projections under transmitted light, thus corresponding abnormal projection areas with specific brightness and chromaticity appear in the third feature image, and the spatial positions of these three types of abnormal areas have a one-to-one correspondence. Internal defects, on the other hand, exist only inside the screen and cannot form morphological, texture, or reflection features on the surface, therefore only presenting areas with abnormal brightness and chromaticity in the third feature image, with no corresponding abnormal areas in the first and second feature images. Based on the above feature differences, through region matching and feature filtering, surface contamination areas and internal defect areas can be accurately separated, clearly distinguishing their location, range, and type.

[0080] Finally, the surface condition of the LCD screen is determined. Integrating the results of separating and judging surface contamination and internal defects, a comprehensive assessment of the LCD screen's surface contamination (such as location, quantity, size, and type of contamination) and internal defects (such as location, extent, and type of defects) is conducted. Combined with preset inspection standards, it is determined whether the LCD screen's surface condition meets the requirements, ultimately outputting a clear inspection result, completing the entire LCD screen surface condition inspection process. Implementing this step significantly improves inspection accuracy and efficiency, reduces the rate of missed and false judgments, adapts to different specifications of LCD screens, and meets the quality inspection needs of high-end electronic equipment.

[0081] As can be seen from the above technical solutions, the LCD screen surface state detection method, system, device, and readable storage medium provided in this application employs dark field scattering imaging to acquire a first feature image to extract the morphological and texture features of surface dirt; employs polarization difference imaging to acquire a second feature image to extract surface reflection characteristics; employs transmitted light imaging to acquire a third feature image while the LCD screen displays a test image, extracting the brightness, chromaticity, and position features of internal defects and surface dirt projections; performs spatial coordinate registration on the three types of feature images to construct a multimodal feature set; and, based on the regional matching relationship of the feature images, separates and determines surface dirt and internal defects to determine the surface state of the LCD screen. This application utilizes the synergistic application of three imaging methods. Leveraging the characteristics of dark-field scattering imaging, it can accurately capture the morphology and texture details of minute surface contaminants, achieving efficient identification of various surface dirt such as dust and fingerprints. Through polarization differential imaging technology, it can clearly extract the reflective characteristics of the LCD screen surface, effectively identifying reflective defects such as surface scratches, compensating for the insufficient sensitivity in identifying minute contaminants. By combining transmitted light imaging with the acquisition of LCD screen test images, it can simultaneously acquire the brightness, color, and position information of internal defects and surface contaminant projections, providing comprehensive data support for subsequent feature differentiation. By performing spatial coordinate registration on the three types of feature images, it can eliminate image misalignment interference caused by different imaging methods, constructing a complete and accurate multimodal feature set. Furthermore, by analyzing the region matching relationships of each image in the multimodal feature set, it can achieve accurate separation of surface contaminants and internal defects, clearly distinguishing their characteristic differences. In summary, this application can significantly improve the detection accuracy and efficiency of LCD screen surface conditions, effectively reducing the rate of missed and false detections, adapting to different specifications of LCD screens, and meeting the quality inspection needs of high-end electronic devices.

[0082] In some embodiments of this application, the process of spatial coordinate registration in step S4 is described. Spatial coordinate registration aims to eliminate spatial misalignment errors in images acquired by dark-field scattering imaging, polarization difference imaging, and transmitted light imaging. By unifying the coordinate system, it establishes regional correlations between various feature images, providing a reliable spatial matching basis for the subsequent accurate separation and determination of surface contaminants and internal defects. Since the three imaging methods correspond to different cameras, and each camera has inherent differences in its installation position and imaging angle, the acquired first, second, and third feature images are located in their respective camera coordinate systems. Without coordinate registration, problems such as feature region misalignment and matching deviations will occur, affecting the accuracy of subsequent determination results. Specifically, this may include:

[0083] ① Before conducting LCD screen surface condition detection, the three cameras corresponding to dark-field scattering imaging, polarization difference imaging, and transmitted light imaging are jointly calibrated to establish a transformation matrix from each camera's coordinate system to the target coordinate system. This is a prerequisite for coordinate registration. During the joint calibration process, a standard calibration component is selected, and imaging data from the three cameras on that component are simultaneously acquired. The intrinsic and extrinsic parameters of each camera are calibrated using a calibration algorithm to eliminate the influence of factors such as camera installation deviation and lens distortion. Through joint calibration, a transformation matrix between each camera's coordinate system and the preset target coordinate system is finally established. This transformation matrix can accurately describe the mapping relationship between any coordinate point in each camera's coordinate system and the corresponding coordinate point in the target coordinate system, providing standardized parameter support for subsequent coordinate transformations of various feature images and ensuring that images acquired by different cameras can be compared and matched under a unified coordinate system.

[0084] ② Transform the region coordinates in the first, second, and third feature images to the target coordinate system: After obtaining the transformation matrix from each camera coordinate system to the target coordinate system, region extraction is performed on the first, second, and third feature images respectively to extract candidate feature regions in each image. In the first feature image, candidate dirty regions are extracted; in the second feature image, candidate surface reflection regions are extracted; and in the third feature image, candidate abnormal regions are extracted. The coordinate information of these candidate regions is then read, and this coordinate information corresponds to the camera coordinate system to which each image belongs. Subsequently, using the pre-established transformation matrix, the coordinates of all candidate regions in the three types of feature images are transformed from their respective camera coordinate system to the preset target coordinate system, achieving a unified coordinate system for the three types of feature images. Through coordinate transformation, it is ensured that candidate regions acquired by different imaging methods can be compared in spatial position, calculated in distance, and matched in feature under the same coordinate system, laying the foundation for establishing multimodal feature correspondences in the future.

[0085] ③For each candidate abnormal region in the third feature image, search for the candidate region with the closest spatial distance in the first feature image and the second feature image. If the distance is less than the preset matching threshold, establish a multimodal feature correspondence, and mark the candidate regions with a corresponding relationship as associated regions: After completing the coordinate transformation, enter the region matching and association stage. The core of this stage is to establish the correspondence between the candidate regions in the third feature image and the first and second feature images. Specifically, for each candidate abnormal region in the third feature image, based on the coordinates of this region in the target coordinate system, search for the candidate region with the closest spatial distance among all candidate dirty regions in the first feature image and all candidate surface reflection regions in the second feature image. By calculating the spatial distance between two candidate regions, determine whether there is a spatial association relationship between them. If the calculated spatial distance is less than the preset matching threshold, it means that the two candidate regions correspond to the same position on the screen surface. At this time, establish the multimodal feature correspondence between the two, and mark these two candidate regions with a corresponding relationship as associated regions. Through this step, the accurate association of candidate regions in the three types of feature images is achieved, and a complete multimodal feature correspondence is constructed, providing a clear regional matching basis for the subsequent separation and determination of surface dirt and internal defects.

[0086] According to the region matching relationships of the feature images in the multimodal feature set, perform separation and determination on surface dirt and internal defects, including:

[0087] If both the first feature image and the second feature image are empty, and the third feature image is non-empty, it is determined as an internal defect;

[0088] If either the first feature image or the second feature image is non-empty, and there are abnormal regions in the third feature image that are associated with the first feature image or the second feature image, it is determined as surface dirt;

[0089] If either the first feature image or the second feature image is non-empty, and there are no abnormal regions in the third feature image that are associated with the first feature image or the second feature image, it is determined as pure surface dirt;

[0090] If the first feature image, the second feature image, and the third feature image are all empty, it is determined as qualified.

[0091] Based on the constructed multimodal feature set and region matching relationships, and through preset judgment rules, surface dirt and internal defects are accurately separated, clarifying the specific state of the screen surface and providing clear judgment results for the quality inspection of LCD screens. This judgment process relies on the regional correlation relationships of each feature image in the multimodal feature set, combined with the feature differences of different defect types, to formulate targeted judgment rules, ensuring the accuracy and reliability of the judgment results and effectively avoiding misjudgment by confusing surface dirt with internal defects. Specifically, based on the region matching relationships of each feature image in the multimodal feature set, surface dirt and internal defects are separated and judged, including the following situations:

[0092] If both the first and second feature images are empty, and the third feature image is not empty, then it is determined to be an internal defect. An empty first feature image indicates that no candidate dirt areas with morphological and textural features were detected on the screen surface. An empty second feature image indicates that no candidate surface reflection areas with reflective properties were detected on the screen surface, indicating that there are no surface defects on the screen surface. A non-empty third feature image indicates that a candidate abnormal region was detected. Based on the coordinate registration and region association results, this candidate abnormal region is not associated with any candidate region in the first or second feature images, indicating that the abnormal region is not generated by surface dirt projection, but originates from an internal defect within the screen. Therefore, this abnormal region is determined to be an internal defect.

[0093] If the first feature image or the second feature image is not empty, and the third feature image contains an abnormal region associated with the first feature image or the second feature image, then it is determined to be surface dirt. A non-empty first feature image indicates that a candidate dirt region has been detected, and a non-empty second feature image indicates that a candidate surface reflection region has been detected. If either one is non-empty, it indicates that there may be a surface defect on the screen surface. At the same time, if the third feature image contains an abnormal region associated with the candidate dirt region or the candidate surface reflection region, combined with the region association rule, it indicates that the abnormal region is a projection of surface dirt under transmitted light imaging, and not an internal defect of the screen. Therefore, the defect corresponding to the associated region is determined to be surface dirt.

[0094] If the first feature image or the second feature image is not empty, and there is no abnormal region associated with the first feature image or the second feature image in the third feature image, then it is determined to be pure surface dirt. The fact that the first feature image or the second feature image is not empty indicates that there are candidate dirt regions or candidate surface reflection regions on the screen surface, i.e., the presence of surface-type defect characteristics. Since no abnormal region associated with the aforementioned candidate regions is detected in the third feature image, it indicates that the surface-type defect does not form a significant projection under transmitted light imaging, exists only on the screen surface, and does not affect the interior of the screen. Therefore, the defect is determined to be pure surface dirt, i.e., surface dirt that exists only on the screen surface and does not produce internal projection.

[0095] The following describes an LCD screen surface condition detection system provided in the embodiments of this application. The LCD screen surface condition detection system described below and the LCD screen surface condition detection method described above can be referred to and correspond to each other.

[0096] See Figure 2 , Figure 2 This is a schematic diagram of an LCD screen surface condition detection system disclosed in an embodiment of this application.

[0097] like Figure 2 As shown, the LCD screen surface condition detection system may include:

[0098] The dark field scattering imaging module 110 is used to illuminate the screen surface at an incident angle lower than the surface of the LCD screen, acquire the image of the scattered light from the screen surface, and obtain the first feature image characterizing the surface dirt after image processing.

[0099] The polarization difference imaging module 120 includes a polarizer set in front of the light source and an analyzer set in front of the camera, used to acquire orthogonal polarization images and parallel polarization images respectively, and obtain a second feature image characterizing the surface reflection characteristics based on the difference processing of the two.

[0100] The transmitted light imaging module 130 is used to acquire transmitted light images while the test screen is displayed on the LCD screen, and obtain a third feature image containing projections of internal defects and surface dirt after image processing.

[0101] The feature fusion determination module 140 is used to spatially register the first feature image, the second feature image and the third feature image in a unified coordinate system, and to separate and determine the surface dirt and internal defects according to the regional matching relationship of each feature image, thereby determining the surface state of the LCD screen.

[0102] As can be seen from the above technical solutions, the LCD screen surface state detection method, system, device, and readable storage medium provided in this application employs dark field scattering imaging to acquire a first feature image to extract the morphological and texture features of surface dirt; employs polarization difference imaging to acquire a second feature image to extract surface reflection characteristics; employs transmitted light imaging to acquire a third feature image while the LCD screen displays a test image, extracting the brightness, chromaticity, and position features of internal defects and surface dirt projections; performs spatial coordinate registration on the three types of feature images to construct a multimodal feature set; and, based on the regional matching relationship of the feature images, separates and determines surface dirt and internal defects to determine the surface state of the LCD screen. This application utilizes the synergistic application of three imaging methods. Leveraging the characteristics of dark-field scattering imaging, it can accurately capture the morphology and texture details of minute surface contaminants, achieving efficient identification of various surface dirt such as dust and fingerprints. Through polarization differential imaging technology, it can clearly extract the reflective characteristics of the LCD screen surface, effectively identifying reflective defects such as surface scratches, compensating for the insufficient sensitivity in identifying minute contaminants. By combining transmitted light imaging with the acquisition of LCD screen test images, it can simultaneously acquire the brightness, color, and position information of internal defects and surface contaminant projections, providing comprehensive data support for subsequent feature differentiation. By performing spatial coordinate registration on the three types of feature images, it can eliminate image misalignment interference caused by different imaging methods, constructing a complete and accurate multimodal feature set. Furthermore, by analyzing the region matching relationships of each image in the multimodal feature set, it can achieve accurate separation of surface contaminants and internal defects, clearly distinguishing their characteristic differences. In summary, this application can significantly improve the detection accuracy and efficiency of LCD screen surface conditions, effectively reducing the rate of missed and false detections, adapting to different specifications of LCD screens, and meeting the quality inspection needs of high-end electronic devices.

[0103] Optionally, the dark field scattering imaging module includes a light source angle self-calibration unit, which is used to select a defect-free area in the acquired image to calculate the gray-scale mean, and use the gray-scale mean as feedback to control the incident angle of the light source until the gray-scale mean is lower than a preset threshold, to determine the optimal dark field angle, and control the dark field scattering imaging module to acquire an image at the optimal dark field angle to generate the first feature image.

[0104] The polarization differential imaging module includes a polarization orthogonal calibration unit, which is used to determine the analyzer angle that minimizes the average gray level of the entire image after each rotation by rotating the analyzer and calculating the average gray level of the entire image. The analyzer angle is then used as the orthogonal polarization acquisition position. The polarization differential imaging module is then controlled to acquire orthogonal polarization images at the orthogonal polarization acquisition position to generate the second feature image.

[0105] Specifically,

[0106] The LCD screen surface condition detection system provided in this application embodiment can be applied to LCD screen surface condition detection equipment. Figure 3 The hardware structure block diagram of the LCD screen surface condition detection device is shown. (Refer to...) Figure 3 The hardware structure of the LCD screen surface condition detection device may include: at least one processor 1, at least one communication interface 2, at least one memory 3 and at least one communication bus 4;

[0107] In this embodiment of the application, the number of processor 1, communication interface 2, memory 3, and communication bus 4 is at least one, and processor 1, communication interface 2, and memory 3 communicate with each other through communication bus 4;

[0108] Processor 1 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0109] Memory 3 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device;

[0110] The memory stores a program, which the processor can call. The program is used for:

[0111] Dark-field scattering imaging was used to acquire the first feature image of the LCD screen surface and extract the morphological and texture features of surface dirt.

[0112] A second feature image of the LCD screen surface is acquired using polarization differential imaging to extract surface reflection characteristics.

[0113] The third feature image is acquired by using transmitted light imaging while the test screen is displayed on the LCD screen, and the brightness, color and position features of the projection of internal defects and surface dirt are extracted.

[0114] The first feature image, the second feature image, and the third feature image are spatially registered to construct a multimodal feature set;

[0115] Based on the region matching relationship of each feature image in the multimodal feature set, surface dirt and internal defects are separated and determined to determine the surface condition of the LCD screen.

[0116] Optionally, the refined and extended functions of the program can be referred to the above description.

[0117] This application embodiment also provides a readable storage medium that can store a program suitable for execution by a processor, the program being used for:

[0118] Dark-field scattering imaging was used to acquire the first feature image of the LCD screen surface and extract the morphological and texture features of surface dirt.

[0119] A second feature image of the LCD screen surface is acquired using polarization differential imaging to extract surface reflection characteristics.

[0120] The third feature image is acquired by using transmitted light imaging while the test screen is displayed on the LCD screen, and the brightness, color and position features of the projection of internal defects and surface dirt are extracted.

[0121] The first feature image, the second feature image, and the third feature image are spatially registered to construct a multimodal feature set;

[0122] Based on the region matching relationship of each feature image in the multimodal feature set, surface dirt and internal defects are separated and determined to determine the surface condition of the LCD screen.

[0123] Optionally, the refined and extended functions of the program can be referred to the above description.

[0124] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0125] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0126] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily 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 this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting the surface condition of an LCD screen, characterized in that, include: Dark-field scattering imaging was used to acquire the first feature image of the LCD screen surface and extract the morphological and texture features of surface dirt. A second feature image of the LCD screen surface is acquired using polarization differential imaging to extract surface reflection characteristics. The third feature image is acquired by using transmitted light imaging while the test screen is displayed on the LCD screen, and the brightness, color and position features of the projection of internal defects and surface dirt are extracted. The first feature image, the second feature image, and the third feature image are spatially registered to construct a multimodal feature set; Based on the region matching relationship of each feature image in the multimodal feature set, surface dirt and internal defects are separated and determined to determine the surface condition of the LCD screen.

2. The method according to claim 1, characterized in that, The method of acquiring a first feature image of the LCD screen surface using dark-field scattering imaging includes: Control the low-angle light source to illuminate the screen at a preset starting angle and acquire the first dark field image; In the first dark field image, a defect-free area is selected to calculate the gray-scale mean. If the gray-scale mean exceeds a set threshold, the incident angle of the light source is gradually reduced, the dark field image is re-acquired, and the gray-scale mean of the corresponding area is calculated until the gray-scale mean is lower than the threshold, and the optimal dark field angle is determined. A second dark field image is acquired at the optimal dark field angle. The second dark field image is then filtered, enhanced, and subjected to local adaptive threshold segmentation to extract candidate dirty regions and generate the first feature image.

3. The method according to claim 1, characterized in that, The method of acquiring a second feature image of the LCD screen surface using polarization differential imaging includes: A polarizer fixed in front of the light source controls the analyzer set in front of the camera to rotate to the orthogonal polarization position, and acquires orthogonal polarization images; The analyzer is rotated to the parallel polarization position to acquire a parallel polarization image; Calculate the difference image between the parallel polarization image and the orthogonal polarization image, and normalize and enhance the contrast of the difference image to obtain the enhanced difference image. Adaptive threshold segmentation and connected component analysis are used to extract candidate surface reflection regions from the enhanced difference image to generate the second feature image.

4. The method according to claim 1, characterized in that, The method of acquiring a third feature image using transmitted light imaging while the test image is displayed on an LCD screen includes: The LCD screen is controlled to display multiple standard test screens sequentially, and multiple frames of transmission images corresponding to each test screen are synchronously acquired through hardware triggering. The brightness of each frame of the transmission image is normalized, and a difference operation is performed between each frame and the standard template image of the corresponding test image to obtain the difference map of each frame. By fusing the difference maps corresponding to each image, a transmission anomaly map is generated. Regions are extracted from the transmission anomaly map to generate the third feature image.

5. The method according to claim 1, characterized in that, The spatial coordinate registration process includes: The three cameras corresponding to dark field scattering imaging, polarization difference imaging and transmitted light imaging are jointly calibrated in advance, and the transformation matrix from each camera coordinate system to the target coordinate system is established. Transform the region coordinates in the first feature image, the second feature image, and the third feature image to the target coordinate system; For each candidate abnormal region in the third feature image, search for the candidate region with the closest spatial distance in the first feature image and the second feature image. If the distance is less than the preset matching threshold, establish a multimodal feature correspondence, and mark the candidate regions with a correspondence as associated regions.

6. The method according to claim 5, characterized in that, The separation determination of surface dirt and internal defects according to the region matching relationship of each feature image in the multimodal feature set includes: If both the first feature image and the second feature image are empty and the third feature image is non-empty, it is determined as an internal defect; If either the first feature image or the second feature image is non-empty and there are abnormal regions in the third feature image associated with the first feature image or the second feature image, it is determined as surface dirt; If either the first feature image or the second feature image is non-empty and there are no abnormal regions in the third feature image associated with the first feature image or the second feature image, it is determined as pure surface dirt; If the first feature image, the second feature image, and the third feature image are all empty, it is determined as qualified.

7. A surface condition detection system for an LCD screen, characterized in that, Including: A dark-field scattering imaging module, which is used to irradiate the screen surface at an incident angle lower than the surface of the LCD screen, collect the scattered light image of the screen surface, and obtain the first feature image representing surface dirt after image processing; A polarization difference imaging module, including a polarizer arranged in front of the light source and an analyzer arranged in front of the camera, which is used to collect orthogonal polarization images and parallel polarization images respectively, and obtain the second feature image representing surface reflection characteristics based on the differential processing of the two; A transmitted light imaging module, which is used to collect the transmitted light image in the state where the LCD screen displays a test pattern, and obtain the third feature image containing the projection of internal defects and surface dirt after image processing; A feature fusion determination module, which is used to spatially register the first feature image, the second feature image, and the third feature image in a unified coordinate system, and separate and determine surface dirt and internal defects according to the region matching relationship of each feature image to determine the surface state of the LCD screen.

8. The system according to claim 7, characterized in that, The dark-field scattering imaging module includes a light source angle self-calibration unit, which is used to select a defect-free region in the collected image to calculate the gray-scale mean value, and use the gray-scale mean value as feedback to control the incident angle of the light source until the gray-scale mean value is lower than the preset threshold, determine the optimal dark-field angle, and control the dark-field scattering imaging module to collect images at the optimal dark-field angle to generate the first feature image; The polarization difference imaging module includes a polarization orthogonality calibration unit, which is used to determine the analyzer angle that makes the average gray-scale of the entire image the lowest as the orthogonal polarization acquisition position by rotating the analyzer and calculating the average gray-scale of the entire image after each rotation, and control the polarization difference imaging module to collect orthogonal polarization images at the orthogonal polarization acquisition position to generate the second feature image.

9. A surface condition detection device for an LCD screen, characterized in that, Including a memory and a processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the LCD screen surface state detection method according to any one of claims 1-6.

10. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements each step of the LCD screen surface state detection method as described in any one of claims 1-6.

Images