Deep Learning-Based Abnormality Detection Method and System for Liquid Crystal Displays

Abstract

The present invention discloses a method and system for detecting abnormal conditions of a liquid crystal display screen based on deep learning. The method for detecting abnormal conditions of the liquid crystal display screen of the present invention includes the steps of: collecting an image of the liquid crystal display screen, preprocessing and enhancing the collected image to obtain enhanced image data; performing preliminary feature extraction on the enhanced image data, and then performing feature optimization processing to obtain a comprehensive feature set; constructing an abnormal condition detection model based on deep learning, performing automated classification detection processing on the comprehensive feature set to obtain a detection result, and realizing the detection of abnormal conditions of the liquid crystal display screen. The present invention enhances the image, greatly improving the brightness, color, and contrast of the image, especially the expressiveness of subtle abnormal areas, making subsequent feature extraction and classification more accurate; optimizing the preliminary features in the image, and realizing accurate detection of abnormal conditions of the liquid crystal display screen.

Description

Technical Field

[0001] The present invention relates to the technical field of liquid crystal display (LCD) abnormality detection, and particularly to a method and system for detecting LCD abnormalities based on deep learning. Background Art

[0002] With the wide application of liquid crystal display (LCD) technology, liquid crystal displays have become an important part of modern electronic devices. From smartphones, tablets, laptops to TVs, industrial displays, etc., the quality of the LCD directly affects the user experience of the final product. However, due to the complexity of the LCD manufacturing process, various defects are likely to occur during production, such as bright dots, dark dots, color differences, uneven brightness, etc. These defects not only affect the display effect, but in some cases can even cause the entire LCD to fail. Therefore, the LCD needs to undergo strict quality inspection before leaving the factory to ensure its good display effect.

[0003] Traditional LCD detection mainly relies on manual visual inspection and simple image processing algorithms. Although manual visual inspection can detect obvious defects, it has low efficiency, insufficient accuracy, and is easily affected by human factors. Especially in mass production, it is difficult to guarantee the efficiency and consistency of manual inspection. And traditional detection algorithms based on image processing usually rely on simple methods such as edge detection and color difference analysis. These methods have limited detection capabilities for complex LCD abnormalities, especially for tiny bright dots, dark dots, and subtle color difference problems. With the improvement of LCD resolution and the continuous increase of user requirements for display effects, traditional detection methods can no longer meet the current detection needs.

[0004] However, the above technologies have at least the following technical problems: In LCD detection, the accuracy of image processing for the LCD is poor and the optimization of feature processing is not comprehensive. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the defects of the above-mentioned prior art, and provide a method and system for detecting LCD abnormalities based on deep learning, which performs enhancement processing on images to greatly improve the brightness, color, and contrast of the images, especially the expressiveness of subtle abnormal areas; and optimizes the preliminary features in the images to form a more expressive feature set among the features, so as to achieve accurate detection of LCD abnormalities.

[0006] To this end, the present invention adopts the following technical solutions.

[0007] In a first aspect, the present invention provides a method for detecting LCD abnormalities based on deep learning, which includes the steps:

[0008] S1. Collect the LCD screen image, preprocess and enhance the collected image to obtain the enhanced image data;

[0009] S2. Perform preliminary feature extraction on the enhanced image data, and then perform feature optimization processing to obtain a comprehensive feature set;

[0010] S3. Construct an anomaly detection model based on deep learning, perform automated classification detection processing on the comprehensive feature set to obtain a detection result, and realize the anomaly detection of the LCD screen.

[0011] Furthermore, in S1, the preprocessed LCD screen image data is enhanced using a recursive adaptive multi-layer enhancement algorithm to obtain the enhanced image data; the recursive adaptive multi-layer enhancement algorithm performs multi-dimensional adjustments on the color and contrast of the LCD screen image to enhance the expressiveness of the subtle anomaly regions in the image and provide higher-quality input for subsequent processing.

[0012] Even further, during the implementation of the recursive adaptive multi-layer enhancement algorithm, perform dynamic color mapping and fusion processing on the preprocessed LCD screen image data; through non-linear operations on different color channels, perform complex interaction processing on the pixel values of the red, green, and blue channels to enhance the color hierarchy and contrast of the image.

[0013] Even further, during the implementation of the recursive adaptive multi-layer enhancement algorithm, perform recursive weighted processing on the image data after dynamic color mapping and fusion processing. Through layer-by-layer recursive calculation, perform layer-by-layer weighted superposition on the pixel values, and prevent noise amplification through non-linear suppression to ensure that the local information of the image is effectively enhanced without generating excessive artifacts or noise accumulation.

[0014] Even further, during the implementation of the recursive adaptive multi-layer enhancement algorithm, perform adaptive local contrast enhancement processing on the image after recursive weighted processing.

[0015] Even further, after recursive weighting and contrast enhancement processing, finally, fuse all levels of the image to ensure that the pixel information at different levels can complement each other to form a comprehensive enhancement result; the goal of multi-layer fusion is to integrate the pixel values at different enhancement levels into the final output image through weighted averaging.

[0016] Furthermore, S2 specifically includes:

[0017] The existing feature engineering techniques are used to perform preliminary feature extraction on the enhanced image data to obtain preliminary image feature data. The preliminary image feature data is then subjected to feature analysis and optimization processing using a non-linear cross-feature interaction and multi-dimensional projection optimization algorithm to obtain a comprehensive feature set. The non-linear cross-feature interaction and multi-dimensional projection optimization algorithm realizes the refined processing of image feature data through complex cross-feature interaction processing, multi-dimensional non-linear projection and normalization processing, as well as multi-dimensional feature cross-optimization.

[0018] Furthermore, in the implementation process of the non-linear cross-feature interaction and multi-dimensional projection optimization algorithm, non-linear cross-feature interaction processing is performed on the preliminary image feature data to form more expressive feature data. After the interaction matrix is generated, in order to map the features that have undergone interaction processing to a multi-dimensional non-linear space, non-linear mapping is used to further amplify the subtle differences of the features, and multi-dimensional non-linear projection processing is performed on them. The projected feature matrix is normalized to eliminate the numerical imbalance between dimensions and ensure that the finally output feature matrix has a consistent numerical range in each dimension. Multi-dimensional feature cross-optimization processing is performed on the normalized feature matrix.

[0019] Further, step S3 specifically includes: constructing an anomaly detection model based on deep learning using a convolutional neural network, processing the comprehensive feature set through the anomaly detection model based on deep learning to obtain a detection result, and the detection result is a binary classification result indicating whether the current state of the liquid crystal screen is normal; when an anomaly of the liquid crystal screen is detected, the anomaly detection model based on deep learning further classifies the anomaly type; for the detected anomaly, the anomaly detection model based on deep learning also evaluates its severity.

[0020] In a second aspect, the present invention provides a liquid crystal screen anomaly detection system based on deep learning for implementing the above-mentioned liquid crystal screen anomaly detection method, which includes:

[0021] An image processing unit: collecting an image of the liquid crystal screen, preprocessing and enhancing the collected image to obtain enhanced image data;

[0022] A feature extraction and optimization processing unit: performing preliminary feature extraction on the enhanced image data, and then performing feature optimization processing to obtain a comprehensive feature set;

[0023] An anomaly detection unit: constructing an anomaly detection model based on deep learning, automatically classifying and detecting the comprehensive feature set to obtain a detection result, and realizing the anomaly detection of the liquid crystal screen.

[0024] The beneficial effects of the present invention are:

[0025] 1. The image is enhanced, significantly improving the brightness, color, and contrast of the image, especially the expressiveness of subtle abnormal areas (such as bright spots, dark spots, and color differences), making subsequent feature extraction and classification more accurate.

[0026] 2. The preliminary features (such as color, texture, edges, etc.) in the image are optimized, enabling the formation of a more expressive feature set among the features, ensuring that the comprehensive feature set can cover all the information of the liquid crystal display (LCD) screen image, especially the tiny abnormal areas, and achieving accurate detection of LCD screen abnormalities. Brief Description of the Drawings

[0027] Figure 1 It is a flowchart of the method for detecting LCD screen abnormalities based on deep learning according to the present invention;

[0028] Figure 2 It is an architecture diagram of the system for detecting LCD screen abnormalities based on deep learning according to the present invention. Detailed Embodiments

[0029] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0030] Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0031] Referring to the attached Figure 1 , which shows a flowchart of the method for detecting LCD screen abnormalities based on deep learning provided by an embodiment of the present invention. The method includes the following steps:

[0032] S1. Collect LCD screen images, perform preprocessing and enhancement processing on the collected images to obtain enhanced image data.

[0033] Professional technicians select a suitable industrial camera to collect images of the LCD screen. The camera should be equipped with appropriate light sources (such as ring light sources, bar light sources) to ensure the lighting uniformity of the LCD screen image, avoid reflection and shadow interference, and at the same time, according to the size of the LCD screen, adopt the full-screen shooting or sub-region shooting method. For large LCD screens, sub-region shooting is used and then image stitching is performed to ensure complete coverage of the detail areas and obtain the LCD screen image.

[0034] Perform preprocessing such as denoising, size adjustment, and normalization on the LCD screen image to obtain the preprocessed LCD screen image data. The preprocessing method is a well-known technical means to those skilled in the art and will not be elaborated here.

[0035] Perform enhancement processing on the preprocessed LCD screen image data using a recursive adaptive multi-layer enhancement algorithm to obtain the enhanced LCD screen image data; the recursive adaptive multi-layer enhancement algorithm performs multi-dimensional adjustment on the color and contrast of the LCD screen image to enhance the expressiveness of subtle abnormal areas (such as bright spots, dark spots, and color differences) in the image and provide higher-quality input for subsequent processing. The specific implementation process is as follows:

[0036] First, perform dynamic color mapping and fusion processing on the preprocessed LCD screen image data; through non-linear operations on different color channels, perform complex interactive processing on the pixel values of the red, green, and blue channels to enhance the color hierarchy and contrast of the image, especially with significant effects in color difference anomaly detection. The specific processing formula is as follows:

[0037]

[0038] Among them, is the pixel value of the red channel, indicating the red brightness of the image at position ; is the pixel value of the green channel, indicating the green brightness of the image at position ; is the pixel value of the blue channel, indicating the blue brightness of the image at position ; is the pixel value after dynamic color mapping, used to output the enhanced image. The above formula makes the pixel values of different channels produce complex non-linear interactions through a combination of logarithmic, square, square root, and fractional operations, aiming to enhance the color contrast of the image and especially magnify the changes in color differences.

[0039] Furthermore, perform recursive weighting processing on the image data after dynamic color mapping and fusion processing. Through layer-by-layer recursive calculation, perform layer-by-layer weighted superposition on the pixel values, and prevent noise amplification through non-linear suppression to ensure that the local information of the image is effectively enhanced without generating excessive artifacts or noise accumulation. The formula for recursive weighting processing is as follows:

[0040]

[0041] Among them, is the pixel value of the image after the -th layer of recursive processing at ; is the image after the -th layer of recursive processing at The pixel value at ; is the weighting coefficient of the th layer, which determines the weight of each layer's recursive processing; is the non - linear suppression coefficient of the th layer during the recursive process, which controls the intensity of recursive enhancement; is the adjustment coefficient in the sine function of the th layer, which controls the periodic change of the sine wave in each layer's recursion; is the attenuation parameter in the exponential function of the th layer, which controls the attenuation speed of the high - level pixel values during the recursive process; By introducing non - linear operations (logarithm, sine, exponential), the recursive process becomes more hierarchical. Each layer in the recursion not only weights according to the pixel values of the previous layer but also goes through the processing of logarithm and sine functions, making the brightness and color information in the image more delicate; The combination of the square in the denominator and the exponential function ensures that the recursion of each layer does not increase infinitely, avoiding the risk of over - amplifying brightness changes and noise, and at the same time realizing the attenuation of the recursive layer number through the exponential function. Specifically, the specific number of recursive layers is set according to the specific scenario and will not be elaborated here.

[0042] Furthermore, for the image after recursive weighting processing perform adaptive local contrast enhancement processing. By adaptively adjusting the contrast in the image, areas with large brightness differences (such as bright or dark spots on the liquid crystal screen) are made more prominent, while areas with small brightness differences are smoothed to avoid noise amplification caused by over - enhancement. The formula for adaptive local contrast enhancement processing is as follows:

[0043]

[0044] where, is the pixel value after adaptive contrast enhancement of the th layer, which is used to output the brightness of the enhanced image; is the average brightness value within the local area of the image, representing the brightness balance of the local neighborhood; is the standard deviation of the local brightness, which is used to measure the degree of dispersion of the brightness values within the local area.

[0045] Furthermore, after multi - layer recursion and contrast enhancement processing, finally, all levels of the image need to be fused to ensure that the pixel information of different levels can complement each other to form a comprehensive enhancement result. The goal of multi - layer fusion is to integrate the pixel values of different enhancement levels into the final output image through weighted average, which can not only maintain the global brightness and contrast but also highlight the local details. The multi - layer fusion formula is as follows:

[0046]

[0047] Among them, is the pixel value of the finally enhanced image; is the pixel value of each layer of the image after local contrast enhancement; is the weighting coefficient of each layer; is the total number of processing layers, which is used to define the number of recursive layers and the number of fusion layers.

[0048] S2. Perform preliminary feature extraction on the enhanced image data, and then perform feature optimization processing to obtain a comprehensive feature set.

[0049] Use existing feature engineering techniques to perform preliminary feature extraction on the enhanced image data to obtain image feature data such as texture features, color features, and edge features as preliminary image feature data. Use the non-linear cross-feature interaction and multi-dimensional projection optimization algorithm to perform feature analysis and optimization processing on the preliminary image feature data to obtain a comprehensive feature set; the non-linear cross-feature interaction and multi-dimensional projection optimization algorithm realizes the refined processing of image feature data through complex cross-feature interaction processing, multi-dimensional non-linear projection and normalization processing, and multi-dimensional feature cross-optimization.

[0050] The specific implementation process is as follows:

[0051] First, in order to combine different features and break the independence between features, perform non-linear cross-feature interaction processing on the preliminary image feature data to form more expressive feature data. Perform point-by-point calculation among all preliminary features and construct an interaction matrix through non-linear operations . Each element in this interaction matrix represents the feature interaction value at a specific pixel point . The specific formula for the interaction processing is as follows:

[0052]

[0053] Among them, is the interaction feature matrix value at the position, which is generated by the non-linear interaction of each feature in the preliminary feature matrix ; is the value of the rd feature extracted preliminarily at the pixel point , which can be a texture, color or edge feature; is the value of the th feature extracted preliminarily at the pixel point , which can be a texture, color or edge feature; is the feature and the feature The interaction coefficient between them controls the association strength between these features; is a small constant used to avoid a zero denominator; is at the total number of features at the position.

[0054] Furthermore, after the interaction matrix is generated, in order to map the features that have undergone interaction processing into a multi-dimensional non-linear space and further amplify the subtle differences of the features by using non-linear mapping, multi-dimensional non-linear projection processing is performed on it. The projection processing formula is as follows:

[0055]

[0056] where, is the value at the pixel point in the projected feature matrix corresponding to the dimension; is the weight coefficient of the projection dimension, which controls the influence of different projection dimensions on the final feature value; is the scaling parameter of the projection, which controls the expansion amplitude of the feature value in different dimensions.

[0057] Furthermore, the projected feature matrix is normalized to eliminate the numerical imbalance between dimensions and ensure that the final output feature matrix has a consistent numerical range in each dimension.

[0058] Furthermore, multi-dimensional feature cross-optimization processing is performed on the normalized feature matrix. By cross-operation, the synergistic effect between features in different dimensions is maximized, and the discrimination ability and expression ability of the feature matrix are further improved. The cross-optimization is performed through a non-linear cross function, and the formula is as follows:

[0059]

[0060] where, is the value at the pixel point in the finally multi-dimensionally optimized feature matrix; is the value at the pixel point in the feature matrix of the dimension after enhancement and normalization; is the cross-optimization weight of the dimension, which controls the contribution of this dimension to the finally optimized feature; is the parameter that controls the feature cross frequency of the dimension; is the number of optimized feature dimensions.

[0061] Finally, constitutes a comprehensive feature set to provide feature data for liquid crystal display abnormal detection.

[0062] S3. Construct an anomaly detection model based on deep learning, perform automated classification detection processing on the comprehensive feature set, obtain the detection result, and achieve anomaly detection of the liquid crystal display screen.

[0063] First, construct an anomaly detection model based on deep learning. The construction process is as follows: Model design: Use a convolutional neural network to perform automated classification processing on the feature data in the comprehensive feature set of the liquid crystal display screen. Take the comprehensive feature set as the input of the anomaly detection model based on deep learning, perform dimensionality reduction through a pooling layer, and then perform classification through a fully connected layer to output the detection result. Model training: In order to enable the anomaly detection model based on deep learning to learn the differences between abnormal and normal liquid crystal display screens, define a loss function (such as the cross-entropy loss function) to measure the gap between the predicted value and the true label of the model. Then, through the backpropagation algorithm, update the parameters (such as the weights and biases of the convolutional kernels) in the anomaly detection model based on deep learning according to the gradient of the loss value. At the same time, to prevent overfitting, add a regularization term, that is, randomly discard some neurons during each layer of training to improve the generalization ability of the model. Model evaluation: Evaluate the model by calculating evaluation metrics such as accuracy, recall rate, and F1 score. After the evaluation passes, the anomaly detection model based on deep learning is constructed.

[0064] Process the comprehensive feature set through the anomaly detection model based on deep learning to obtain the detection result. The detection result is a binary classification result, indicating whether the current state of the liquid crystal display screen is normal. Through the calculation output of the model, the liquid crystal display screen image will be classified as "normal" or "abnormal". This result is based on the feature patterns learned by the model during training. By analyzing the input comprehensive feature set through the anomaly detection model based on deep learning, such as features like color, texture, and edges, it can identify tiny defects in the liquid crystal display screen. Normal result: If the detection result shows that the state of the liquid crystal display screen is normal, it means that no obvious bright spots, dark spots, color differences, etc. are detected during the detection. After feature classification by the model, it is considered that the comprehensive feature set meets the standards of a normal liquid crystal display screen. Abnormal result: If the model's judgment result is "abnormal", it means that some features are detected to be different from those of a normal liquid crystal display screen. The anomalies may include bright spots, dark spots, or uneven color differences, etc.

[0065] When an LCD screen anomaly is detected, the deep learning-based anomaly detection model further classifies the anomaly type. Based on the multi-classification capabilities of the deep learning anomaly detection model, anomalies can be divided into: Bright spots: The brightness values ​​of certain areas of the LCD screen are significantly higher than normal areas. This phenomenon is caused by pixel failure or backlight anomaly in the liquid crystal layer. The detection results will mark these bright spots and give their specific locations. Dark spots: The brightness values ​​of certain areas on the LCD screen are lower than normal and appear as darker areas. The model can automatically identify these dark spots and locate them by analyzing the brightness distribution of the image. Color difference: Color difference often refers to the color display of certain parts of the LCD screen being inconsistent with the surrounding areas, usually manifested as color shift or saturation abnormality. The model can detect these color difference problems and determine their severity by analyzing the features in the color channel. Uneven brightness: There is a large unevenness in the brightness of different areas on the LCD screen, which is manifested as one area being too bright or too dark. The model will detect uneven brightness and mark the problem areas through a detailed analysis of brightness information.

[0066] Finally, for the detected anomalies, the deep learning-based anomaly detection model will also evaluate their severity. By analyzing the size of the abnormal area, the degree of brightness deviation or color deviation, each anomaly is quantitatively evaluated and the following information is provided to the user: Minor anomalies: Minor bright spots or dark spots. Such problems may have little effect on the display effect of the LCD screen. Moderate anomalies: Obvious bright spots, dark spots or color differences. These problems may have a certain impact on the user experience and require further correction. Severe anomalies: Severe color differences, large areas of uneven brightness or dark spots may cause the LCD screen to not work properly and must be replaced or repaired. Help operators make maintenance or replacement decisions in a timely manner. Realize abnormal detection of LCD screens.

[0067] The beneficial effects of this embodiment are:

[0068] 1. The image is enhanced by a recursive adaptive multi-layer enhancement algorithm, which greatly improves the brightness, color and contrast of the image, especially the expression of subtle abnormal areas (such as bright spots, dark spots, and color differences). This enhancement process not only strengthens the local details in the image, but also effectively reduces noise and ensures the global consistency of the image, greatly improving the detection ability of tiny defects and making subsequent feature extraction and classification more accurate.

[0069] 2. Through the non - linear cross - feature interaction and multi - dimensional projection optimization algorithm, the preliminary features in the image (such as color, texture, edges, etc.) are deeply interacted and optimized. This algorithm breaks the independence between features, enabling the features to form a more expressive feature set through complex non - linear interaction operations. Further, through multi - dimensional projection, normalization, and feature cross - optimization, the feature data has higher discrimination and expression capabilities, ensuring that the comprehensive feature set can cover all the information of the LCD screen image, especially the tiny abnormal areas.

[0070] Refer to the appendix Figure 2 , which shows the architecture diagram of the deep - learning - based LCD screen anomaly detection system provided by another embodiment of the present invention, used to implement the above - mentioned LCD screen anomaly detection method, and it includes:

[0071] Image processing unit: Collect the LCD screen image, pre - process and enhance the collected image to obtain the enhanced image data;

[0072] Feature extraction and optimization processing unit: Perform preliminary feature extraction on the enhanced image data, and then perform feature optimization processing to obtain a comprehensive feature set;

[0073] Anomaly detection unit: Construct a deep - learning - based anomaly detection model, perform automated classification detection processing on the comprehensive feature set to obtain a detection result, and achieve LCD screen anomaly detection.

[0074] It should be noted that each unit in the above - mentioned deep - learning - based LCD screen anomaly detection system can be implemented in whole or in part by software, hardware, and their combination. The above - mentioned units can be embedded in the processor of the computer device in hardware form or be independent of it, or be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above - mentioned modules. For the specific limitations of a deep - learning - based LCD screen anomaly detection system, refer to the limitations of a deep - learning - based LCD screen anomaly detection method in the above text. The two have the same functions and effects and will not be elaborated here.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for detecting abnormalities of a liquid crystal screen based on deep learning, characterized in that: Includes steps: S1. Collecting LCD screen images, preprocessing and enhancing the collected images to obtain enhanced image data; S2. Perform preliminary feature extraction on the enhanced image data, and then perform feature optimization processing to obtain a comprehensive feature set, which specifically includes: Perform preliminary feature extraction on the enhanced image data to obtain preliminary image feature data, and use nonlinear cross-feature interaction and multidimensional projection optimization algorithm to perform feature analysis and optimization on the preliminary image feature data to obtain a comprehensive feature set; the nonlinear cross-feature interaction and multidimensional projection optimization algorithm achieves refined processing of image feature data through complex cross-feature interaction processing, multidimensional nonlinear projection and normalization processing, and multidimensional feature cross-optimization; The specific formula for cross-feature interaction processing is as follows: in, is The interactive feature matrix value of the position is determined by the preliminary feature matrix The nonlinear interaction of the features in is generated; It is the first Features at the pixel The value at is a texture, color or edge feature; It is the first Features at the pixel The value at is a texture, color or edge feature; It is a feature With features The interaction coefficient between them controls the strength of the association between these features; is a small constant used to avoid the denominator being zero; is The total number of features at the location; S3. Construct an anomaly detection model based on deep learning, perform automatic classification detection processing on the comprehensive feature set, and obtain a detection result; In S1, the pre-processed LCD screen image data is enhanced using a recursive adaptive multi-layer enhancement algorithm to obtain enhanced image data; the recursive adaptive multi-layer enhancement algorithm performs multi-dimensional adjustment on the color and contrast of the LCD screen image; In the process of implementing the recursive adaptive multi-layer enhancement algorithm, the pre-processed LCD screen image data is subjected to dynamic color mapping and fusion processing; through nonlinear operations of different color channels, the pixel values ​​of the red, green and blue channels are subjected to complex interactive processing to enhance the color level and contrast of the image; the image data after dynamic color mapping and fusion processing is subjected to recursive weighted processing, and the pixel values ​​are weightedly superimposed layer by layer through layer-by-layer recursive calculation, and noise amplification is prevented through nonlinear suppression; the image after recursive weighted processing is subjected to adaptive local contrast enhancement processing.

2. The method for detecting abnormalities of a liquid crystal screen based on deep learning according to claim 1, characterized in that: After recursive weighting and contrast enhancement processing, it is necessary to fuse all levels of images to ensure that the pixel information at different levels can complement each other and form a comprehensive enhancement result; the goal of multi-layer fusion is to integrate the pixel values ​​of different enhancement levels into the final output image by weighted averaging.

3. The method for detecting abnormalities of a liquid crystal screen based on deep learning according to claim 1, characterized in that: In the process of implementing the nonlinear cross-feature interaction and multi-dimensional projection optimization algorithm, the preliminary image feature data is processed by nonlinear cross-feature interaction to form more expressive feature data; After the interaction matrix is ​​generated, in order to map the features that have been interactively processed to a multidimensional nonlinear space, nonlinear mapping is used to further amplify the subtle differences in the features, and multidimensional nonlinear projection processing is performed on them; the projected feature matrix is ​​normalized to eliminate the numerical imbalance between dimensions and ensure that the final output feature matrix has a consistent numerical range in each dimension; and the normalized feature matrix is ​​subjected to multidimensional feature cross-optimization processing.

4. The method for detecting abnormalities of a liquid crystal screen based on deep learning according to claim 1, characterized in that: The S3 specifically includes: using a convolutional neural network to build an anomaly detection model based on deep learning, processing the comprehensive feature set through the anomaly detection model based on deep learning, and obtaining a detection result, wherein the detection result is a binary classification result, indicating whether the current state of the LCD screen is normal; when an abnormality of the LCD screen is detected, the anomaly detection model based on deep learning further classifies the anomaly type; for the detected anomaly, the anomaly detection model based on deep learning also evaluates its severity.

5. A liquid crystal screen anomaly detection system based on deep learning, used to implement the liquid crystal screen anomaly detection method according to any one of claims 1 to 4, characterized in that: include: Image processing unit: collects LCD screen images, performs pre-processing and enhancement processing on the collected images, and obtains enhanced image data; Feature extraction and optimization processing unit: perform preliminary feature extraction on the enhanced image data, and then perform feature optimization processing to obtain a comprehensive feature set; Anomaly detection unit: Build an anomaly detection model based on deep learning, perform automatic classification detection processing on the comprehensive feature set, obtain detection results, and realize LCD screen anomaly detection.

Images