A display panel module defect detection method and system

Abstract

The present invention provides a module defect detection method and system for a display panel, and relates to the technical field of panel defect detection. The method flow is: feature region interception, Gaussian filtering and image enhancement are performed on a grayscale panel image to obtain an enhanced training image and an enhanced detection image; singular value decomposition, threshold segmentation, opening operation and edge detection are performed on a color panel image to obtain a foreground training image and a foreground detection image; model training is performed based on the enhanced training image and the foreground training image to obtain a defect detection model; defect detection is performed on the enhanced detection image and the foreground detection image using the defect detection model to obtain a defect detection result; priority and confidence verification is performed on the defect detection result to obtain a final defect detection result. The present invention adopts a combination of grayscale images and color images to significantly improve the model defect detection accuracy, and solves the problem that the existing panel defect detection is prone to missed detection and false detection.

Description

Technical Field

[0001] The present invention relates to the technical field of panel defect detection, and in particular to a method and system for detecting module defects of a display panel. Background Art

[0002] The display panel industry is booming, and market demand continues to grow. Improving the production efficiency and increasing production capacity of display panels while reducing production costs and improving product quality is a major development trend. Defect detection methods with artificial intelligence as the core are gradually being applied to various process stages of display panel production, replacing manual visual inspection while effectively improving the stability and effectiveness of quality control.

[0003] During the module shipment inspection stage of display panels, the number of display panel product images is huge, and the accuracy of product defect detection is extremely high. Due to the high resolution of panel images, small defect size, and extremely slight defect features, the use of end-to-end deep learning target detection models alone will cause more missed and misjudgments in defect detection, reducing the product shipment yield. Summary of the invention

[0004] The present invention provides a method and system for detecting module defects of a display panel, which solves the problem that the existing display panel defect detection technology performs defect detection based on an end-to-end target detection model, which is prone to missed detection and false detection.

[0005] In a first aspect, an embodiment of the present invention provides a method for detecting module defects of a display panel, the method comprising the following process:

[0006] Acquire a panel image of the display panel, and construct a training sample set and a detection sample set based on the panel image;

[0007] Performing feature region interception, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the detection sample set to obtain enhanced training images and enhanced detection images;

[0008] Performing singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set to obtain a foreground training image and a foreground detection image;

[0009] Performing feature annotation on the enhanced training image and the foreground training image, and inputting the feature-annotated enhanced training image and the foreground training image into the two-stage target detection model for model training to obtain a defect detection model;

[0010] Using the defect detection model to perform defect detection on the enhanced detection image and the foreground detection image to obtain several defect detection results;

[0011] Several defect detection results are verified in a manner combining defect priority and confidence to obtain the final defect detection result.

[0012] In the above embodiment, the present invention first scales the grayscale image of the display panel (Panel) module, uses Gaussian filtering to denoise the cropped image, uses gamma transformation and histogram equalization to enhance the cropped image, and can significantly visualize the slight defect features while reducing the image size; at the same time, singular value decomposition and normalization are performed on the color image of the display panel (Panel) module, and the image is denoised again using binary threshold segmentation and opening operation, and the foreground and background images are segmented using edge detection to filter invalid pixel areas. Then, the processed image is trained using a two-stage target detection algorithm to obtain a defect detection model. Finally, the defect detection model is used to perform defect detection on the above-mentioned processed image, and the detection results of each image are summarized. The defect priority and confidence are combined to calculate and output the final unique defect detection result. The present invention uses a combination of grayscale images and color images to significantly improve the model defect detection accuracy, assist factories in improving their shipment yield, and bring obvious economic benefits.

[0013] As some optional implementations of the present application, the process of performing feature region interception, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the detection sample set is as follows:

[0014] Perform ROI interception on the grayscale panel images in the training sample set and the detection sample set to obtain the corresponding ROI image;

[0015] Performing Gaussian filtering on the ROI images corresponding to the grayscale panel images in the training sample set and the detection sample set, and switching the Gaussian filtered ROI images to a high magnification to obtain the corresponding denoised images;

[0016] Gamma transformation and histogram equalization are performed on the denoised images corresponding to the grayscale panel images in the training sample set and the detection sample set to obtain the final enhanced training image and enhanced detection image.

[0017] In the above embodiment, the present invention scales the grayscale panel image of the display panel module, uses a Gaussian filter to denoise the cropped image, and uses a gamma transform and histogram equalization to enhance the cropped image, thereby significantly visualizing slight defect features while reducing the image size.

[0018] As some optional implementations of the present application, the process of performing singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set is as follows:

[0019] Perform singular value decomposition and regularization processing on the color panel images in the training sample set and the detection sample set to obtain the corresponding enhanced images;

[0020] Performing binary threshold segmentation and opening operation on the enhanced images corresponding to the color panel images in the training sample set and the detection sample set to obtain the corresponding binary images;

[0021] Edge detection is performed on the binary images corresponding to the color panel images in the training sample set and the detection sample set to obtain the final foreground training image and foreground detection image.

[0022] In the above embodiment, the present invention performs singular value classification and normalization on the color image of the display panel (Panel) module, and uses binary threshold segmentation and opening operation to denoise the panel image, use edge detection to segment the foreground and background images, filter out invalid pixel areas, and reduce the image size.

[0023] As some optional implementations of the present application, the two-stage target detection model includes Faster RCNN and Mask RCNN.

[0024] In the above embodiment, the present invention adopts a two-stage target detection model to perform defect detection, and achieves high-precision target detection through two-stage processing.

[0025] As some optional implementation methods of the present application, the process of inputting the feature-annotated enhanced training image and the foreground training image into the two-stage target detection model for model training is as follows:

[0026] The feature-annotated enhanced training image and the foreground training image are respectively input into the two-stage object detection model, and the candidate regions are generated from the enhanced training image and the foreground training image by the two-stage object detection model;

[0027] Classify the candidate area to determine whether the candidate area contains the target object and the type of the target object, and use the bounding box regression technology to adjust the position and size of the candidate area to locate the target object;

[0028] The classification loss and regression loss of the two-stage target detection model are calculated to obtain the classification loss value and the regression loss value, and the model weights are updated according to the classification loss value and the regression loss value to optimize the model performance.

[0029] In the above embodiment, the present invention trains the two-stage target detection model so that the model can learn and detect defect features.

[0030] As some optional implementations of the present application, the process of using the defect detection model to perform defect detection on the enhanced detection image and the foreground detection image is as follows:

[0031] The enhanced inspection image is input into the defect detection model, and the defect features are located and classified by the defect detection model to obtain a number of grayscale defect target frames;

[0032] The foreground detection image is input into the defect detection model, and the defect features are located and classified by the defect detection model to obtain several color defect target frames.

[0033] In the above-mentioned embodiment, the present invention performs defect detection through a trained defect detection model, and is able to obtain a number of grayscale defect target frames and color defect target frames.

[0034] As some optional implementation methods of the present application, the process of verifying several defect detection results in a manner combining defect priority and confidence is as follows:

[0035] Obtain the defect type, location, size, and confidence level corresponding to the grayscale defect target frame and the color defect target frame;

[0036] Defects detected in the grayscale defect target frame and the color defect target frame are prioritized according to defect severity to obtain a defect priority classification result; wherein the defect severity is comprehensively determined according to the defect type, location and size;

[0037] The grayscale defect target frame and the color defect target frame are screened according to the defect priority classification results, and the final defect detection result is output in combination with the confidence level.

[0038] In the above embodiment, the present invention screens the grayscale defect target frame and the color defect target frame in a manner combining defect priority and confidence, which can significantly improve the model defect detection accuracy.

[0039] As some optional implementations of the present application, the target detection model adopts Faster RCNN or YOLO.

[0040] In a second aspect, the present invention provides a module defect detection system for a display panel, the system comprising:

[0041] A panel image acquisition unit, the panel image acquisition unit is used to acquire a panel image of the display panel, and construct a training sample set and a detection sample set based on the panel image;

[0042] A grayscale image processing unit, the grayscale image processing unit is used to perform feature region interception, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the detection sample set to obtain enhanced training images and enhanced detection images;

[0043] A color image processing unit, the color image processing unit is used to perform singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set to obtain a foreground training image and a foreground detection image;

[0044] A detection model training unit, wherein the detection model training unit is used to perform feature annotation on the enhanced training image and the foreground training image, and input the feature-annotated enhanced training image and the foreground training image into a two-stage target detection model for model training to obtain a defect detection model;

[0045] A panel defect detection unit, wherein the panel defect detection unit uses a defect detection model to perform defect detection on the enhanced detection image and the foreground detection image to obtain a number of defect detection results;

[0046] A defect comprehensive verification unit verifies a plurality of defect detection results in a manner combining defect priority and confidence to obtain a final defect detection result.

[0047] In a third aspect, the present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the module defect detection method for a display panel when executing the computer program.

[0048] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for detecting module defects of a display panel is implemented.

[0049] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements the module defect detection method for a display panel.

[0050] The beneficial effects of the present invention are as follows: the present invention first scales the grayscale image of the display panel (Panel) module, uses Gaussian filtering to denoise the cropped image, uses gamma transformation and histogram equalization to enhance the cropped image, and can significantly visualize the slight defect features while reducing the image size; at the same time, singular value decomposition and normalization are performed on the color image of the display panel (Panel) module, and the image is denoised again using binary threshold segmentation and opening operation, and the foreground and background images are segmented using edge detection, which can filter the invalid pixel area. Then, the processed image is trained using a two-stage target detection algorithm to obtain a defect detection model. Finally, the defect detection model is used to perform defect detection on the above-mentioned processed image, and the detection results of each image are summarized, and the final unique defect detection result is calculated and output by combining defect priority and confidence. The present invention uses a combination of grayscale images and color images to significantly improve the model defect detection accuracy, assist factories in improving their shipment yield, and bring obvious economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0052] Figure 1 is a schematic diagram of the computer device structure of the hardware operating environment described in the embodiment of the present invention;

[0053] Figure 2 is a flow chart of a module defect detection method according to an embodiment of the present invention;

[0054] Figure 3 is the original pixel distribution diagram of the panel image described in the embodiment of the present invention;

[0055] Figure 4 is a pixel distribution diagram of the panel image after Gaussian filtering according to the embodiment of the present invention;

[0056] Figure 5 is a pixel distribution diagram after gamma transformation of the panel image according to the embodiment of the present invention;

[0057] Figure 6 is a pixel distribution diagram after histogram equalization of the panel image according to the embodiment of the present invention;

[0058] Figure 7 It is a structural block diagram of the module defect detection system described in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0060] In order to solve the problem that the existing display panel defect detection technology is based on an end-to-end target detection model for defect detection, which is prone to missed detection and false detection. This application provides a display panel module defect detection method and system. Before introducing the specific technical solution of this application, the hardware operating environment involved in the embodiment of this application is first introduced.

[0061] See also Figure 1 , Figure 1 A schematic diagram of the computer device structure of the hardware operating environment involved in the embodiment of the present application.

[0062] like Figure 1 As shown, the computer device may include: a processor, such as a central processing unit (CPU), a communication bus, a user interface, a network interface, and a memory. Among them, the communication bus is used to realize the connection and communication between these components. The user interface may include a display screen (Display), an input unit such as a keyboard (Keyboard), and optionally, the user interface may also include a standard wired interface and a wireless interface. The network interface may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity interface). The memory may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory. The memory may optionally be a storage device independent of the aforementioned processor.

[0063] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the computer device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0064] like Figure 1 As shown, a memory as a storage medium may include an operating system, a data storage module, a network communication module, a user interface module, and an electronic program module.

[0065] exist Figure 1In the computer device shown, the network interface is mainly used for data communication with the network server; the user interface is mainly used for data interaction with the user; the processor and memory in the computer device of the present application can be set in the computer device, and the computer device calls the computer program product stored in the electronic program module through the processor, and executes the method for module defect detection of the display panel provided in the embodiment of the present application.

[0066] Based on the hardware environment of the aforementioned embodiment, the embodiment of the present application provides a module defect detection method for a display panel, please refer to Figure 2 , Figure 2 The flowchart of the module defect detection method is as follows:

[0067] (1) Acquire a panel image of a display panel, and construct a training sample set and a detection sample set based on the panel image; wherein the panel images included in the training sample set are panel images with defects, and the panel images included in the detection sample set are panel images to be detected.

[0068] (2) Perform feature region extraction, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the test sample set to obtain enhanced training images and enhanced test images.

[0069] In an embodiment of the present invention, the process of performing feature region extraction, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the detection sample set is as follows:

[0070] (2.1) Perform ROI capture on the grayscale panel images in the training sample set and the test sample set to obtain the corresponding ROI image; wherein, ROI capture is used to extract a specific area from the panel image for subsequent processing, and a rectangle, circle or other shape can be drawn on the panel image to specify the region of interest.

[0071] (2.2) Performing Gaussian filtering on the ROI images corresponding to the grayscale panel images in the training sample set and the detection sample set, and switching the Gaussian filtered ROI images to a high magnification to obtain the corresponding denoised images; wherein Gaussian filtering can eliminate Gaussian noise in the panel images.

[0072] (2.3) Perform gamma transformation and histogram equalization on the denoised images corresponding to the grayscale panel images in the training sample set and the test sample set to obtain the final enhanced training image and enhanced test image. The gamma transformation and histogram equalization can improve the contrast and visual effect of the panel image. Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 3 is the original pixel distribution map of the panel image, Figure 4 is the pixel distribution diagram of the panel image after Gaussian filtering, Figure 5 is the pixel distribution diagram of the panel image after gamma transformation, Figure 6 This is a pixel distribution diagram after histogram equalization of the panel image. By performing Gaussian filtering, gamma transformation, and histogram equalization on the grayscale panel image, the overall grayscale value of the image is more dispersed and the defect features are more obvious.

[0073] (3) Perform singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set to obtain a foreground training image and a foreground detection image.

[0074] In an embodiment of the present invention, the process of performing singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set is as follows:

[0075] (3.1) Perform singular value decomposition and regularization on the color panel images in the training sample set and the test sample set to obtain the corresponding enhanced images. Singular value decomposition can achieve panel image compression and denoising; regularization can improve data consistency and reliability, reduce data redundancy, and improve data storage and processing efficiency, which is convenient for subsequent data analysis, training and application of machine learning models.

[0076] (3.2) Perform binary threshold segmentation and opening operation on the enhanced images corresponding to the color panel images in the training sample set and the detection sample set to obtain the corresponding binary images.

[0077] (3.3) Perform edge detection on the binary images corresponding to the color panel images in the training sample set and the detection sample set to obtain the final foreground training image and foreground detection image. Among them, edge detection processing can greatly reduce the amount of data, eliminate irrelevant information, retain important structural attributes of the image, and extract the outline and key features of the panel image by identifying the edge information in the panel image, providing a basis for subsequent defect detection.

[0078] (4) Feature annotating the enhanced training image and the foreground training image, and inputting the feature-annotated enhanced training image and the foreground training image into a two-stage target detection model for model training to obtain a defect detection model. The two-stage target detection model includes Faster RCNN and Mask RCNN.

[0079] In an embodiment of the present invention, the two-stage target detection model divides the target detection task into two stages for processing. The first stage is to generate candidate areas, that is, areas that may contain targets; the second stage is to classify and regress bounding boxes of these candidate areas to determine the target category and precise location.

[0080] Specifically, the process of inputting the feature-annotated enhanced training image and the foreground training image into the two-stage target detection model for model training is as follows:

[0081] (4.1) The feature-annotated enhanced training image and foreground training image are respectively input into the two-stage target detection model, and the candidate regions are generated from the enhanced training image and the foreground training image through the two-stage target detection model.

[0082] (4.2) Classify the candidate region to determine whether the candidate region contains the target object and the type of the target object, and use the bounding box regression technology to adjust the position and size of the candidate region to locate the target object.

[0083] (4.3) Calculate the classification loss and regression loss of the two-stage target detection model to obtain the classification loss value and the regression loss value, and update the model weights according to the classification loss value and the regression loss value to optimize the model performance.

[0084] (5) The defect detection model is used to perform defect detection on the enhanced detection image and the foreground detection image to obtain several defect detection results.

[0085] In the embodiment of the present invention, the process of using the defect detection model to perform defect detection on the enhanced detection image and the foreground detection image is as follows:

[0086] (5.1) The enhanced detection image is input into the defect detection model, and the defect features are located and classified by the defect detection model to obtain several grayscale defect target frames.

[0087] (5.2) The foreground detection image is input into the defect detection model, and the defect features are located and classified by the defect detection model to obtain several color defect target boxes.

[0088] (6) Verify several defect detection results in a manner combining defect priority and confidence to obtain the final defect detection result.

[0089] In an embodiment of the present invention, the process of verifying several defect detection results in a manner combining defect priority and confidence is as follows:

[0090] (6.1) Obtain the defect type, location, size and confidence corresponding to the grayscale defect target box and the color defect target box.

[0091] (6.2) Prioritize the defects detected in the grayscale defect target frame and the color defect target frame according to the defect severity to obtain the defect priority classification result. Among them, the defect severity is comprehensively determined according to the defect type, location and size; that is, the defect type, location and size can be assigned weights respectively, and then the comprehensive assignment is calculated, and the defects are prioritized according to the comprehensive assignment.

[0092] (6.3) The grayscale defect target frame and the color defect target frame are screened according to the defect priority classification results, and the final defect detection result is output in combination with the confidence level.

[0093] In summary, the present invention first scales the grayscale image of the display panel (Panel) module, uses Gaussian filtering to denoise the cropped image, uses gamma transformation and histogram equalization to enhance the cropped image, and can significantly visualize the slight defect features while reducing the image size; at the same time, singular value decomposition and normalization are performed on the color image of the display panel (Panel) module, and the image is denoised again using binary threshold segmentation and opening operation, and the foreground and background images are segmented using edge detection, which can filter invalid pixel areas. Then, the processed image is trained using a two-stage target detection algorithm to obtain a defect detection model. Finally, the defect detection model is used to perform defect detection on the above-mentioned processed image, summarize the detection results of each image frame, and use a combination of defect priority and confidence to calculate and output the final unique defect detection result. The present invention uses a combination of grayscale images and color images to significantly improve the model defect detection accuracy, assist factories in improving their shipment yield, and bring obvious economic benefits.

[0094] In addition, in one embodiment, based on the same inventive concept as the above embodiment, the embodiment of the present invention provides a module defect detection system for a display panel, and the system corresponds to the method of embodiment 1 one by one, please refer to Figure 7 , Figure 7 The structure diagram of the module defect detection system is as follows:

[0095] A panel image acquisition unit, the panel image acquisition unit is used to acquire a panel image of the display panel, and construct a training sample set and a detection sample set based on the panel image;

[0096] A grayscale image processing unit, the grayscale image processing unit is used to perform feature region interception, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the detection sample set to obtain enhanced training images and enhanced detection images;

[0097] A color image processing unit, the color image processing unit is used to perform singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set to obtain a foreground training image and a foreground detection image;

[0098] A detection model training unit, wherein the detection model training unit is used to perform feature annotation on the enhanced training image and the foreground training image, and input the feature-annotated enhanced training image and the foreground training image into a two-stage target detection model for model training to obtain a defect detection model;

[0099] A panel defect detection unit, wherein the panel defect detection unit uses a defect detection model to perform defect detection on the enhanced detection image and the foreground detection image to obtain a number of defect detection results;

[0100] A defect comprehensive verification unit verifies a plurality of defect detection results in a manner combining defect priority and confidence to obtain a final defect detection result.

[0101] It should be noted that each unit in the module defect detection system of the display panel in this embodiment corresponds one-to-one to each step in the module defect detection method of the display panel in the aforementioned embodiment. Therefore, the specific implementation method and technical effects achieved by this embodiment can refer to the implementation method of the module defect detection method of the display panel in the aforementioned embodiment, and will not be repeated here.

[0102] In addition, in one embodiment, the present application also provides a computer device, which includes a processor, a memory, and a computer program stored in the memory, and the computer program implements the method in the aforementioned embodiment when executed by the processor.

[0103] In addition, in one embodiment, the present application further provides a computer storage medium, on which a computer program is stored, and the computer program implements the method in the aforementioned embodiment when executed by a processor.

[0104] In some embodiments, the computer readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or various devices including one or any combination of the above memories. The computer may be various computing devices including intelligent terminals and servers.

[0105] In some embodiments, executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0106] As an example, executable instructions may, but need not, correspond to a file in a file system, may be stored as part of a file that stores other programs or data, such as in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions).

[0107] By way of example, executable instructions may be deployed to be executed on one computing device, or on multiple computing devices located at one site, or on multiple computing devices distributed across multiple sites and interconnected by a communication network.

[0108] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0109] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0110] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a multimedia terminal device (which can be a mobile phone, a computer, a television receiver, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0111] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for detecting module defects of a display panel, characterized in that: The method comprises the following steps: Acquire a panel image of a display panel, and construct a training sample set and a detection sample set based on the plurality of panel images; Performing feature region interception, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the detection sample set to obtain enhanced training images and enhanced detection images; Performing singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set to obtain a foreground training image and a foreground detection image; Performing feature annotation on the enhanced training image and the foreground training image, and inputting the feature-annotated enhanced training image and the foreground training image into the two-stage target detection model for model training to obtain a defect detection model; Using the defect detection model to perform defect detection on the enhanced detection image and the foreground detection image to obtain several defect detection results; Several defect detection results are verified in a manner combining defect priority and confidence to obtain the final defect detection result.

2. The method for detecting module defects of a display panel according to claim 1, characterized in that: The process of feature region extraction, Gaussian filtering and image enhancement for the grayscale panel images in the training sample set and the test sample set is as follows: Perform ROI interception on the grayscale panel images in the training sample set and the detection sample set to obtain the corresponding ROI image; Performing Gaussian filtering on the ROI images corresponding to the grayscale panel images in the training sample set and the detection sample set, and switching the Gaussian filtered ROI images to a high magnification to obtain the corresponding denoised images; Gamma transformation and histogram equalization are performed on the denoised images corresponding to the grayscale panel images in the training sample set and the detection sample set to obtain the final enhanced training image and enhanced detection image.

3. The method for detecting module defects of a display panel according to claim 1, characterized in that: The process of performing singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the test sample set is as follows: Perform singular value decomposition and regularization processing on the color panel images in the training sample set and the detection sample set to obtain the corresponding enhanced images; Performing binary threshold segmentation and opening operation on the enhanced images corresponding to the color panel images in the training sample set and the detection sample set to obtain the corresponding binary images; Edge detection is performed on the binary images corresponding to the color panel images in the training sample set and the detection sample set to obtain the final foreground training image and foreground detection image.

4. The method for detecting module defects of a display panel according to claim 1, characterized in that: The two-stage target detection model includes Faster RCNN and Mask RCNN.

5. The method for detecting module defects of a display panel according to claim 1, characterized in that: The process of inputting the feature-annotated enhanced training images and foreground training images into the two-stage object detection model for model training is as follows: The feature-annotated enhanced training image and the foreground training image are respectively input into the two-stage object detection model, and the candidate regions are generated from the enhanced training image and the foreground training image by the two-stage object detection model; Classify the candidate area to determine whether the candidate area contains the target object and the type of the target object, and use the bounding box regression technology to adjust the position and size of the candidate area to locate the target object; The classification loss and regression loss of the two-stage target detection model are calculated to obtain the classification loss value and the regression loss value, and the model weights are updated according to the classification loss value and the regression loss value to optimize the model performance.

6. The method for detecting module defects of a display panel according to claim 1, characterized in that: The process of using the defect detection model to perform defect detection on the enhanced detection image and the foreground detection image is as follows: The enhanced inspection image is input into the defect detection model, and the defect features are located and classified by the defect detection model to obtain a number of grayscale defect target frames; The foreground detection image is input into the defect detection model, and the defect features are located and classified by the defect detection model to obtain several color defect target frames.

7. The method for detecting module defects of a display panel according to claim 6, characterized in that: The process of verifying several defect detection results in a combination of defect priority and confidence is as follows: Obtain the defect type, location, size, and confidence level corresponding to the grayscale defect target frame and the color defect target frame; Defects detected in the grayscale defect target frame and the color defect target frame are prioritized according to defect severity to obtain a defect priority classification result; wherein the defect severity is comprehensively determined according to the defect type, location and size; The grayscale defect target frame and the color defect target frame are screened according to the defect priority classification results, and the final defect detection result is output in combination with the confidence level.

8. A module defect detection system for a display panel, characterized in that: The system comprises: A panel image acquisition unit, the panel image acquisition unit is used to acquire a panel image of the display panel, and construct a training sample set and a detection sample set based on the panel image; A grayscale image processing unit, the grayscale image processing unit is used to perform feature region interception, Gaussian filtering and image enhancement on the grayscale panel images in the training sample set and the detection sample set to obtain enhanced training images and enhanced detection images; A color image processing unit, the color image processing unit is used to perform singular value decomposition, threshold segmentation, opening operation and edge detection on the color panel images in the training sample set and the detection sample set to obtain a foreground training image and a foreground detection image; A detection model training unit, wherein the detection model training unit is used to perform feature annotation on the enhanced training image and the foreground training image, and input the feature-annotated enhanced training image and the foreground training image into a two-stage target detection model for model training to obtain a defect detection model; A panel defect detection unit, wherein the panel defect detection unit uses a defect detection model to perform defect detection on the enhanced detection image and the foreground detection image to obtain a number of defect detection results; A defect comprehensive verification unit verifies a plurality of defect detection results in a manner combining defect priority and confidence to obtain a final defect detection result.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for detecting module defects of a display panel described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for detecting module defects of a display panel as described in any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for detecting module defects of a display panel described in any one of claims 1 to 7 is implemented.

Images