OLED irregular circuit coating and process differentiation true and false defect detection method

Abstract

The present application relates to panel true and false defect detection technical field, a kind of OLED irregular circuit plating and process differentiation true and false defect detection method, comprising: confirming OLED panel set and industrial camera, OLED panel is extracted from OLED panel set in turn, and the following operations are executed to the extracted OLED panel: receiving true and false defect detection instruction, obtaining brightness normalization image based on true and false defect detection instruction, industrial camera and OLED panel, the brightness normalization image is denoised, and denoised panel image is obtained, defect classification result set is obtained based on denoised panel image, defect classification result set is summarized, and multiple defect classification result set is obtained, and true and false defect detection of OLED irregular circuit plating and process differentiation is completed based on defect classification result set.The present application can improve the quality and reliability of OLED product.

Description

Technical Field

[0001] This invention relates to the field of panel defect detection technology, and in particular to a method for detecting genuine and fake defects in OLED irregular circuit coating and process differences. Background Technology

[0002] OLEDs, as light-emitting devices based on organic materials, possess advantages such as self-illumination, high contrast, fast response, and thinness. Irregular circuit coating processes refer to circuit structures that may exhibit irregular shapes. Genuine defects originate from process abnormalities, material flaws, or operational errors, directly affecting product performance and quality; false defects, on the other hand, are mostly caused by factors such as testing equipment errors, methodological limitations, or environmental interference, and are considered misjudgments.

[0003] Currently, OLED defect detection mainly relies on traditional methods such as optics and electronics. However, optical detection is susceptible to interference from light reflection and refraction, leading to false alarms; electrical detection, due to its complex irregular circuit structure, struggles to accurately locate and classify defects. Furthermore, existing methods have limited ability to distinguish between genuine and false defects. Therefore, improving the quality and reliability of OLED products is an urgent technical problem that needs to be solved. Summary of the Invention

[0004] This invention provides a method for detecting genuine and fake defects in OLED irregular circuit coating and process differences, and a computer-readable storage medium, the main purpose of which is to improve the quality and reliability of OLED products.

[0005] To achieve the above objectives, the present invention provides a method for detecting genuine and fake defects in OLED irregular circuit coating and process differences, comprising:

[0006] The OLED panel set and industrial camera have been identified. The OLED panel set includes multiple OLED panels.

[0007] OLED panels are extracted sequentially from the OLED panel collection, and the following operations are performed on each extracted OLED panel:

[0008] Receive true and false defect detection commands, acquire brightness normalized images based on true and false defect detection commands, industrial cameras and OLED panels, denoise the brightness normalized images, and obtain denoised panel images.

[0009] Defect classification result sets are obtained based on denoised panel images, and the defect classification result sets are summarized to obtain multiple defect classification result sets. Among them, the OLED panel corresponds one-to-one with the defect classification result set.

[0010] Based on the defect classification result set, we completed the detection of true and false defects in OLED irregular circuit coating and process differences.

[0011] Optionally, the step of acquiring a brightness-normalized image based on true / false defect detection commands, an industrial camera, and an OLED panel includes:

[0012] Based on the instructions for detecting true and false defects, an industrial camera is used to photograph the OLED panel to obtain an initial sequence of OLED panel images.

[0013] Based on the initial OLED panel image sequence, an aligned OLED panel image sequence is obtained. Then, a multi-frame averaging fusion operation is performed on the aligned OLED panel image sequence to obtain the average panel image.

[0014] Global brightness normalization is performed on the average panel image to obtain a brightness-normalized image.

[0015] Optionally, obtaining the aligned OLED panel image sequence based on the initial OLED panel image sequence includes:

[0016] A reference panel image is identified based on the initial OLED panel image sequence, and the reference panel image is removed from the initial OLED panel image sequence to obtain a sequence of panel images to be registered. The reference panel image is the first initial OLED panel image in the initial OLED panel image sequence, and the sequence of panel images to be registered includes multiple panel images to be registered.

[0017] Extract the panel images to be registered sequentially from the sequence of panel images to be registered, and perform the following operations on each of the panel images to be registered:

[0018] The position of the maximum peak is obtained based on the reference panel image and the panel image to be registered, and the position of the maximum peak is used as the relative global translation amount.

[0019] A translation transformation matrix is ​​constructed based on the relative global translation amount. An image realignment operation is then performed on the panel image to be registered based on the translation transformation matrix to obtain the aligned OLED panel image.

[0020] By combining the aligned OLED panel images and the reference panel image, an aligned OLED panel image sequence is obtained. The aligned OLED panel image sequence includes multiple aligned OLED panel images, and each aligned OLED panel image includes multiple grayscale pixels.

[0021] Optionally, obtaining the maximum peak position based on the reference panel image and the panel image to be registered includes:

[0022] Perform windowing operations on the reference panel image and the panel image to be registered, respectively, to obtain a windowed reference panel image and a windowed panel image to be registered;

[0023] Calculate the reference complex matrix and the complex matrix to be registered for the windowed reference panel image and the windowed panel image to be registered, respectively;

[0024] The cross-power spectrum is calculated based on the reference complex matrix and the complex matrix to be registered. An inverse Fourier transform is performed on the cross-power spectrum to obtain the impulse response diagram. The position of the maximum peak is extracted from the impulse response diagram.

[0025] Optionally, the step of performing a multi-frame averaging fusion operation on the aligned OLED panel image sequence to obtain an average panel image includes:

[0026] Sequentially extract aligned OLED panel images from the aligned OLED panel image sequence, and create a blank two-dimensional matrix based on the aligned OLED panel images;

[0027] According to the preset pixel extraction order, grayscale pixels are extracted from the aligned OLED panel image in sequence, and the grayscale pixel coordinates and grayscale pixel values ​​are determined based on the grayscale pixels.

[0028] The target matching position is determined from the blank two-dimensional matrix based on the grayscale pixel coordinates. The target matching value is determined based on the target matching position. The grayscale pixel value is added to the target matching value to obtain the updated matching value. The updated matching value is then used to replace the target matching value to obtain the accumulated pixel value.

[0029] Return to the step of extracting grayscale pixels from the aligned OLED panel image sequentially according to the preset pixel extraction order, until all grayscale pixels in the aligned OLED panel image have been extracted;

[0030] Summarize the accumulated pixel values ​​to obtain an accumulated two-dimensional matrix. Use the accumulated two-dimensional matrix as a blank two-dimensional matrix and return to the step of extracting aligned OLED panel images from the aligned OLED panel image sequence sequentially until all aligned OLED panel images in the aligned OLED panel image sequence have been extracted.

[0031] The summation of the two-dimensional matrices yields a comprehensive summation two-dimensional matrix, which includes multiple comprehensive summation pixel values.

[0032] The number of OLED panel images is obtained from the aligned OLED panel image sequence, and the average number of panel images is obtained from the comprehensive summation of the two-dimensional matrix and the number of OLED panel images.

[0033] Optionally, the step of performing global brightness normalization on the average panel image to obtain a brightness-normalized image includes:

[0034] Obtain the short side size of the average panel image, determine the Gaussian kernel size based on the short side size, determine the target Gaussian kernel based on the Gaussian kernel size, and perform a Gaussian blur operation on the average panel image using the target Gaussian kernel to obtain the background brightness image.

[0035] Extract the aggregated pixel values ​​sequentially from the average panel image, and perform the following operations on each aggregated pixel value:

[0036] The accumulated pixel values ​​are converted to floating-point numbers to obtain floating-point pixel values. The background brightness pixel values ​​are then determined from the background brightness image based on the floating-point pixel values.

[0037] The normalized pixel value is calculated based on the floating-point pixel value, the background brightness pixel value, and the preset target brightness constant. The normalized pixel value is then converted to an integer type to obtain an integer normalized pixel value.

[0038] By summing the integer normalized pixel values, a brightness-normalized image is obtained.

[0039] Optionally, the step of obtaining the defect classification result set based on the denoised panel image includes:

[0040] Once the grayscale image of the panel template is identified, it is quickly matched with the denoised panel image to obtain the aligned panel template image and the aligned denoised panel image.

[0041] A panel difference map is obtained based on the aligned panel template image and the aligned denoised panel image. The set of circuit functional regions is identified based on the grayscale image of the panel template. The set of circuit functional regions includes multiple circuit functional regions.

[0042] The panel difference map is divided according to the circuit functional area set to obtain the circuit sub-region set to be detected. The circuit sub-region set to be detected includes: multiple circuit sub-regions to be detected, and each circuit sub-region to be detected includes: multiple difference map circuit pixels, and the difference map circuit pixels correspond to a difference map circuit pixel value.

[0043] For each sub-region of the circuit to be tested within the set of sub-regions, the following operations are performed:

[0044] A potential defect map is obtained based on the sub-region of the circuit to be inspected, and an AI review region set is generated based on the potential defect map.

[0045] A pre-built deep learning classification model is used to perform precise judgment on each AI review region in the AI ​​review region set to obtain the defect classification result, in which the defect classification result is either a real defect or normal.

[0046] By summarizing the defect classification results, a defect classification result set corresponding to the sub-region set of the circuit to be tested is obtained.

[0047] Optionally, obtaining the potential defect map based on the sub-region of the circuit to be detected includes:

[0048] Extract the difference map circuit pixels sequentially from the sub-region of the circuit to be detected, and perform the following operations on each extracted difference map circuit pixel:

[0049] Set the process error range, determine the target difference map circuit pixel value based on the difference map circuit pixel value, and determine whether the target difference map circuit pixel value is within the process error range.

[0050] If the target difference map circuit pixel value is within the process error range, then the difference map circuit pixel is marked normally using the preset normal pixel value to obtain normal circuit pixels;

[0051] If the target difference map circuit pixel value is not within the process error range, then the upper limit and lower limit of the process error range are obtained. If the target difference map circuit pixel value is less than the lower limit of the range, then the difference map circuit pixel is marked normally using the normal pixel value to obtain the normal circuit pixel. If the target difference map circuit pixel value is greater than the upper limit of the range, then the difference map circuit pixel is marked abnormally using the preset abnormal pixel value to obtain the abnormal circuit pixel.

[0052] By summarizing the pixels of normal and abnormal circuits, a potential defect map is obtained.

[0053] Optionally, setting the process error range includes:

[0054] Obtain a standard sub-region image set based on the sub-region image, and calculate the standard gray average value and standard gray standard deviation of each standard sub-region image in the standard sub-region image set to obtain the standard gray average value set and the standard gray standard deviation set.

[0055] The mean gray value and standard deviation of the region are calculated based on the standard gray value mean set and the standard gray value standard deviation set.

[0056] The process error range is constructed based on the regional gray mean, regional gray standard deviation, and preset quality control coefficient.

[0057] To achieve the above objectives, the present invention also provides a system for detecting genuine and fake defects in OLED irregular circuit coating and process differences, comprising:

[0058] The system initialization module is used to identify the OLED panel set and the industrial camera. The OLED panel set includes multiple OLED panels.

[0059] The image acquisition and processing module is used to sequentially extract OLED panels from the OLED panel collection and perform the following operations on each extracted OLED panel: receive true and false defect detection instructions, acquire brightness normalized images based on the true and false defect detection instructions, industrial camera and OLED panel, and denoise the brightness normalized images to obtain denoised panel images.

[0060] The panel defect classification module is used to obtain a defect classification result set based on the denoised panel image, summarize the defect classification result set, and obtain multiple defect classification result sets, in which the OLED panel corresponds one-to-one with the defect classification result set;

[0061] The defect detection completion module is used to complete the detection of true and false defects in OLED irregular circuit coating and process differences based on the defect classification result set.

[0062] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0063] Memory, storing at least one instruction;

[0064] The processor executes the instructions stored in the memory to implement the above-described method for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation.

[0065] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned method for detecting genuine and false defects in OLED irregular circuit coating and process differentiation.

[0066] To address the problems described in the background art, this invention identifies an OLED panel set and an industrial camera. The OLED panel set includes multiple OLED panels. OLED panels are extracted sequentially from the OLED panel set, and the following operations are performed on each extracted OLED panel: This invention sequentially extracts OLED panels for inspection, enabling batch processing of multiple OLED panels and improving inspection efficiency. Compared to inspecting individual panels one by one, this method can complete the inspection of a large number of panels in a shorter time, meeting the needs of large-scale production. Simultaneously, sequential extraction and inspection ensure the orderliness and standardization of the inspection process, facilitating the management and traceability of inspection results. It receives true and false defect detection commands, acquires brightness-normalized images based on these commands, the industrial camera, and the OLED panels, and denoises the brightness-normalized images to obtain denoised panel images. The brightness-normalized images of this invention can eliminate brightness differences between different OLED panels and between different areas of the same panel. In actual production, due to manufacturing processes and environmental factors, the brightness of OLED panels may fluctuate. This brightness difference can interfere with defect detection, leading to misjudgments or missed detections. By acquiring brightness-normalized images, the panel brightness can be adjusted to a relatively uniform level, highlighting defect features and improving the accuracy of defect detection. Simultaneously, images may be subject to various noise interferences during acquisition and transmission. Noise can blur images, obscuring the true characteristics of defects and affecting defect identification and classification. Denoising processing can effectively remove this noise, making the image clearer, enhancing defect identifiability, and providing high-quality image data for subsequent defect classification and detection. This invention acquires defect classification result sets from noisy panel images, summarizes these result sets, and obtains multiple defect classification result sets. Each OLED panel corresponds one-to-one with a defect classification result set. Different types of defects may be caused by different reasons and have varying impacts on product quality. Accurate classification results facilitate in-depth analysis and processing of defects, enabling targeted improvement measures. By summarizing the defect classification results for each OLED panel to form a result set corresponding to each panel, the defect status of each panel can be easily recorded and managed, facilitating subsequent quality traceability and analysis. Based on the defect classification result sets, the invention enables the detection of true and false defects in OLED irregular circuit coating and process differences. Therefore, this invention can improve the quality and reliability of OLED products. Attached Figure Description

[0067] Figure 1 This is a flowchart illustrating a method for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation, provided in an embodiment of the present invention.

[0068] Figure 2A functional block diagram of a system for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation provided in an embodiment of the present invention;

[0069] Figure 3 This is a schematic diagram of an electronic device that implements the method for detecting real and fake defects in OLED irregular circuit coating and process differentiation, according to an embodiment of the present invention.

[0070] Explanation of reference numerals in the attached figures:

[0071] 1. Electronic equipment; 10. Processor; 11. Memory; 12. Bus; 100. OLED irregular circuit coating and process differentiation defect detection system; 101. System initialization module; 102. Image acquisition and processing module; 103. Panel defect classification module; 104. Defect detection completion module.

[0072] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0073] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0074] This application provides a method for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation. The execution subject of this method includes, but is not limited to, at least one electronic device that can be configured to execute the method provided in this application, such as a server or a terminal. In other words, the method for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0075] Reference Figure 1 The diagram shown is a flowchart illustrating a method for detecting genuine and fake defects in OLED irregular circuit coating and process differences, according to an embodiment of the present invention. In this embodiment, the method for detecting genuine and fake defects in OLED irregular circuit coating and process differences includes:

[0076] S1. Identify the OLED panel set and industrial camera. The OLED panel set includes multiple OLED panels.

[0077] It should be explained that OLED is an abbreviation for Organic Light Emitting Diode. An OLED panel refers to a display panel based on organic light-emitting diode technology, composed of multiple layers of organic materials and electrodes, capable of directly emitting light, thus possessing self-emissive properties. An industrial camera refers to a highly specialized camera; this invention is primarily used for image acquisition and analysis of OLED panels during industrial production processes. It can quickly and accurately capture images of OLED panels and transmit image data to computers or other image processing equipment for subsequent processing. This invention's industrial camera is equipped with a monochrome (grayscale) image sensor, aiming to achieve higher spatial resolution and signal-to-noise ratio to meet the stringent requirements of the inspection system for imaging accuracy of microscopic circuit structures. Simultaneously, it also improves the throughput of the entire inspection system, meeting mass production needs.

[0078] S2. Sequentially extract OLED panels from the OLED panel collection, and perform the following operations on each extracted OLED panel: receive true and false defect detection instructions, and acquire brightness normalized images based on the true and false defect detection instructions, industrial camera and OLED panel.

[0079] Specifically, the acquisition of a brightness-normalized image based on true / false defect detection commands, an industrial camera, and an OLED panel includes:

[0080] Based on the instructions for detecting true and false defects, an industrial camera is used to photograph the OLED panel to obtain an initial sequence of OLED panel images.

[0081] Based on the initial OLED panel image sequence, an aligned OLED panel image sequence is obtained. Then, a multi-frame averaging fusion operation is performed on the aligned OLED panel image sequence to obtain the average panel image.

[0082] Global brightness normalization is performed on the average panel image to obtain a brightness-normalized image.

[0083] It should be explained that the "true / false defect detection command" refers to a manually initiated command used to start defect detection on the OLED panel. Upon receiving the command, the system triggers an industrial camera to capture an image of the OLED panel and initiates a series of subsequent image processing steps. The initial OLED panel image refers to the unprocessed image obtained by the industrial camera after receiving the command. The initial OLED panel image sequence is a collection of all initial OLED panel images. The reason for obtaining the initial OLED panel image sequence in this invention is that any industrial camera introduces noise during shooting. This noise manifests as random, isolated bright and dark pixels in the image, which is particularly noticeable under low light or short exposure times. If only one initial OLED panel image is captured, these noise points will be mistakenly identified as real defects by subsequent algorithms, leading to misjudgment. The brightness normalized image refers to the image obtained by performing global brightness normalization processing on the average panel image. The purpose of brightness normalization is to eliminate brightness variations caused by factors such as uneven lighting and differences in camera response, making the brightness distribution of the image more uniform and helping to improve the accuracy and reliability of defect detection.

[0084] Specifically, obtaining the aligned OLED panel image sequence based on the initial OLED panel image sequence includes:

[0085] A reference panel image is identified based on the initial OLED panel image sequence, and the reference panel image is removed from the initial OLED panel image sequence to obtain a sequence of panel images to be registered. The reference panel image is the first initial OLED panel image in the initial OLED panel image sequence, and the sequence of panel images to be registered includes multiple panel images to be registered.

[0086] Extract the panel images to be registered sequentially from the sequence of panel images to be registered, and perform the following operations on each of the panel images to be registered:

[0087] The position of the maximum peak is obtained based on the reference panel image and the panel image to be registered, and the position of the maximum peak is used as the relative global translation amount.

[0088] A translation transformation matrix is ​​constructed based on the relative global translation amount. An image realignment operation is then performed on the panel image to be registered based on the translation transformation matrix to obtain the aligned OLED panel image.

[0089] By combining the aligned OLED panel images and the reference panel image, an aligned OLED panel image sequence is obtained. The aligned OLED panel image sequence includes multiple aligned OLED panel images, and each aligned OLED panel image includes multiple grayscale pixels.

[0090] It should be explained that the reference panel image refers to the first grayscale image selected from the initial OLED panel image sequence. It serves as a reference image for comparison and alignment with other images. The sequence of panel images to be registered is the image sequence obtained by removing the reference panel image from the initial OLED panel image sequence. Each image in the sequence is a panel image to be registered. The relative global translation is the position of the maximum peak extracted from the impulse response map, representing the translation amount of the OLED panel image to be registered relative to the reference panel image in the horizontal and vertical directions. The translation transformation matrix is ​​a 2x3 matrix used to represent the image translation operation. In the step of constructing the translation transformation matrix based on the relative global translation, the translation transformation matrix is ​​expressed as: ,in, The x-coordinate represents the location of the maximum peak value. The vertical coordinate represents the position of the maximum peak. Image realignment is an operation that translates and aligns the panel image to be registered with the reference panel image according to the translation transformation matrix. The image realignment operation of this invention can be achieved by calling the `warpAffine` function in standard computer vision libraries such as OpenCV, and importing the translation transformation matrix and the image to be registered, using bilinear interpolation. This step is existing technology and will not be described further here. The aligned OLED panel image is the image obtained after the image realignment operation. The aligned OLED panel image sequence refers to the image sequence obtained by summarizing all aligned OLED panel images and the reference panel image, and this sequence has the same arrangement position as the initial OLED panel image in the initial OLED panel image sequence. The reason why this invention uses the maximum peak position as the relative global translation amount is that, through frequency domain processing in the phase correlation method, the content of the reference panel image and the panel image to be registered is filtered out, and the global translation relationship between the two images is finally represented in the impulse response diagram as a sharp pulse (maximum peak value). The coordinates of this peak in the impulse response plot directly correspond to the pixel offsets of the image to be registered relative to the reference image in the horizontal and vertical directions. Therefore, without complex calculations, the position of the maximum peak can be directly located and read to obtain a high-precision relative global translation. The phase correlation method is an efficient registration algorithm that locates the maximum peak point in the impulse response map of two images by calculating the inverse transform of the phase difference in the frequency domain. The coordinates of this point are the precise translation between the images.

[0091] Specifically, obtaining the maximum peak position based on the reference panel image and the panel image to be registered includes:

[0092] Perform windowing operations on the reference panel image and the panel image to be registered, respectively, to obtain a windowed reference panel image and a windowed panel image to be registered;

[0093] Calculate the reference complex matrix and the complex matrix to be registered for the windowed reference panel image and the windowed panel image to be registered, respectively;

[0094] The cross-power spectrum is calculated based on the reference complex matrix and the complex matrix to be registered. An inverse Fourier transform is performed on the cross-power spectrum to obtain the impulse response diagram. The position of the maximum peak is extracted from the impulse response diagram.

[0095] It should be explained that the windowing operation performed on the reference panel image and the panel image to be registered, respectively, refers to applying a window function (such as a Hanning window or a Hamming window) to both the reference panel image and the panel image to be registered. The window function can mitigate the negative impact of boundary discontinuities in the reference panel image and the panel image to be registered on the FFT, resulting in a higher quality cross-power spectrum and sharper, more easily detectable peaks in the impulse response plot. The formulas for calculating the reference complex matrix and the complex matrix to be registered in the step of calculating the reference complex matrix and the complex matrix to be registered for the windowed reference panel image and the windowed panel image to be registered, respectively, are as follows:

[0096] ,

[0097] ,

[0098] in, Represents the base complex matrix, This indicates that a Fourier transform is performed on the windowed reference panel image. This represents the windowed reference panel image. Represents the complex matrix to be registered. This indicates that a Fourier transform is performed on the windowed panel image to be registered. This represents the windowed panel image to be registered. The step of calculating the cross power spectrum based on the reference complex matrix and the complex matrix to be registered is as follows: obtain the complex conjugate to be registered based on the complex matrix to be registered, and calculate the cross power spectrum based on the complex conjugate to be registered. The calculation formula is as follows:

[0099] ,

[0100] in, Represents the cross power spectrum. Indicates the complex conjugate to be registered. Indicates the modulo symbol.

[0101] Furthermore, the impulse response graph refers to the image obtained from the cross-power spectrum through inverse Fourier transform. The maximum peak position refers to the coordinates of the point with the largest value in the impulse response graph. These coordinates are obtained by establishing an image coordinate system with the lower left corner of the impulse response graph as the origin.

[0102] Specifically, the step of performing a multi-frame averaging and fusion operation on the aligned OLED panel image sequence to obtain an average panel image includes:

[0103] Sequentially extract aligned OLED panel images from the aligned OLED panel image sequence, and create a blank two-dimensional matrix based on the aligned OLED panel images;

[0104] According to the preset pixel extraction order, grayscale pixels are extracted from the aligned OLED panel image in sequence, and the grayscale pixel coordinates and grayscale pixel values ​​are determined based on the grayscale pixels.

[0105] The target matching position is determined from the blank two-dimensional matrix based on the grayscale pixel coordinates. The target matching value is determined based on the target matching position. The grayscale pixel value is added to the target matching value to obtain the updated matching value. The updated matching value is then used to replace the target matching value to obtain the accumulated pixel value.

[0106] Return to the step of extracting grayscale pixels from the aligned OLED panel image sequentially according to the preset pixel extraction order, until all grayscale pixels in the aligned OLED panel image have been extracted;

[0107] Summarize the accumulated pixel values ​​to obtain an accumulated two-dimensional matrix. Use the accumulated two-dimensional matrix as a blank two-dimensional matrix and return to the step of extracting aligned OLED panel images from the aligned OLED panel image sequence sequentially until all aligned OLED panel images in the aligned OLED panel image sequence have been extracted.

[0108] The summation of the two-dimensional matrices yields a comprehensive summation two-dimensional matrix, which includes multiple comprehensive summation pixel values.

[0109] The number of OLED panel images is obtained from the aligned OLED panel image sequence, and the average number of panel images is obtained from the comprehensive summation of the two-dimensional matrix and the number of OLED panel images.

[0110] It should be explained that the blank 2D matrix is ​​a matrix of the same size as the aligned OLED panel image, initially with all elements set to 0, used to store accumulated pixel values. The pixel extraction order refers to a pre-defined sequence used to extract grayscale pixels one by one from the aligned OLED panel image, ensuring an orderly extraction process. A grayscale pixel is a single pixel in an image, its value representing the grayscale intensity of that point. For grayscale images, each pixel has only one grayscale value, typically ranging from 0 to 255. Grayscale pixel coordinates refer to the position of the grayscale pixel in the image, facilitating the extraction of the corresponding grayscale pixel value from other aligned OLED panel images. The target matching position is the position in the blank 2D matrix corresponding to the grayscale pixel coordinates, used to determine the position for accumulating grayscale pixel values ​​in the blank 2D matrix. The target matching value is the currently stored value at the target matching position. The updated matching value is the value obtained by adding the grayscale pixel value to the target matching value. The accumulated pixel value is the value at the target matching position after replacing the target matching value. The accumulated two-dimensional matrix refers to the matrix that stores the accumulated grayscale pixel values ​​of all grayscale pixels after one round of processing of aligned OLED panel images. It is used to store the accumulated result after processing each aligned OLED panel image. The comprehensive accumulated two-dimensional matrix refers to the final accumulated matrix obtained after processing all aligned OLED panel images, containing the sum of the accumulated grayscale pixel values ​​in all aligned OLED panel images. The number of OLED panel images refers to the number of aligned OLED panel images in the aligned OLED panel image sequence. Obtaining the average panel image based on the comprehensive accumulated two-dimensional matrix and the number of OLED panel images means dividing each accumulated grayscale pixel value in the comprehensive accumulated two-dimensional matrix by the number of OLED panel images to obtain a new two-dimensional matrix, and using this new two-dimensional matrix as the average panel image. The purpose of performing multi-frame averaging fusion on the aligned OLED panel image sequence in this invention is to generate an image with a higher signal-to-noise ratio.

[0111] Specifically, the step of performing global brightness normalization on the average panel image to obtain a brightness-normalized image includes:

[0112] Obtain the short side size of the average panel image, determine the Gaussian kernel size based on the short side size, determine the target Gaussian kernel based on the Gaussian kernel size, and perform a Gaussian blur operation on the average panel image using the target Gaussian kernel to obtain the background brightness image.

[0113] Extract the aggregated pixel values ​​sequentially from the average panel image, and perform the following operations on each aggregated pixel value:

[0114] The accumulated pixel values ​​are converted to floating-point numbers to obtain floating-point pixel values. The background brightness pixel values ​​are then determined from the background brightness image based on the floating-point pixel values.

[0115] The normalized pixel value is calculated based on the floating-point pixel value, the background brightness pixel value, and the preset target brightness constant. The normalized pixel value is then converted to an integer type to obtain an integer normalized pixel value.

[0116] By summing the integer normalized pixel values, a brightness-normalized image is obtained.

[0117] It should be explained that the shorter side dimension of the image refers to the length of the shorter side of the width and height of the average panel image. This is used to determine the size of the Gaussian kernel, ensuring it is large enough to capture the overall brightness variations of the average panel image. Determining the Gaussian kernel size based on the shorter side dimension means dividing the shorter side dimension by a preset ratio (such as 4 or 5). It should be noted that the width and height of the target Gaussian kernel must be positive odd numbers. If the calculated Gaussian kernel size is even, add 1 to make it odd, as libraries like OpenCV require the width and height of the Gaussian kernel to be odd. The target Gaussian kernel is the Gaussian kernel used to perform the Gaussian blur operation. Performing the Gaussian blur operation means applying a Gaussian blur to the image using the target Gaussian kernel. The purpose is to generate a background brightness image that preserves the overall brightness variations of the average panel image while removing details. For example, in Python, this Gaussian blur operation can be performed using the `cv2.GaussianBlur` function from the OpenCV library. The background brightness image refers to the image after Gaussian blur processing, which preserves the overall brightness variations of the image. Floating-point type conversion refers to converting pixel values ​​from integer types to floating-point types to avoid data overflow and precision loss when calculating normalized pixel values, thereby ensuring the correctness of brightness normalization.

[0118] S3. Denoise the brightness-normalized image to obtain the denoised panel image.

[0119] It should be explained that the denoising of the brightness normalized image refers to denoising the brightness normalized image using mean filtering. The denoised panel image refers to the image after denoising processing. The purpose of denoising is to remove noise from the brightness normalized image, reduce the interference of noise on defect detection, improve the accuracy of detection, and enhance the quality of the brightness normalized image for subsequent defect detection.

[0120] S4. Obtain the defect classification result set based on the denoised panel image, summarize the defect classification result set, and obtain multiple defect classification result sets, among which the OLED panel corresponds one-to-one with the defect classification result set.

[0121] Specifically, the method for obtaining the defect classification result set based on the denoised panel image includes:

[0122] Once the grayscale image of the panel template is identified, it is quickly matched with the denoised panel image to obtain the aligned panel template image and the aligned denoised panel image.

[0123] A panel difference map is obtained based on the aligned panel template image and the aligned denoised panel image. The set of circuit functional regions is identified based on the grayscale image of the panel template. The set of circuit functional regions includes multiple circuit functional regions.

[0124] The panel difference map is divided according to the circuit functional area set to obtain the circuit sub-region set to be detected. The circuit sub-region set to be detected includes: multiple circuit sub-regions to be detected, and each circuit sub-region to be detected includes: multiple difference map circuit pixels, and the difference map circuit pixels correspond to a difference map circuit pixel value.

[0125] For each sub-region of the circuit to be tested within the set of sub-regions, the following operations are performed:

[0126] A potential defect map is obtained based on the sub-region of the circuit to be inspected, and an AI review region set is generated based on the potential defect map.

[0127] A pre-built deep learning classification model is used to perform precise judgment on each AI review region in the AI ​​review region set to obtain the defect classification result, in which the defect classification result is either a real defect or normal.

[0128] By summarizing the defect classification results, a defect classification result set corresponding to the sub-region set of the circuit to be tested is obtained.

[0129] It should be explained that the panel template grayscale image is a standard, defect-free OLED panel image used for comparison with the actual captured denoised panel image. The rapid matching of the panel template grayscale image and the denoised panel image involves using feature extraction algorithms (such as Canny edge detection, SIFT, etc.) to extract multiple feature points from both the panel template grayscale image and the denoised panel image, respectively. Rapid matching is then performed based on these feature points. An aligned panel template image refers to the panel template grayscale image that is spatially aligned with the denoised panel image after rapid matching. An aligned denoised panel image refers to the denoised panel image that is spatially aligned with the panel template grayscale image after rapid matching. The purpose of this rapid matching is to ensure that the panel template grayscale image and the denoised panel image are spatially aligned so that subsequent difference detection can be performed accurately. The process of obtaining a panel difference map based on an aligned panel template image and an aligned denoised panel image refers to sequentially extracting template pixel values ​​from the aligned panel template image, identifying the target denoised panel pixel values ​​from the aligned denoised panel image based on the template pixel values, calculating the absolute pixel values ​​of the template pixel points and the target denoised panel pixel points, summarizing the absolute pixel values, and obtaining the panel difference map.

[0130] Importantly, template pixel value refers to the pixel value of a pixel in the aligned panel template image. Target denoised panel pixel refers to the pixel value of the pixel corresponding to the template pixel extracted from the aligned denoised panel image. Absolute pixel value refers to the absolute value of the target denoised panel pixel value minus the template pixel value. The circuit functional region set is a collection of multiple circuit functional regions on the OLED panel. For example, circuit functional regions include power line regions, pad regions, etc. Dividing the panel difference map according to the circuit functional region set means determining the position and range of each circuit functional region in the circuit functional region set, and dividing the difference map into multiple sub-regions based on these positions and ranges. The circuit sub-region to be detected refers to a single sub-region in the circuit sub-region set, used for further defect detection and classification. The step of generating the AI ​​verification region set based on the potential defect map is as follows: obtain the abnormal coordinate set corresponding to all abnormal circuit pixel values ​​in the potential defect map, divide the circuit sub-region to be detected according to the abnormal coordinate set and the preset clipping size to obtain the AI ​​verification region set, wherein the AI ​​verification region corresponds one-to-one with the abnormal coordinates. The process of using a pre-built deep learning classification model to refine the classification of each AI verification region in the AI ​​verification region set involves inputting each AI verification region into the deep learning classification model to obtain multiple probability values. The model then extracts the highest probability value and uses the defect type of the AI ​​verification region corresponding to this highest probability value as the defect classification result. Anomaly coordinates refer to the coordinates of pixels marked as anomalies in the potential defect map. These coordinates are all located in the image coordinate system of the potential defect map. The anomaly coordinate set is the collection of all anomaly coordinates. The cropping size is a pre-set size for each region when extracting AI verification regions, used to ensure that the image size input to the deep learning model is consistent. The probability value is the confidence score of each category output by the deep learning classification model when classifying each AI verification region. By comparing these probability values, the most likely defect type for each AI verification region can be determined. The defect classification result set is the collection of all defect classification results. The circuit pixels in the difference map refer to the pixels in the sub-region of the circuit to be detected.

[0131] Specifically, the process of obtaining a potential defect map based on a sub-region of the circuit to be detected includes:

[0132] Extract the difference map circuit pixels sequentially from the sub-region of the circuit to be detected, and perform the following operations on each extracted difference map circuit pixel:

[0133] Set the process error range, determine the target difference map circuit pixel value based on the difference map circuit pixel value, and determine whether the target difference map circuit pixel value is within the process error range.

[0134] If the target difference map circuit pixel value is within the process error range, then the difference map circuit pixel is marked normally using the preset normal pixel value to obtain normal circuit pixels;

[0135] If the target difference map circuit pixel value is not within the process error range, then the upper limit and lower limit of the process error range are obtained. If the target difference map circuit pixel value is less than the lower limit of the range, then the difference map circuit pixel is marked normally using the normal pixel value to obtain the normal circuit pixel. If the target difference map circuit pixel value is greater than the upper limit of the range, then the difference map circuit pixel is marked abnormally using the preset abnormal pixel value to obtain the abnormal circuit pixel.

[0136] By summarizing the pixels of normal and abnormal circuits, a potential defect map is obtained.

[0137] It should be explained that the difference map circuit pixel value refers to the grayscale value of each pixel in the panel difference map, representing the difference between the denoised panel image and the panel template grayscale image. By analyzing these difference map circuit pixel values, potential defects in the OLED panel can be identified. If the target difference map circuit pixel value is within the process error range, it means that the change in the target difference map circuit pixel value is within the normal range and may be caused by process errors or normal variations, rather than a defect. Normal circuit pixels are pixels whose difference map circuit pixel values ​​are within the process error range. In the potential defect map, normal circuit pixels are marked as 0, indicating that the pixel has no defect. If the target difference map circuit pixel value is not within the process error range, it means that the change in the target difference map circuit pixel value exceeds the normal range and may indicate a defect, requiring further analysis and marking. The upper and lower limits of the range refer to the maximum and minimum allowable values ​​of the process error range, respectively. Anomaly marking is an operation used to mark difference map circuit pixels as 255. Anomaly circuit pixels are pixels whose difference map circuit pixel values ​​are not within the process error range. The potential defect map is an image generated after analyzing and marking the target difference map circuit pixel values. It displays the location of potential defects in the OLED panel. The potential defect map is used for further defect detection and classification. The target difference map circuit pixel value is the difference map circuit pixel value compared to the currently extracted difference map circuit pixel. The normal pixel value is the pixel value corresponding to the pre-defined normal circuit pixel. The abnormal pixel value is the pixel value corresponding to the pre-defined abnormal circuit pixel. For example, the normal pixel value is 0, and the abnormal pixel value is 255.

[0138] Specifically, the setting of the process error range includes:

[0139] Obtain a standard sub-region image set based on the sub-region image, and calculate the standard gray average value and standard gray standard deviation of each standard sub-region image in the standard sub-region image set to obtain the standard gray average value set and the standard gray standard deviation set.

[0140] The mean gray value and standard deviation of the region are calculated based on the standard gray value mean set and the standard gray value standard deviation set.

[0141] The process error range is constructed based on the regional gray mean, regional gray standard deviation, and preset quality control coefficient.

[0142] It needs to be explained that a standard sub-region image refers to the grayscale image corresponding to a sub-region image from an image of a known qualified OLED panel. A standard sub-region image set refers to the collection of all sub-region images acquired from a large number of known qualified OLED panels for the same specific sub-region location. The standard grayscale average value is the arithmetic mean of the grayscale values ​​of all pixels in a standard sub-region image, reflecting the overall brightness of that area on a single panel. The standard grayscale standard deviation is the standard deviation of the grayscale values ​​of all pixels in the same standard sub-region image, reflecting the uniformity within that area on a single panel. The standard grayscale average value set is the set composed of all standard grayscale average values. The standard grayscale standard deviation set is the set composed of all standard grayscale standard deviations. By processing multiple standard sub-region images, raw data on the fluctuation of that sub-region under normal production conditions can be obtained, ensuring that the subsequently established process error range is not based on empirical guesswork but is supported by a large amount of actual data, guaranteeing objectivity and accuracy. The region grayscale average value is the average of all standard grayscale average values. The regional grayscale standard deviation refers to the standard deviation of all standard grayscale standard deviation sets. The quality control coefficient is a pre-set coefficient used to define the stringency of control. In the step of constructing the process error interval based on the regional grayscale mean, regional grayscale standard deviation, and the pre-set quality control coefficient, the process error interval is defined as [regional grayscale mean - quality control coefficient * regional grayscale standard deviation, regional grayscale mean + quality control coefficient * regional grayscale standard deviation]. The process error interval is dynamically generated based on actual production data and can adapt to different production lines and different batches of process parameters. By adjusting the quality control coefficient, the stringency of detection can be easily controlled. Increasing the quality control coefficient can reduce false alarms (missing some minor anomalies), while decreasing the quality control coefficient can increase the detection rate (but may also increase false alarms).

[0143] S5. Based on the defect classification result set, complete the detection of true and false defects in OLED irregular circuit coating and process differences.

[0144] It should be explained that the defect classification result set contains detailed defect information for each OLED panel. By comprehensively analyzing and judging this information, true and false defects can be accurately distinguished. Given the irregular circuit coating and process variations in OLEDs, some situations that appear to be defects may actually be normal process characteristics, or minor anomalies caused by process fluctuations may be misjudged as defects. Testing based on the defect classification result set, combined with process knowledge and experience, allows for accurate judgment of these situations, avoiding misjudgments and omissions, improving the accuracy and reliability of testing, and ensuring product quality meets requirements.

[0145] To address the problems described in the background art, this invention identifies an OLED panel set and an industrial camera. The OLED panel set includes multiple OLED panels. OLED panels are extracted sequentially from the OLED panel set, and the following operations are performed on each extracted OLED panel: This invention sequentially extracts OLED panels for inspection, enabling batch processing of multiple OLED panels and improving inspection efficiency. Compared to inspecting individual panels one by one, this method can complete the inspection of a large number of panels in a shorter time, meeting the needs of large-scale production. Simultaneously, sequential extraction and inspection ensure the orderliness and standardization of the inspection process, facilitating the management and traceability of inspection results. It receives true and false defect detection commands, acquires brightness-normalized images based on these commands, the industrial camera, and the OLED panels, and denoises the brightness-normalized images to obtain denoised panel images. The brightness-normalized images of this invention can eliminate brightness differences between different OLED panels and between different areas of the same panel. In actual production, due to manufacturing processes and environmental factors, the brightness of OLED panels may fluctuate. This brightness difference can interfere with defect detection, leading to misjudgments or missed detections. By acquiring brightness-normalized images, the panel brightness can be adjusted to a relatively uniform level, highlighting defect features and improving the accuracy of defect detection. Simultaneously, images may be subject to various noise interferences during acquisition and transmission. Noise can blur images, obscuring the true characteristics of defects and affecting defect identification and classification. Denoising processing can effectively remove this noise, making the image clearer, enhancing defect identifiability, and providing high-quality image data for subsequent defect classification and detection. This invention acquires defect classification result sets from noisy panel images, summarizes these result sets, and obtains multiple defect classification result sets. Each OLED panel corresponds one-to-one with a defect classification result set. Different types of defects may be caused by different reasons and have varying impacts on product quality. Accurate classification results facilitate in-depth analysis and processing of defects, enabling targeted improvement measures. By summarizing the defect classification results for each OLED panel to form a result set corresponding to each panel, the defect status of each panel can be easily recorded and managed, facilitating subsequent quality traceability and analysis. Based on the defect classification result sets, the invention enables the detection of true and false defects in OLED irregular circuit coating and process differences. Therefore, this invention can improve the quality and reliability of OLED products.

[0146] like Figure 2 The diagram shown is a functional block diagram of a system for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation, provided by an embodiment of the present invention.

[0147] The OLED irregular circuit coating and process differentiation defect detection system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the OLED irregular circuit coating and process differentiation defect detection system 100 may include a system initialization module 101, an image acquisition and processing module 102, a panel defect classification module 103, and a defect detection completion module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0148] The system initialization module 101 is used to identify the OLED panel set and the industrial camera, wherein the OLED panel set includes: multiple OLED panels;

[0149] The image acquisition and processing module 102 is used to sequentially extract OLED panels from the OLED panel collection and perform the following operations on each extracted OLED panel: receive a true / false defect detection command, acquire a brightness normalized image based on the true / false defect detection command, an industrial camera and the OLED panel, and denoise the brightness normalized image to obtain a denoised panel image.

[0150] The panel defect classification module 103 is used to obtain a defect classification result set based on the denoised panel image, summarize the defect classification result set, and obtain multiple defect classification result sets, wherein the OLED panel corresponds one-to-one with the defect classification result set;

[0151] The defect detection completion module 104 is used to complete the detection of true and false defects in OLED irregular circuit coating and process differences based on the defect classification result set.

[0152] In detail, the modules in the OLED irregular circuit coating and process differentiation true and false defect detection system 100 described in this embodiment of the invention adopt the same methods as described above when in use. Figure 1 The method described herein is the same as the method for detecting real and fake defects in OLED irregular circuit coating and process differentiation, and can produce the same technical effect, so it will not be elaborated here.

[0153] like Figure 3 The diagram shown is a structural schematic of an electronic device that implements a method for detecting true and false defects in OLED irregular circuit coating and process differentiation, according to an embodiment of the present invention.

[0154] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and capable of running on the processor 10, such as a method program for detecting real and fake defects in OLED irregular circuit coating and process differentiation.

[0155] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a method for detecting true and false defects in OLED irregular circuit coating and process differentiation, but also to temporarily store data that has been output or will be output.

[0156] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., OLED irregular circuit coating and process differentiation defect detection methods), and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0157] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0158] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0159] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0160] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0161] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0162] The program for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation, stored in the memory 11 of the electronic device 1, is a combination of multiple instructions. When run in the processor 10, it can achieve the following:

[0163] The OLED panel set and industrial camera have been identified. The OLED panel set includes multiple OLED panels.

[0164] OLED panels are extracted sequentially from the OLED panel collection, and the following operations are performed on each extracted OLED panel:

[0165] Receive true and false defect detection commands, acquire brightness normalized images based on true and false defect detection commands, industrial cameras and OLED panels, denoise the brightness normalized images, and obtain denoised panel images.

[0166] Defect classification result sets are obtained based on denoised panel images, and the defect classification result sets are summarized to obtain multiple defect classification result sets. Among them, the OLED panel corresponds one-to-one with the defect classification result set.

[0167] Based on the defect classification result set, we completed the detection of true and false defects in OLED irregular circuit coating and process differences.

[0168] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 3 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0169] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0170] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0171] The OLED panel set and industrial camera have been identified. The OLED panel set includes multiple OLED panels.

[0172] OLED panels are extracted sequentially from the OLED panel collection, and the following operations are performed on each extracted OLED panel:

[0173] Receive true and false defect detection commands, acquire brightness normalized images based on true and false defect detection commands, industrial cameras and OLED panels, denoise the brightness normalized images, and obtain denoised panel images.

[0174] Defect classification result sets are obtained based on denoised panel images, and the defect classification result sets are summarized to obtain multiple defect classification result sets. Among them, the OLED panel corresponds one-to-one with the defect classification result set.

[0175] Based on the defect classification result set, we completed the detection of true and false defects in OLED irregular circuit coating and process differences.

[0176] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0177] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0178] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0179] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation, characterized in that, The method includes: The OLED panel set and industrial camera have been identified. The OLED panel set includes multiple OLED panels. OLED panels are extracted sequentially from the OLED panel collection, and the following operations are performed on each extracted OLED panel: Receive true and false defect detection commands, acquire brightness normalized images based on true and false defect detection commands, industrial cameras and OLED panels, denoise the brightness normalized images, and obtain denoised panel images. Defect classification result sets are obtained based on denoised panel images, and the defect classification result sets are summarized to obtain multiple defect classification result sets. Among them, the OLED panel corresponds one-to-one with the defect classification result set. Based on the defect classification result set, we completed the detection of true and false defects in OLED irregular circuit coating and process differences.

2. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differences as described in claim 1, characterized in that, The process of acquiring a brightness-normalized image based on true / false defect detection commands, an industrial camera, and an OLED panel includes: Based on the instructions for detecting true and false defects, an industrial camera is used to photograph the OLED panel to obtain an initial sequence of OLED panel images. Based on the initial OLED panel image sequence, an aligned OLED panel image sequence is obtained. Then, a multi-frame averaging fusion operation is performed on the aligned OLED panel image sequence to obtain the average panel image. Global brightness normalization is performed on the average panel image to obtain a brightness-normalized image.

3. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differences as described in claim 2, characterized in that, The process of obtaining the aligned OLED panel image sequence based on the initial OLED panel image sequence includes: A reference panel image is identified based on the initial OLED panel image sequence, and the reference panel image is removed from the initial OLED panel image sequence to obtain a sequence of panel images to be registered. The reference panel image is the first initial OLED panel image in the initial OLED panel image sequence, and the sequence of panel images to be registered includes multiple panel images to be registered. Extract the panel images to be registered sequentially from the sequence of panel images to be registered, and perform the following operations on each of the panel images to be registered: The position of the maximum peak is obtained based on the reference panel image and the panel image to be registered, and the position of the maximum peak is used as the relative global translation amount. A translation transformation matrix is ​​constructed based on the relative global translation amount. An image realignment operation is then performed on the panel image to be registered based on the translation transformation matrix to obtain the aligned OLED panel image. By combining the aligned OLED panel images and the reference panel image, an aligned OLED panel image sequence is obtained. The aligned OLED panel image sequence includes multiple aligned OLED panel images, and each aligned OLED panel image includes multiple grayscale pixels.

4. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differences as described in claim 3, characterized in that, The process of obtaining the maximum peak position based on the reference panel image and the panel image to be registered includes: Perform windowing operations on the reference panel image and the panel image to be registered, respectively, to obtain a windowed reference panel image and a windowed panel image to be registered; Calculate the reference complex matrix and the complex matrix to be registered for the windowed reference panel image and the windowed panel image to be registered, respectively; The cross-power spectrum is calculated based on the reference complex matrix and the complex matrix to be registered. An inverse Fourier transform is performed on the cross-power spectrum to obtain the impulse response diagram. The position of the maximum peak is extracted from the impulse response diagram.

5. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differences as described in claim 4, characterized in that, The step of performing a multi-frame averaging fusion operation on the aligned OLED panel image sequence to obtain an average panel image includes: Sequentially extract aligned OLED panel images from the aligned OLED panel image sequence, and create a blank two-dimensional matrix based on the aligned OLED panel images; According to the preset pixel extraction order, grayscale pixels are extracted from the aligned OLED panel image in sequence, and the grayscale pixel coordinates and grayscale pixel values ​​are determined based on the grayscale pixels. The target matching position is determined from the blank two-dimensional matrix based on the grayscale pixel coordinates. The target matching value is determined based on the target matching position. The grayscale pixel value is added to the target matching value to obtain the updated matching value. The updated matching value is then used to replace the target matching value to obtain the accumulated pixel value. Return to the step of extracting grayscale pixels from the aligned OLED panel image sequentially according to the preset pixel extraction order, until all grayscale pixels in the aligned OLED panel image have been extracted; Summarize the accumulated pixel values ​​to obtain an accumulated two-dimensional matrix. Use the accumulated two-dimensional matrix as a blank two-dimensional matrix and return to the step of extracting aligned OLED panel images from the aligned OLED panel image sequence sequentially until all aligned OLED panel images in the aligned OLED panel image sequence have been extracted. The summation of the two-dimensional matrices yields a comprehensive summation two-dimensional matrix, which includes multiple comprehensive summation pixel values. The number of OLED panel images is obtained from the aligned OLED panel image sequence, and the average number of panel images is obtained from the comprehensive summation of the two-dimensional matrix and the number of OLED panel images.

6. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differences as described in claim 5, characterized in that, The step of performing global brightness normalization on the average panel image to obtain a brightness-normalized image includes: Obtain the short side size of the average panel image, determine the Gaussian kernel size based on the short side size, determine the target Gaussian kernel based on the Gaussian kernel size, and perform a Gaussian blur operation on the average panel image using the target Gaussian kernel to obtain the background brightness image. Extract the aggregated pixel values ​​sequentially from the average panel image, and perform the following operations on each aggregated pixel value: The accumulated pixel values ​​are converted to floating-point numbers to obtain floating-point pixel values. The background brightness pixel values ​​are then determined from the background brightness image based on the floating-point pixel values. The normalized pixel value is calculated based on the floating-point pixel value, the background brightness pixel value, and the preset target brightness constant. The normalized pixel value is then converted to an integer type to obtain an integer normalized pixel value. By summing the integer normalized pixel values, a brightness-normalized image is obtained.

7. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differences as described in claim 6, characterized in that, The defect classification result set obtained based on the denoised panel image includes: Once the grayscale image of the panel template is identified, it is quickly matched with the denoised panel image to obtain the aligned panel template image and the aligned denoised panel image. A panel difference map is obtained based on the aligned panel template image and the aligned denoised panel image. The set of circuit functional regions is identified based on the grayscale image of the panel template. The set of circuit functional regions includes multiple circuit functional regions. The panel difference map is divided according to the circuit functional area set to obtain the circuit sub-region set to be detected. The circuit sub-region set to be detected includes: multiple circuit sub-regions to be detected, and each circuit sub-region to be detected includes: multiple difference map circuit pixels, and the difference map circuit pixels correspond to a difference map circuit pixel value. For each sub-region of the circuit to be tested within the set of sub-regions, the following operations are performed: A potential defect map is obtained based on the sub-region of the circuit to be inspected, and an AI review region set is generated based on the potential defect map. A pre-built deep learning classification model is used to perform precise judgment on each AI review region in the AI ​​review region set to obtain the defect classification result, in which the defect classification result is either a real defect or normal. By summarizing the defect classification results, a defect classification result set corresponding to the sub-region set of the circuit to be tested is obtained.

8. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation as described in claim 7, characterized in that, The process of obtaining a potential defect map based on a sub-region of the circuit to be detected includes: Extract the difference map circuit pixels sequentially from the sub-region of the circuit to be detected, and perform the following operations on each extracted difference map circuit pixel: Set the process error range, determine the target difference map circuit pixel value based on the difference map circuit pixel value, and determine whether the target difference map circuit pixel value is within the process error range. If the target difference map circuit pixel value is within the process error range, then the difference map circuit pixel is marked normally using the preset normal pixel value to obtain normal circuit pixels; If the target difference map circuit pixel value is not within the process error range, then the upper limit and lower limit of the process error range are obtained. If the target difference map circuit pixel value is less than the lower limit of the range, then the difference map circuit pixel is marked normally using the normal pixel value to obtain the normal circuit pixel. If the target difference map circuit pixel value is greater than the upper limit of the range, then the difference map circuit pixel is marked abnormally using the preset abnormal pixel value to obtain the abnormal circuit pixel. By summarizing the pixels of normal and abnormal circuits, a potential defect map is obtained.

9. The method for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation as described in claim 8, characterized in that, The setting of the process error range includes: Obtain a standard sub-region image set based on the sub-region image, and calculate the standard gray average value and standard gray standard deviation of each standard sub-region image in the standard sub-region image set to obtain the standard gray average value set and the standard gray standard deviation set. The mean gray value and standard deviation of the region are calculated based on the standard gray value mean set and the standard gray value standard deviation set. The process error range is constructed based on the regional gray mean, regional gray standard deviation, and preset quality control coefficient.

10. A system for detecting genuine and fake defects in OLED irregular circuit coating and process differentiation, characterized in that, The system includes: The system initialization module is used to identify the OLED panel set and the industrial camera. The OLED panel set includes multiple OLED panels. The image acquisition and processing module is used to sequentially extract OLED panels from the OLED panel collection and perform the following operations on each extracted OLED panel: receive true and false defect detection instructions, acquire brightness normalized images based on the true and false defect detection instructions, industrial camera and OLED panel, and denoise the brightness normalized images to obtain denoised panel images. The panel defect classification module is used to obtain a defect classification result set based on the denoised panel image, summarize the defect classification result set, and obtain multiple defect classification result sets, in which the OLED panel corresponds one-to-one with the defect classification result set; The defect detection completion module is used to complete the detection of true and false defects in OLED irregular circuit coating and process differences based on the defect classification result set.

Images