A synchronous detection method for multi-type OLED panels

Abstract

The present application relates to the technical field of panel synchronous detection, and a kind of synchronous detection method of multiple type OLED panel, comprising: determining glass substrate and the set of OLED panel of type to be detected, grid division is carried out to glass substrate, obtain standard grid set, the glass substrate in line is conveyed to the detection station of pre-construction, obtain the glass substrate to be detected, the global panel image is obtained by global shooting to the glass substrate to be detected, the global panel image is segmented, obtain segmentation panel image set, obtain panel image feature group set, according to segmentation panel image set and panel image feature group set confirm out panel information parameter set, obtain detection data group set based on panel information parameter set, obtain accurate defect coordinate set according to detection data group set, and the synchronous detection of multiple type OLED panel is completed based on accurate defect coordinate set.The present application can improve the detection efficiency and accuracy of multiple type OLED panel.

Description

Technical Field

[0001] This invention relates to the field of panel synchronous detection technology, and in particular to a method for synchronous detection of multiple types of OLED panels. Background Technology

[0002] Multi-type OLED panels refer to OLED panels with different characteristics (such as size, shape, model, and arrangement). An OLED panel is 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. Simultaneous detection is the process of detecting multiple targets at the same time.

[0003] Traditional testing methods typically involve inspecting each panel individually, which is inefficient and cannot meet the needs of large-scale production. Furthermore, traditional methods often use a fixed-size grid to divide the glass substrate, resulting in low substrate utilization. Therefore, improving the testing efficiency and accuracy of various OLED panel types is an urgent technical challenge. Summary of the Invention

[0004] This invention provides a method for simultaneous detection of multiple types of OLED panels and a computer-readable storage medium, the main purpose of which is to improve the detection efficiency and accuracy of multiple types of OLED panels.

[0005] To achieve the above objectives, the present invention provides a method for simultaneous detection of multiple types of OLED panels, comprising:

[0006] The glass substrate and the set of OLED panels of the type to be tested are determined. The glass substrate is divided into grids according to the set of OLED panels of the type to be tested to obtain a standard grid set. The set of OLED panels of the type to be tested includes multiple different types of OLED panels of the type to be tested.

[0007] In a standard grid set, each OLED panel of the type to be tested in the OLED panel set to be tested is laid flat to obtain an array glass substrate, wherein the array glass substrate contains: an array of OLED panels of the type to be tested;

[0008] Receive simultaneous detection instructions for multiple types of OLED panels, and according to the instructions, use a pre-built conveying device to transport the arranged glass substrates to the pre-built detection station to obtain the glass substrates to be tested.

[0009] A global image of the glass substrate to be inspected is obtained by taking a global picture. The global image of the global panel is then segmented to obtain a set of segmented panel images. The set of segmented panel images includes multiple segmented panel images, and each segmented panel image contains an OLED panel of a certain arrangement type.

[0010] Obtain the panel image feature set, and determine the panel information parameter set based on the segmented panel image set and the panel image feature set;

[0011] The detection data set is obtained based on the panel information parameter set, and the accurate defect coordinate set is obtained based on the detection data set.

[0012] Simultaneous inspection of multiple types of OLED panels was achieved based on a precise defect coordinate set.

[0013] Optionally, the step of dividing the glass substrate into a grid according to the OLED panel set of the type to be detected to obtain a standard grid set includes:

[0014] Obtain the panel area set and allowable rotation angle based on the OLED panel set of the type to be tested;

[0015] Obtain the substrate length and substrate width of the glass substrate, and identify the maximum panel area group and the minimum panel area group from the panel area set. The maximum panel area group includes one or more maximum panel areas, and the minimum panel area group includes one or more minimum panel areas.

[0016] The number of maximum and minimum panel areas is determined based on the maximum and minimum panel area groups.

[0017] The largest and smallest panel area groups are removed from the panel area set to obtain the remaining panel area set. The average panel area is then calculated based on the remaining panel area set.

[0018] The grid area threshold is calculated based on the average panel area, the number of panels with the largest area, and the number of panels with the smallest area.

[0019] The initial number of grid rows and columns are determined based on the grid area threshold, substrate length, and substrate width.

[0020] A set of candidate mesh partitioning schemes is generated based on the allowed rotation angle, the initial number of mesh rows, and the initial number of mesh columns. The area fit set is then calculated based on the set of candidate mesh partitioning schemes.

[0021] Obtain the optimal area fit from the area fit set, take the candidate mesh partitioning scheme corresponding to the optimal area fit as the optimal mesh partitioning scheme, and determine the standard mesh set based on the optimal mesh partitioning scheme.

[0022] Optionally, generating a set of candidate mesh partitioning schemes based on the allowed rotation angle, the initial number of mesh rows, and the initial number of mesh columns includes:

[0023] Determine the row number fine-tuning amount and column number fine-tuning amount, fix the initial grid row number to obtain a fixed grid row number, and adjust the initial grid column number using the column number fine-tuning amount to obtain multiple fine-tuned grid column numbers;

[0024] The first candidate grid partitioning scheme group is obtained by combining the fixed number of grid rows with each of the multiple fine-tuning grid column numbers;

[0025] The initial number of grid columns is fixed to obtain a fixed number of grid columns. The initial number of grid rows is adjusted using the row number fine-tuning amount to obtain multiple fine-tuning grid row numbers.

[0026] The fixed number of grid columns is combined with each of the multiple fine-tuning grid row numbers to obtain the second candidate grid partitioning scheme group;

[0027] Based on the allowed rotation angle, the number of rows and columns of the initial grid are interchanged to obtain the third candidate grid partitioning scheme;

[0028] The first candidate mesh partitioning scheme group, the second candidate mesh partitioning scheme group, and the third candidate mesh partitioning scheme are summarized to obtain the candidate mesh partitioning scheme set.

[0029] Optionally, the calculation of the area fit set based on the candidate mesh partitioning scheme set includes:

[0030] For each candidate mesh partitioning scheme in the candidate mesh partitioning scheme set, perform the following operations:

[0031] The area of ​​each grid cell is calculated based on the candidate grid partitioning scheme. The area fit is then calculated based on the grid cell area, the maximum panel area, and the minimum panel area. The formula for calculating the area fit is as follows:

[0032] ,

[0033] in, Indicates area fit. This indicates taking the minimum value. Represents the area of ​​a grid cell. Indicates the maximum panel area. This indicates the preset maximum panel area weight. Indicates the minimum panel area. This indicates the preset minimum panel area weight;

[0034] The area fit scores are aggregated to obtain an area fit score set, which includes multiple area fit scores, and each area fit score corresponds one-to-one with a candidate grid partitioning scheme.

[0035] Optionally, the acquisition of the panel image feature set includes:

[0036] Once the set of known OLED panel types is identified, perform the following operations on each known OLED panel in the set:

[0037] Collect a set of panel images of known types of OLED panels. The set of panel images includes multiple panel images, and each panel image has different lighting conditions and shooting angles.

[0038] Panel images are extracted sequentially from the panel image set. Based on the panel images, image structure data and panel size data are obtained. Based on the image structure data and panel size data, panel image features are obtained.

[0039] The panel image features are summarized to obtain a panel image feature group, wherein the panel image feature group includes multiple panel image features, and the panel image features correspond one-to-one with the panel images;

[0040] By summarizing the panel image feature groups, we obtain the panel image feature group set corresponding to the known type OLED panel set, where there is a one-to-one correspondence between the known type OLED panel and the panel image feature group.

[0041] Optionally, the step of determining the panel information parameter set based on the segmented panel image set and the panel image feature set includes:

[0042] Extract segmentation panel images sequentially from the segmentation panel image set, and perform the following operations on each extracted segmentation panel image:

[0043] Panel image feature groups are extracted sequentially from the panel image feature group set to obtain the key feature groups of the segmented panel image. The matching similarity is calculated based on the key feature groups and the panel image feature groups. The matching similarity is then summarized to obtain the matching similarity set.

[0044] The maximum matching similarity is identified from the matching similarity set, the matching panel image feature group is identified based on the maximum matching similarity, and the panel type of OLED panels arranged in the segmented panel image is identified based on the matching panel image feature group.

[0045] The OLED panels of different arrangement types are identified according to their panel types to obtain a unique panel identifier. The physical coordinates of the panels are obtained based on the segmented panel images and the OLED panels of different arrangement types.

[0046] By using panel type, panel unique identifier, and panel physical coordinates as panel information parameters, and summarizing the panel information parameters, we obtain the panel information parameter set corresponding to the segmented panel image set.

[0047] Optionally, the step of obtaining the detection data set based on the panel information parameter set includes:

[0048] The panel information parameter set is classified into categories to obtain a panel type set, which includes multiple panel type groups, and each panel type group includes one or more OLED panels of different arrangement types.

[0049] Obtain a panel defect detection model set, which includes multiple panel defect detection models, and each panel defect detection model corresponds to a different type of panel;

[0050] Panel type groups are extracted sequentially from the panel type group set, the target segmented image set is identified from the segmented panel image set based on the panel type group, and the target panel defect detection model is identified from the panel defect detection model set based on the panel type group.

[0051] The target segmentation image set is synchronously detected using the target panel defect detection model to obtain a detection data set. The detection data set includes multiple detection data, and each detection data corresponds one-to-one with the target segmentation image.

[0052] Summarize the detection data sets to obtain the detection data set corresponding to the panel type set.

[0053] Optionally, the simultaneous detection of the target segmentation image set using the target panel defect detection model to obtain a detection data set includes:

[0054] Perform a size scaling operation on each target segmentation image in the target segmentation image set to obtain a standard segmentation image set, wherein the standard segmentation image set includes multiple standard segmentation images with the same image size;

[0055] Input the standard segmented image set into the target panel defect detection model to obtain the segmented image set to be detected;

[0056] Multiple computational streams are created, and the set of images to be detected and segmented is assigned to multiple computational streams, resulting in multiple computational streams to be detected, where each computational stream corresponds one-to-one with an image to be detected and segmented.

[0057] Parallel inference computation is performed on multiple streams to obtain an initial detection data set and the input model order of the target segmentation image set is obtained.

[0058] The pre-built result aggregation thread reorders the initial detection data set according to the input model order to obtain the detection data set.

[0059] Optionally, obtaining the precise defect coordinate set based on the detection data set includes:

[0060] Obtain panel defect dataset from the detection data set. The panel defect dataset includes: multiple panel defect data, and the panel defect data includes: defect type, defect quantity, detection screen type and initial defect coordinates.

[0061] Perform the following operation on each panel defect data in the panel defect dataset:

[0062] Align the pre-built crosshair cursor with the initial defect coordinates in the panel defect data to obtain the positioning optical camera, wherein the positioning optical camera includes the crosshair cursor;

[0063] The current area image is obtained by using a positioning optical camera to capture images of OLED panels of the same arrangement type corresponding to panel defect data;

[0064] Defect features are identified in the current region image to obtain the coordinates of the defect center. The horizontal and vertical offsets of the defect center coordinates from the initial defect coordinates are calculated.

[0065] The updated defect coordinates are calculated based on the horizontal offset, vertical offset, and initial defect coordinates. The crosshair is aligned with the updated defect coordinates to obtain the crosshair coordinates. The OLED panel corresponding to the panel defect data is photographed to obtain the updated defect center coordinates.

[0066] Determine whether the updated defect center coordinates are equal to the crosshair cursor coordinates;

[0067] If the updated defect center coordinates are not equal to the crosshair coordinates, the updated defect center coordinates are used as the initial defect coordinates, and the process returns to the step of using the positioning optical camera to photograph the OLED panel corresponding to the arrangement type of the panel defect data, until the updated defect center coordinates are equal to the crosshair coordinates.

[0068] If the updated defect center coordinates are equal to the crosshair cursor coordinates, then the updated defect center coordinates are used as the precise defect coordinates.

[0069] By summarizing the precise defect coordinates, a precise defect coordinate set is obtained.

[0070] To achieve the above objectives, the present invention also provides a synchronous detection system for multiple types of OLED panels, comprising:

[0071] The OLED panel preparation module is used to determine the glass substrate and the set of OLED panels of the type to be tested. The glass substrate is divided into grids according to the set of OLED panels of the type to be tested to obtain a standard grid set. The set of OLED panels of the type to be tested includes multiple different types of OLED panels of the type to be tested.

[0072] The inspection station preparation module is used to lay out each type of OLED panel in the standard grid set to obtain an array of glass substrates. The array of glass substrates includes: an array of OLED panels of different types; receiving synchronous inspection instructions for multiple types of OLED panels; and using a pre-built conveying device to transport the array of glass substrates to the pre-built inspection station according to the synchronous inspection instructions for multiple types of OLED panels, thereby obtaining the glass substrate to be inspected.

[0073] The image acquisition and segmentation module is used to capture a global image of the glass substrate to be inspected, and to segment the global panel image to obtain a segmented panel image set. The segmented panel image set includes multiple segmented panel images, and each segmented panel image contains an OLED panel of an arrangement type.

[0074] The defect precise location module is used to acquire a panel image feature set, determine the panel information parameter set based on the segmented panel image set and the panel image feature set, acquire a detection data set based on the panel information parameter set, acquire a precise defect coordinate set based on the detection data set, and complete the synchronous detection of multiple types of OLED panels based on the precise defect coordinate set.

[0075] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0076] Memory, storing at least one instruction;

[0077] The processor executes the instructions stored in the memory to implement the synchronous detection method for multiple types of OLED panels described above.

[0078] 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 synchronous detection method for multiple types of OLED panels.

[0079] To address the problems described in the background section, this invention determines a glass substrate and a set of OLED panels of different types to be tested. The glass substrate is then divided into a grid based on the OLED panel set to be tested, resulting in a standard grid set. The OLED panel set includes multiple OLED panels of different types. This grid division ensures that the different types of panels are distributed regularly on the glass substrate, avoiding chaotic panel arrangement and improving the organization and orderliness of subsequent testing processes. Within the standard grid set, each OLED panel of a specific type is laid flat to obtain an arranged glass substrate. The arrangement of glass substrates includes: an array of OLED panels of various types. This invention lays multiple different types of OLED panels to be tested flat on a standard grid, achieving integration of multiple panel types on the same glass substrate. This allows for simultaneous testing of multiple panel types, avoiding the cumbersome process of testing different panel types separately, greatly improving testing efficiency. The invention receives synchronous testing instructions for multiple OLED panel types and, based on these instructions, uses a pre-built conveying device to transport the arranged glass substrate to a pre-built testing station, obtaining the glass substrate to be tested. This invention achieves this by receiving synchronous testing instructions and using a conveying device to transport the arranged glass substrate... The substrate is conveyed to the inspection station, automating the inspection process, reducing manual intervention, and minimizing the impact of human factors on the inspection results. This also improves the accuracy and stability of the inspection. A global image of the glass substrate to be inspected is captured, and this global image is then segmented to obtain a set of segmented panel images. Each segmented panel image includes multiple segmented panel images, and each segmented panel image contains an OLED panel of a single arrangement type. This invention segments the global panel image into a set of segmented panel images containing OLED panels of a single arrangement type, making subsequent analysis of each panel more convenient and accurate. Each segmented panel image can be processed independently. This invention improves the targeting of detection by combining panel image feature sets with segmented panel image sets to accurately extract information parameters for each panel. Based on these panel information parameter sets, a detection data set is obtained, and then a precise defect coordinate set is derived. This precise defect coordinate set allows for accurate location of defects on the panel, which is crucial for subsequent repair and processing, improving repair efficiency and accuracy. Furthermore, the precise defect coordinate set enables simultaneous detection of multiple types of OLED panels. Therefore, this invention can improve the detection efficiency and accuracy of multiple types of OLED panels. Attached Figure Description

[0080] Figure 1This is a flowchart illustrating a method for simultaneous detection of multiple types of OLED panels according to an embodiment of the present invention.

[0081] Figure 2 A functional block diagram of a synchronous detection system for multiple types of OLED panels provided in an embodiment of the present invention;

[0082] Figure 3 This is a schematic diagram of the structure of an electronic device for implementing the synchronous detection method for multiple types of OLED panels according to an embodiment of the present invention.

[0083] Explanation of reference numerals in the attached figures:

[0084] 1. Electronic equipment; 10. Processor; 11. Memory; 12. Bus; 100. Synchronous inspection system for multiple types of OLED panels; 101. OLED panel preparation module; 102. Inspection station preparation module; 103. Image acquisition and segmentation module; 104. Defect precision location module.

[0085] 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

[0086] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0087] This application provides a method for simultaneous detection of multiple types of OLED panels. The executing entity of this method includes, but is not limited to, at least one electronic device configured to execute the method provided in this application, such as a server or a terminal. In other words, the method for simultaneous detection of multiple types of OLED panels 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.

[0088] Reference Figure 1 The diagram shown is a flowchart illustrating a method for simultaneous detection of multiple types of OLED panels according to an embodiment of the present invention. In this embodiment, the method for simultaneous detection of multiple types of OLED panels includes:

[0089] S1. Determine the glass substrate and the set of OLED panels of the type to be tested. Divide the glass substrate into grids according to the set of OLED panels of the type to be tested to obtain a standard grid set. The set of OLED panels of the type to be tested includes multiple different types of OLED panels of the type to be tested.

[0090] It should be explained that the glass substrate refers to the transparent glass base used to manufacture OLED panels. Multiple OLED panels are arranged on the glass substrate, and these panels need to be inspected during the production process. The OLED panel to be inspected refers to the OLED panel that requires defect detection. An OLED panel is a 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. Examples of OLED panels include Apple's notched screen and Samsung's punch-hole screen.

[0091] Specifically, the step of dividing the glass substrate into grids according to the OLED panel set of the type to be tested to obtain a standard grid set includes:

[0092] Obtain the panel area set and allowable rotation angle based on the OLED panel set of the type to be tested;

[0093] Obtain the substrate length and substrate width of the glass substrate, and identify the maximum panel area group and the minimum panel area group from the panel area set. The maximum panel area group includes one or more maximum panel areas, and the minimum panel area group includes one or more minimum panel areas.

[0094] The number of maximum and minimum panel areas is determined based on the maximum and minimum panel area groups.

[0095] The largest and smallest panel area groups are removed from the panel area set to obtain the remaining panel area set. The average panel area is then calculated based on the remaining panel area set.

[0096] The grid area threshold is calculated based on the average panel area, the number of panels with the largest area, and the number of panels with the smallest area.

[0097] The initial number of grid rows and columns are determined based on the grid area threshold, substrate length, and substrate width.

[0098] A set of candidate mesh partitioning schemes is generated based on the allowed rotation angle, the initial number of mesh rows, and the initial number of mesh columns. The area fit set is then calculated based on the set of candidate mesh partitioning schemes.

[0099] Obtain the optimal area fit from the area fit set, take the candidate mesh partitioning scheme corresponding to the optimal area fit as the optimal mesh partitioning scheme, and determine the standard mesh set based on the optimal mesh partitioning scheme.

[0100] It should be explained that the panel area set refers to the collection of panel areas containing all OLED panel types to be tested. Panel area refers to the area of ​​each OLED panel type to be tested. Allowable rotation angle refers to the allowed rotation angle of the panel during layout. For example, allowed rotation angles are 0 degrees, 90 degrees, etc. Substrate length refers to the length of the glass substrate. Substrate width refers to the width of the glass substrate. In this invention, the substrate length and substrate width are used to determine the range of grid division. The maximum panel area group refers to the set of OLED panels of the largest tested type identified from the panel area set, used to determine the minimum grid size to ensure that the largest panel can be accommodated. The minimum panel area group refers to the set of OLED panels of the smallest tested type identified from the panel area set, used to optimize grid division to ensure that the smallest panel can also be placed reasonably. The number of maximum panel areas and the number of minimum panel areas refer to the number of maximum panel areas in the maximum panel area group and the number of minimum panel areas in the minimum panel area group, respectively. The remaining panel area set refers to the set of panel areas remaining after removing the maximum and minimum panel area groups from the panel area set. The average panel area refers to the average of the remaining panel areas, ensuring that the mesh can accommodate most panels. In the step of calculating the mesh area threshold based on the average panel area, the maximum number of panel areas, and the minimum number of panel areas, the mesh area threshold = (average panel area) / (maximum panel area) * ... (Maximum panel area + Minimum panel area) / 2. The grid area threshold is used to balance the layout requirements of panels of different sizes, ensuring that the grid can accommodate panels with the largest panel area without wasting space on panels with the smallest panel area. The step of determining the initial number of grid rows and columns based on the grid area threshold, substrate length, and substrate width is as follows: The initial number of grid rows and columns are calculated using the following formulas:

[0101] ,

[0102] ,

[0103] in, Indicates the initial number of grid rows. Indicates the substrate length. Indicates the width of the substrate. Indicates the grid area threshold. Indicates the initial number of grid columns. The symbol indicates rounding down. Optimal area fit refers to the area fit with the largest value in the area fit set. Optimal grid partitioning scheme refers to the candidate grid partitioning scheme corresponding to the optimal area fit. Determining the standard grid set based on the optimal grid partitioning scheme means confirming the grids corresponding to the optimal grid partitioning scheme as the standard grid set. The grid area threshold of this invention is not a fixed value preset by humans, but rather a customized, optimized value generated based on the different OLED panel sets of the type to be detected each time. It comprehensively considers the influence of the positive value of the panel area, generates a candidate grid partitioning scheme set, and then filters it through quantified area fit. This avoids the traditional fixed grid method using a uniform grid size to handle substrates of different sizes. To accommodate all substrates, the largest substrate size is usually used as the grid standard, leading to underutilized areas when processing smaller panels, causing resource waste and improving the utilization efficiency of the glass substrate.

[0104] Specifically, the step of generating a candidate mesh partitioning scheme set based on the allowed rotation angle, the initial number of mesh rows, and the initial number of mesh columns includes:

[0105] Determine the row number fine-tuning amount and column number fine-tuning amount, fix the initial grid row number to obtain a fixed grid row number, and adjust the initial grid column number using the column number fine-tuning amount to obtain multiple fine-tuned grid column numbers;

[0106] The first candidate grid partitioning scheme group is obtained by combining the fixed number of grid rows with each of the multiple fine-tuning grid column numbers;

[0107] The initial number of grid columns is fixed to obtain a fixed number of grid columns. The initial number of grid rows is adjusted using the row number fine-tuning amount to obtain multiple fine-tuning grid row numbers.

[0108] The fixed number of grid columns is combined with each of the multiple fine-tuning grid row numbers to obtain the second candidate grid partitioning scheme group;

[0109] Based on the allowed rotation angle, the number of rows and columns of the initial grid are interchanged to obtain the third candidate grid partitioning scheme;

[0110] The first candidate mesh partitioning scheme group, the second candidate mesh partitioning scheme group, and the third candidate mesh partitioning scheme are summarized to obtain the candidate mesh partitioning scheme set.

[0111] It should be explained that the row number fine-tuning amount is a value used to adjust the increment or decrement of the initial grid row number. The column number fine-tuning amount is a value used to adjust the increment or decrement of the initial grid column number. In this invention, both the row number fine-tuning amount and the column number fine-tuning amount are preset manually. Fixed grid row number refers to the initial grid row number that remains unchanged when generating candidate grid partitioning schemes. Fixing the initial grid row number means keeping the initial grid row number unchanged when generating candidate grid partitioning schemes. Adjusting the initial grid row number using the row number fine-tuning amount means adding or subtracting the row number fine-tuning amount from the initial grid row number. For example, if the row number fine-tuning amount is 1 and the initial grid row number is 10, adding the row number fine-tuning amount to the initial grid row number results in a fine-tuned grid row number of 11. Subtracting the row number fine-tuning amount from the initial grid row number results in a fine-tuned grid row number of 9. The fine-tuned grid row numbers of 11 and 9 are used as multiple fine-tuned grid row numbers. Fine-tuning the number of grid columns refers to the new number of columns generated by adjusting the initial number of grid columns. The first candidate grid partitioning scheme group refers to a group of candidate grid partitioning schemes composed of a fixed number of grid rows and multiple fine-tuned grid column numbers. The fixed number of grid columns refers to the initial number of grid columns that remains unchanged when generating candidate grid partitioning schemes.

[0112] Importantly, the method for adjusting the initial grid row count using row number fine-tuning is the same as the method for adjusting the initial grid column count using column number fine-tuning, and will not be repeated here. The second candidate grid partitioning scheme group refers to a group of candidate grid partitioning schemes composed of a fixed grid column count and multiple fine-tuned grid row counts. The row and column number swapping operation refers to the operation of exchanging the initial grid row count and the initial grid column count. For example, if the initial grid row count is 8 and the initial grid column count is 10, then swapping the initial grid row count and the initial grid column count results in an initial grid row count of 10 and an initial grid column count of 8. The third candidate grid partitioning scheme refers to a candidate grid partitioning scheme constructed by swapping the initial grid row count and the initial grid column count. The candidate grid partitioning scheme set refers to the set formed by summarizing the first candidate grid partitioning scheme group, the second candidate grid partitioning scheme group, and the third candidate grid partitioning scheme. In this invention, the initial grid row count and the initial grid column count are not optimal solutions. By fixing one dimension (rows) and fine-tuning another dimension (columns), a series of layout schemes with similar total grid counts but different shapes for each grid cell can be generated. Increasing the initial number of grid columns increases the total number of grids, but the width of each grid decreases while the height remains the same (because the number of rows is fixed, the total height of the glass substrate is fixed, and the number of rows determines the height of the grid). This approach is more suitable for placing tall and narrow panels. Reducing the number of columns decreases the total number of grids, but the width of each grid increases while the height remains the same. This approach is more suitable for placing wide and short panels. By generating candidate grid division schemes, the scheme whose grid shape best matches the shape of the current panel to be arranged is selected, thereby maximizing substrate utilization.

[0113] Specifically, the calculation of the area fit set based on the candidate mesh partitioning scheme set includes:

[0114] For each candidate mesh partitioning scheme in the candidate mesh partitioning scheme set, perform the following operations:

[0115] The area of ​​each grid cell is calculated based on the candidate grid partitioning scheme. The area fit is then calculated based on the grid cell area, the maximum panel area, and the minimum panel area. The formula for calculating the area fit is as follows:

[0116] ,

[0117] in, Indicates area fit. This indicates taking the minimum value. Represents the area of ​​a grid cell. Indicates the maximum panel area. This indicates the preset maximum panel area weight. Indicates the minimum panel area. This indicates the preset minimum panel area weight;

[0118] The area fit scores are aggregated to obtain an area fit score set, which includes multiple area fit scores, and each area fit score corresponds one-to-one with a candidate grid partitioning scheme.

[0119] It should be explained that in the step of calculating the grid cell area based on the candidate grid partitioning scheme, the grid cell area = substrate length. Substrate width / Number of grid rows in candidate mesh generation scheme The number of grid columns in the candidate grid generation scheme. The maximum panel area weight is a pre-set value used to measure the importance of the largest panel area in the area fit calculation, ensuring that the fit of larger panels is given sufficient consideration when calculating area fit. The minimum panel area weight is a pre-set value used to measure the importance of the smallest panel area in the area fit calculation, ensuring that the fit of smaller panels is also taken into account when calculating area fit.

[0120] Furthermore, in this invention Furthermore, the area of ​​the largest panel has a greater weight than that of the smallest panel because ensuring that the large panels can be placed is the primary task in layout design. If the large panels cannot be placed, the entire layout will fail. Small panels, due to their smaller size, are more flexible and can be placed in the gaps between the large panels. Therefore, the formula prioritizes compatibility with the large panels. In This indicates that the smaller value between the grid cell area and the maximum panel area should be used. The panel size cannot exceed the grid boundaries, and each grid cell can only hold one panel. If the grid cell area is larger than the maximum panel area, the portion of the grid cell exceeding the maximum panel area is considered unusable space for that panel. Divide by The purpose is to convert the effective area calculated by the molecule into a ratio, and the ratio is in the range of [0, 1], thereby normalizing the effective area.

[0121] The core design objective of the area fit formula in this invention is to prioritize layout feasibility rather than maximizing material utilization. The logic is as follows: when the grid cell area (D) is greater than or equal to the maximum panel area (D_max), the ratio is always 1, indicating that the grid is qualified and can accommodate a large panel, thus preventing the design from failing due to the inability to place the maximum panel; while when... hour, The ratio decreases proportionally, and this mechanism significantly reduces the total score of those candidate grid partitioning schemes, making them easier to exclude in subsequent comparisons of merits. and Similarly, this will not be elaborated upon here.

[0122] S2. In the standard grid set, each OLED panel of the type to be tested in the OLED panel set to be tested is laid flat to obtain an array glass substrate, wherein the array glass substrate contains: the array of OLED panels of the type to be tested.

[0123] It should be explained that "tiling" refers to placing each OLED panel of the type to be tested in a non-overlapping arrangement on a glass substrate. "Arranged glass substrate" refers to a glass substrate that has undergone tiling treatment, on which multiple OLED panels of the type to be tested are arranged. "Arranged type OLED panel set" refers to a collection consisting of all arranged type OLED panels. "Arranged type OLED panel" refers to an OLED panel of the type to be tested that has been arranged according to a standard grid set, with one standard grid corresponding to one OLED panel of the type to be tested. In this invention, each arranged type OLED panel on the arranged glass substrate is in an illuminated state.

[0124] S3. Receive multi-type OLED panel synchronous detection instructions. According to the multi-type OLED panel synchronous detection instructions, use a pre-built conveying device to convey the arranged glass substrate to the pre-built detection station to obtain the glass substrate to be detected.

[0125] It should be explained that the multi-type OLED panel synchronous inspection instruction refers to the instruction issued by the operator to the system to simultaneously inspect a set of OLED panels of different arrangement types. A conveyor device refers to a device used to move arranged glass substrates from one location to another. For example, a conveyor belt is a conveyor device. An inspection station is a dedicated area for OLED panel inspection, providing the necessary equipment and environment for performing the inspection task. A glass substrate to be inspected refers to a glass substrate that has been conveyed to the inspection station and is ready for inspection.

[0126] S4. Take a global picture of the glass substrate to be tested to obtain a global panel image. Segment the global panel image to obtain a segmented panel image set. The segmented panel image set includes multiple segmented panel images, and each segmented panel image contains an OLED panel of an arrangement type.

[0127] It should be explained that the "global imaging" of the glass substrate under inspection refers to capturing the entire substrate using a high-resolution, high-speed camera. The global panel image refers to an image containing all OLED panels of all arrangement types on the entire glass substrate. The "segmentation" of the global panel image refers to using image processing algorithms (such as edge detection and contour finding) to divide the global panel image into multiple independent segmented panel images. A segmented panel image is a single image containing only one arrangement type of OLED panel, separated from the global panel image.

[0128] S5. Obtain the panel image feature set, and determine the panel information parameter set based on the segmented panel image set and the panel image feature set.

[0129] Specifically, the acquisition of the panel image feature set includes:

[0130] Once the set of known OLED panel types is identified, perform the following operations on each known OLED panel in the set:

[0131] Collect a set of panel images of known types of OLED panels. The set of panel images includes multiple panel images, and each panel image has different lighting conditions and shooting angles.

[0132] Panel images are extracted sequentially from the panel image set. Based on the panel images, image structure data and panel size data are obtained. Based on the image structure data and panel size data, panel image features are obtained.

[0133] The panel image features are summarized to obtain a panel image feature group, wherein the panel image feature group includes multiple panel image features, and the panel image features correspond one-to-one with the panel images;

[0134] By summarizing the panel image feature groups, we obtain the panel image feature group set corresponding to the known type OLED panel set, where there is a one-to-one correspondence between the known type OLED panel and the panel image feature group.

[0135] It should be explained that "known type OLED panel" refers to OLED panels with known models and specifications. For example, Apple's notched screen and Samsung's punch-hole screen are known types of OLED panels. A panel image refers to an image of an OLED panel captured by a camera or other imaging device. In this invention, each panel image corresponds to different lighting conditions and shooting angles to ensure sufficient information is obtained from different conditions. The steps for obtaining structured image data and panel size data based on panel images are as follows: The panel image is converted to a grayscale image to reduce data volume and facilitate processing; the grayscale image is then converted to a binary image to facilitate edge and panel contour extraction; a pre-built denoising algorithm (such as median filtering) is used to denoise the binary image, resulting in a denoised binary image; an edge detection algorithm (such as Canny edge detection) is used to extract edge information from the denoised binary image, which helps determine the panel's contour and features; a contour extraction algorithm (such as OpenCV's findContours) is used to extract the contours from the panel image, which helps determine the panel's shape and structure; and a texture analysis method (such as Gray-Level Co-occurrence Matrix (GLCM)) is used to extract the texture features of the panel image, which helps identify the panel's surface characteristics. The contour feature data, texture feature data, and edge data are then used as structured image data. A high-precision measurement tool (such as a laser measuring instrument) is used to directly measure the dimensions of a known type of OLED panel, including length, width, and thickness. For example, texture features include pixel distribution patterns and texture direction. The process of obtaining panel size data from panel images refers to the integration of panel size data and structured image data to obtain panel image features. For example, panel image features include: (contour feature data, texture feature data, edge data, and panel size data). A panel image feature set refers to the collection of all panel image features for a known type of OLED panel. A panel image feature set is the collection of panel image feature sets for all known types of OLED panels.

[0136] Specifically, the step of determining the panel information parameter set based on the segmented panel image set and the panel image feature set includes:

[0137] Extract segmentation panel images sequentially from the segmentation panel image set, and perform the following operations on each extracted segmentation panel image:

[0138] Panel image feature groups are extracted sequentially from the panel image feature group set to obtain the key feature groups of the segmented panel image. The matching similarity is calculated based on the key feature groups and the panel image feature groups. The matching similarity is then summarized to obtain the matching similarity set.

[0139] The maximum matching similarity is identified from the matching similarity set, the matching panel image feature group is identified based on the maximum matching similarity, and the panel type of OLED panels arranged in the segmented panel image is identified based on the matching panel image feature group.

[0140] The OLED panels of different arrangement types are identified according to their panel types to obtain a unique panel identifier. The physical coordinates of the panels are obtained based on the segmented panel images and the OLED panels of different arrangement types.

[0141] By using panel type, panel unique identifier, and panel physical coordinates as panel information parameters, and summarizing the panel information parameters, we obtain the panel information parameter set corresponding to the segmented panel image set.

[0142] It should be explained that the method for obtaining the key feature groups of the segmented panel image is the same as the method for obtaining panel image feature groups based on the panel image, and will not be repeated here. The calculation of matching similarity based on the key feature groups and panel image feature groups refers to using a feature matching algorithm (such as the FLANN matcher or BFMatcher) to compare feature points in the key feature groups with feature points in the panel image feature groups. Multiple matching pairs are obtained, and the similarity between these pairs is calculated using the Euclidean distance formula. The sum of the similarities between these pairs is taken as the matching similarity. The matching similarity set refers to the set of matching similarities between a segmented panel image and all panel image feature groups in the panel image feature group set. The maximum matching similarity is the matching similarity with the highest value in the matching similarity set. The matching panel image feature group refers to the panel image feature group most similar to the key feature group of the segmented panel image. The panel type refers to the type of panel in the segmented panel image. The panel unique identifier is a unique identifier used to accurately track each type of OLED panel. For example, obtaining the physical coordinates of an OLED panel based on its segmented image and arrangement type refers to determining the position of the segmented panel image using image processing algorithms (such as edge detection and contour finding), mapping the position of the segmented panel image to the physical coordinate system of the glass substrate, and thus obtaining the panel's physical coordinates. Panel information parameters are key information describing each OLED panel, including panel type, unique panel identifier, and panel physical coordinates. The panel information parameter set is a collection of panel information parameters.

[0143] Importantly, the core purpose of this invention in calculating similarity is to identify unknown panels segmented from panel images. By comparing the features of the unknown panel with the features of all known panels in the database one by one, a matching similarity set can be obtained. The maximum matching similarity in the matching similarity set indicates that the current unknown panel is closest to the model of a known type of OLED panel in the panel image feature set. Thus, the unknown panel in the segmented panel image is identified as that model. This automated identification process aims to provide a basis for using the same defect model to detect panels of the same model when simultaneously detecting the target segmented image set, thereby improving detection efficiency.

[0144] S6. Obtain the detection data set based on the panel information parameter set, and obtain the accurate defect coordinate set based on the detection data set.

[0145] Specifically, the acquisition of the detection data set based on the panel information parameter set includes:

[0146] The panel information parameter set is classified into categories to obtain a panel type set, which includes multiple panel type groups, and each panel type group includes one or more OLED panels of different arrangement types.

[0147] Obtain a panel defect detection model set, which includes multiple panel defect detection models, and each panel defect detection model corresponds to a different type of panel;

[0148] Panel type groups are extracted sequentially from the panel type group set, the target segmented image set is identified from the segmented panel image set based on the panel type group, and the target panel defect detection model is identified from the panel defect detection model set based on the panel type group.

[0149] The target segmentation image set is synchronously detected using the target panel defect detection model to obtain a detection data set. The detection data set includes multiple detection data, and each detection data corresponds one-to-one with the target segmentation image.

[0150] Summarize the detection data sets to obtain the detection data set corresponding to the panel type set.

[0151] It should be explained that panel type refers to a type with the same characteristics (such as size, shape, etc.), used to distinguish panels with different characteristics. A panel defect detection model is an algorithm used to detect panel defects. By using a specialized detection model, defects on each panel, such as scratches, blemishes, and pixel faults, can be identified more accurately. The target segmentation image set is a collection of segmented images extracted from the segmented panel image set, associated with a specific panel type group. The target segmentation image set is used to match the corresponding target panel defect detection model for defect detection. The target panel defect detection model is a panel defect detection model extracted from the panel defect detection model set, corresponding to a specific panel type group. Synchronous detection refers to the process of simultaneously performing defect detection on multiple target segmented images in the target segmentation image set. Through synchronous detection, a large number of images can be processed quickly, improving detection speed. Detection data is the data generated by the target panel defect detection model after detecting each image in the target segmentation image set. This data includes information such as the location, type, and severity of defects. A detection data set refers to the collection of all detection data generated after classification based on the panel information parameter set, calling the corresponding panel defect detection models to perform parallel batch detection on panels of the same category. The acquisition of panel defect detection model set refers to acquiring a large number of defect images of different types of panels, training a deep learning model (such as a convolutional neural network) using the defect images of a large number of different types of panels, and then obtaining panel defect detection models corresponding to different types of panels.

[0152] Specifically, the simultaneous detection of the target segmentation image set using the target panel defect detection model to obtain a detection data set includes:

[0153] Perform a size scaling operation on each target segmentation image in the target segmentation image set to obtain a standard segmentation image set, wherein the standard segmentation image set includes multiple standard segmentation images with the same image size;

[0154] Input the standard segmented image set into the target panel defect detection model to obtain the segmented image set to be detected;

[0155] Multiple computational streams are created, and the set of images to be detected and segmented is assigned to multiple computational streams, resulting in multiple computational streams to be detected, where each computational stream corresponds one-to-one with an image to be detected and segmented.

[0156] Parallel inference computation is performed on multiple streams to obtain an initial detection data set and the input model order of the target segmentation image set is obtained.

[0157] The pre-built result aggregation thread reorders the initial detection data set according to the input model order to obtain the detection data set.

[0158] It should be explained that the size scaling operation refers to the operation of adjusting the size of each target segmentation image in the target segmentation image set to the size required for model input using image processing libraries (such as OpenCV or Pillow). The standard segmentation image set refers to the collection of segmentation images with the same size after the size scaling operation, providing a uniform input size for the model and ensuring correct processing. The segmentation image set to be detected refers to the collection of standard segmentation images prepared for inference computation after the standard segmentation image set is input into the target panel defect detection model. Creating multiple computational streams refers to creating multiple independent computational streams using a GPU programming framework (such as CUDA or cuDNN), each capable of independently performing inference computation. Multiple computational streams to be computed refer to multiple computational streams already assigned to the segmentation image set to be detected, each processing the assigned images independently, improving parallel processing capabilities. The parallel inference computation operation refers to executing inference computations of multiple computational streams simultaneously on the GPU. The initial detection data set refers to the collection of preliminary detection data obtained from the inference computations of multiple computational streams. The input model order refers to the order in which the target segmentation image set is input into the target panel defect detection model, ensuring that the final output detection results are consistent with the order of the input images. The result aggregation thread is a dedicated thread responsible for sorting and organizing the detection results returned from multiple computation streams. Reordering refers to the operation of sorting the initial detection data sets according to the input model order. This invention breaks through the limitations of traditional serial batch processing, enabling parallel inference of multiple images on the GPU, greatly utilizing the GPU's computing power and improving detection efficiency.

[0159] Specifically, obtaining the precise defect coordinate set based on the detection data set includes:

[0160] Obtain panel defect dataset from the detection data set. The panel defect dataset includes: multiple panel defect data, and the panel defect data includes: defect type, defect quantity, detection screen type and initial defect coordinates.

[0161] Perform the following operation on each panel defect data in the panel defect dataset:

[0162] Align the pre-built crosshair cursor with the initial defect coordinates in the panel defect data to obtain the positioning optical camera, wherein the positioning optical camera includes the crosshair cursor;

[0163] The current area image is obtained by using a positioning optical camera to capture images of OLED panels of the same arrangement type corresponding to panel defect data;

[0164] Defect features are identified in the current region image to obtain the coordinates of the defect center. The horizontal and vertical offsets of the defect center coordinates from the initial defect coordinates are calculated.

[0165] The updated defect coordinates are calculated based on the horizontal offset, vertical offset, and initial defect coordinates. The crosshair is aligned with the updated defect coordinates to obtain the crosshair coordinates. The OLED panel corresponding to the panel defect data is photographed to obtain the updated defect center coordinates.

[0166] Determine whether the updated defect center coordinates are equal to the crosshair cursor coordinates;

[0167] If the updated defect center coordinates are not equal to the crosshair coordinates, the updated defect center coordinates are used as the initial defect coordinates, and the process returns to the step of using the positioning optical camera to photograph the OLED panel corresponding to the arrangement type of the panel defect data, until the updated defect center coordinates are equal to the crosshair coordinates.

[0168] If the updated defect center coordinates are equal to the crosshair cursor coordinates, then the updated defect center coordinates are used as the precise defect coordinates.

[0169] By summarizing the precise defect coordinates, a precise defect coordinate set is obtained.

[0170] It should be explained that the panel defect dataset refers to a collection containing defect data from multiple panels. Defect type refers to the category of defects, such as scratches, blemishes, pixel failures, etc. Defect quantity refers to the number of defects on each panel. Detection image type refers to the specific color displayed on the OLED panel for detecting specific defects. Initial defect coordinates are the coordinates of defects in the segmented panel image identified by the defect detection model. A movable optical camera refers to an optical camera that can be moved to a designated position to capture a specific area of ​​the panel and obtain a high-resolution image. A positioning optical camera refers to an optical camera moved to the initial defect coordinate position. A crosshair refers to a cross-shaped marker in the positioning optical camera used to precisely align with the defect location. The current area image refers to the image of the area containing defects captured by the positioning optical camera. Defect feature recognition of the current area image refers to using image processing algorithms (such as edge detection, contour finding) or deep learning models to analyze the current area image and identify defect features such as shape, size, and color. Aligning the pre-constructed crosshair with the initial defect coordinates in the panel defect data refers to moving the optical camera along a fixed track to align the crosshair in the optical camera with the initial defect coordinates. The optical camera of this invention moves on a fixed track during each shot, similar to a sliding rail.

[0171] Importantly, the defect center coordinates refer to the coordinates of the center position of the defect in the current image region. The center position coordinates refer to the coordinates of the geometric center of an object (such as a defect, object, shape, etc.) within its image region. In the step of calculating the horizontal and vertical offsets between the defect center coordinates and the initial defect coordinates, the horizontal offset = the horizontal coordinate of the defect center coordinates - the horizontal coordinate of the initial defect coordinates, and the vertical offset = the vertical coordinate of the defect center coordinates - the vertical coordinate of the initial defect coordinates. The step of calculating the updated defect coordinates based on the horizontal and vertical offsets and the initial defect coordinates means adding the horizontal offset to the horizontal coordinate of the initial defect coordinates and adding the vertical offset to the vertical coordinate of the initial defect coordinates to obtain the new coordinates. The crosshair coordinates refer to the coordinates after moving the crosshair to the updated defect coordinates. The updated defect center coordinates are the corrected defect center coordinates used to further accurately determine the defect position. The accurate defect coordinates are the updated defect center coordinates equal to the crosshair coordinates. The existence of horizontal and vertical offsets in this invention is because the defect feature recognition of the current image region can only obtain a rough defect position, and its accuracy is insufficient to meet the requirements of precise positioning. To achieve sub-pixel level positioning accuracy, it is necessary to continuously calculate the offset between the crosshair and the defect center coordinates, and then perform further precise positioning and compensation correction on the initially identified defect center coordinates.

[0172] S7. Simultaneous inspection of multiple types of OLED panels based on a precise defect coordinate set.

[0173] It should be explained that the precise defect coordinate set refers to the collection of all precise defect coordinates. Using the precise defect coordinate set, simultaneous inspection of multiple types of OLED panels can be completed quickly and accurately. Inspectors can directly address problematic panels based on the defect coordinates, improving inspection and repair efficiency. Simultaneously, simultaneous inspection and accurate defect location enable timely detection and handling of defects in the panels, ensuring the product quality of multiple types of OLED panels and improving the production yield.

[0174] To address the problems described in the background section, this invention determines a glass substrate and a set of OLED panels of different types to be tested. The glass substrate is then divided into a grid based on the OLED panel set to be tested, resulting in a standard grid set. The OLED panel set includes multiple OLED panels of different types. This grid division ensures that the different types of panels are distributed regularly on the glass substrate, avoiding chaotic panel arrangement and improving the organization and orderliness of subsequent testing processes. Within the standard grid set, each OLED panel of a specific type is laid flat to obtain an arranged glass substrate. The arrangement of glass substrates includes: an array of OLED panels of various types. This invention lays multiple different types of OLED panels to be tested flat on a standard grid, achieving integration of multiple panel types on the same glass substrate. This allows for simultaneous testing of multiple panel types, avoiding the cumbersome process of testing different panel types separately, greatly improving testing efficiency. The invention receives synchronous testing instructions for multiple OLED panel types and, based on these instructions, uses a pre-built conveying device to transport the arranged glass substrate to a pre-built testing station, obtaining the glass substrate to be tested. This invention achieves this by receiving synchronous testing instructions and using a conveying device to transport the arranged glass substrate... The substrate is conveyed to the inspection station, automating the inspection process, reducing manual intervention, and minimizing the impact of human factors on the inspection results. This also improves the accuracy and stability of the inspection. A global image of the glass substrate to be inspected is captured, and this global image is then segmented to obtain a set of segmented panel images. Each segmented panel image includes multiple segmented panel images, and each segmented panel image contains an OLED panel of a single arrangement type. This invention segments the global panel image into a set of segmented panel images containing OLED panels of a single arrangement type, making subsequent analysis of each panel more convenient and accurate. Each segmented panel image can be processed independently. This invention improves the targeting of detection by combining panel image feature sets with segmented panel image sets to accurately extract information parameters for each panel. Based on these panel information parameter sets, a detection data set is obtained, and then a precise defect coordinate set is derived. This precise defect coordinate set allows for accurate location of defects on the panel, which is crucial for subsequent repair and processing, improving repair efficiency and accuracy. Furthermore, the precise defect coordinate set enables simultaneous detection of multiple types of OLED panels. Therefore, this invention can improve the detection efficiency and accuracy of multiple types of OLED panels.

[0175] like Figure 2The diagram shown is a functional block diagram of a synchronous detection system for multiple types of OLED panels provided in an embodiment of the present invention.

[0176] The synchronous detection system 100 for multiple types of OLED panels described in this invention can be installed in an electronic device. Depending on the functions implemented, the synchronous detection system 100 for multiple types of OLED panels may include an OLED panel preparation module 101, a detection station preparation module 102, an image acquisition and segmentation module 103, and a defect precision location 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.

[0177] The OLED panel preparation module 101 is used to determine the glass substrate and the set of OLED panels of the type to be tested, and to divide the glass substrate into grids according to the set of OLED panels of the type to be tested to obtain a standard grid set. The set of OLED panels of the type to be tested includes multiple different types of OLED panels of the type to be tested.

[0178] The testing station preparation module 102 is used to lay out each type of OLED panel in the OLED panel set to be tested in the standard grid set to obtain an arranged glass substrate. The arranged glass substrate includes: an arranged OLED panel set, receiving a multi-type OLED panel synchronous testing instruction, and according to the multi-type OLED panel synchronous testing instruction, using a pre-built conveying device to convey the arranged glass substrate to the pre-built testing station to obtain the glass substrate to be tested.

[0179] The image acquisition and segmentation module 103 is used to capture a global image of the glass substrate to be inspected, and to segment the global panel image to obtain a segmented panel image set. The segmented panel image set includes multiple segmented panel images, and each segmented panel image contains an OLED panel of an arrangement type.

[0180] The defect precise location module 104 is used to acquire a panel image feature set, determine a panel information parameter set based on the segmented panel image set and the panel image feature set, acquire a detection data set based on the panel information parameter set, acquire a precise defect coordinate set based on the detection data set, and complete the synchronous detection of multiple types of OLED panels based on the precise defect coordinate set.

[0181] In detail, the modules in the synchronous detection system 100 for multiple types of OLED panels described in this embodiment of the invention employ the same methods as described above. Figure 1 The method used is the same as the synchronous detection method for multiple types of OLED panels described above, and it can produce the same technical effect, so it will not be repeated here.

[0182] like Figure 3 The diagram shown is a schematic representation of an electronic device for implementing a synchronous detection method for multiple types of OLED panels according to an embodiment of the present invention.

[0183] 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 executable on the processor 10, such as a synchronous detection method program for multiple types of OLED panels.

[0184] 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 synchronous detection method program for multiple types of OLED panels, but also to temporarily store data that has been output or will be output.

[0185] 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., a synchronous detection method program for multiple types of OLED panels) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0186] 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.

[0187] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3 The 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.

[0188] 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.

[0189] 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.

[0190] 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.

[0191] The synchronous detection method program for multiple types of OLED panels 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:

[0192] The glass substrate and the set of OLED panels of the type to be tested are determined. The glass substrate is divided into grids according to the set of OLED panels of the type to be tested to obtain a standard grid set. The set of OLED panels of the type to be tested includes multiple different types of OLED panels of the type to be tested.

[0193] In a standard grid set, each OLED panel of the type to be tested in the OLED panel set to be tested is laid flat to obtain an array glass substrate, wherein the array glass substrate contains: an array of OLED panels of the type to be tested;

[0194] Receive simultaneous detection instructions for multiple types of OLED panels, and according to the instructions, use a pre-built conveying device to transport the arranged glass substrates to the pre-built detection station to obtain the glass substrates to be tested.

[0195] A global image of the glass substrate to be inspected is obtained by taking a global picture. The global image of the global panel is then segmented to obtain a set of segmented panel images. The set of segmented panel images includes multiple segmented panel images, and each segmented panel image contains an OLED panel of a certain arrangement type.

[0196] Obtain the panel image feature set, and determine the panel information parameter set based on the segmented panel image set and the panel image feature set;

[0197] The detection data set is obtained based on the panel information parameter set, and the accurate defect coordinate set is obtained based on the detection data set.

[0198] Simultaneous inspection of multiple types of OLED panels was achieved based on a precise defect coordinate set.

[0199] 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.

[0200] 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).

[0201] 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:

[0202] The glass substrate and the set of OLED panels of the type to be tested are determined. The glass substrate is divided into grids according to the set of OLED panels of the type to be tested to obtain a standard grid set. The set of OLED panels of the type to be tested includes multiple different types of OLED panels of the type to be tested.

[0203] In a standard grid set, each OLED panel of the type to be tested in the OLED panel set to be tested is laid flat to obtain an array glass substrate, wherein the array glass substrate contains: an array of OLED panels of the type to be tested;

[0204] Receive simultaneous detection instructions for multiple types of OLED panels, and according to the instructions, use a pre-built conveying device to transport the arranged glass substrates to the pre-built detection station to obtain the glass substrates to be tested.

[0205] A global image of the glass substrate to be inspected is obtained by taking a global picture. The global image of the global panel is then segmented to obtain a set of segmented panel images. The set of segmented panel images includes multiple segmented panel images, and each segmented panel image contains an OLED panel of a certain arrangement type.

[0206] Obtain the panel image feature set, and determine the panel information parameter set based on the segmented panel image set and the panel image feature set;

[0207] The detection data set is obtained based on the panel information parameter set, and the accurate defect coordinate set is obtained based on the detection data set.

[0208] Simultaneous inspection of multiple types of OLED panels was achieved based on a precise defect coordinate set.

[0209] 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.

[0210] 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.

[0211] 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.

[0212] 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.

[0213] 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 synchronously detecting multi-type OLED panels, characterized in that, The method comprises: determining a glass substrate and a set of to-be-detected type OLED panels, performing grid division on the glass substrate according to the set of to-be-detected type OLED panels, and obtaining a set of standard grids, wherein the set of to-be-detected type OLED panels comprises a plurality of different kinds of to-be-detected type OLED panels; in the set of standard grids, tiling each to-be-detected type OLED panel in the set of to-be-detected type OLED panels to obtain an arranged glass substrate, wherein the arranged glass substrate comprises a set of arranged type OLED panels; receiving a multi-type OLED panel synchronous detection instruction, and transmitting the arranged glass substrate to a pre-constructed detection station by using a pre-constructed conveying device according to the multi-type OLED panel synchronous detection instruction to obtain a to-be-detected glass substrate; performing global shooting on the to-be-detected glass substrate to obtain a global panel image, and performing segmentation on the global panel image to obtain a set of segmented panel images, wherein the set of segmented panel images comprises a plurality of segmented panel images, and each segmented panel image comprises an arranged type OLED panel; obtaining a set of panel image feature groups, and confirming a set of panel information parameters according to the set of segmented panel images and the set of panel image feature groups; wherein the obtaining of the set of panel image feature groups comprises: confirming a set of known type OLED panels, and performing the following operations on each known type OLED panel in the set of known type OLED panels: collecting a set of panel images of the known type OLED panel, wherein the set of panel images comprises a plurality of panel images, and the illumination conditions and shooting angles corresponding to each panel image are different; sequentially extracting panel images from the set of panel images, obtaining image structured data and panel size data based on the panel images, and obtaining panel image features based on the image structured data and the panel size data; summarizing the panel image features to obtain a panel image feature group, wherein the panel image feature group comprises a plurality of panel image features, and each panel image feature corresponds to a panel image; summarizing the panel image feature groups to obtain a set of panel image feature groups corresponding to the set of known type OLED panels, wherein each known type OLED panel corresponds to a panel image feature group; wherein the confirming of the set of panel information parameters according to the set of segmented panel images and the set of panel image feature groups comprises: sequentially extracting segmented panel images from the set of segmented panel images, and performing the following operations on each extracted segmented panel image: sequentially extracting panel image feature groups from the set of panel image feature groups, obtaining a key feature group of the segmented panel image, calculating a matching similarity according to the key feature group and the panel image feature group, summarizing the matching similarity to obtain a set of matching similarities; confirming a maximum matching similarity from the set of matching similarities, confirming a matching panel image feature group according to the maximum matching similarity, and confirming a panel type of the arranged type OLED panel in the segmented panel image according to the matching panel image feature group; identifying the arranged type OLED panel according to the panel type to obtain a panel unique identifier, and obtaining panel physical coordinates according to the segmented panel image and the arranged type OLED panel. The panel type, the panel unique identifier and the panel physical coordinates are taken as panel information parameters, and the panel information parameters are summarized to obtain a panel information parameter set corresponding to the segmented panel image set; The detection data group set is obtained based on the panel information parameter set, and the accurate defect coordinate set is obtained according to the detection data group set; The detection data group set is obtained based on the panel information parameter set, and the accurate defect coordinate set is obtained according to the detection data group set; The panel information parameter set is classified by type to obtain a panel type group set, wherein the panel type group set includes a plurality of panel type groups, and each panel type group includes one or more arrangement type OLED panels; A panel defect detection model set is obtained, wherein the panel defect detection model set includes a plurality of panel defect detection models, and each panel defect detection model corresponds to a different type of panel; The panel type groups are extracted from the panel type group set in sequence, the target segmented image set is determined from the panel type group in the segmented panel image set, and the target panel defect detection model is determined from the panel type group in the panel defect detection model set; The target panel defect detection model is used to synchronously detect the target segmented image set to obtain detection data, wherein the detection data includes a plurality of detection data, and the detection data corresponds to the target segmented image one by one; The detection data group set corresponding to the panel type group set is obtained by summarizing the detection data group; The detection data group set is obtained based on the panel information parameter set, and the accurate defect coordinate set is obtained according to the detection data group set; The panel defect data set is obtained from the detection data group set, wherein the panel defect data set includes a plurality of panel defect data, and the panel defect data includes defect type, defect quantity, detection picture type and initial defect coordinates; The following operations are performed on each panel defect data in the panel defect data set: The pre-constructed crosshair is aligned with the initial defect coordinates in the panel defect data to obtain a positioning optical camera, wherein the positioning optical camera contains the crosshair; The arrangement type OLED panel corresponding to the panel defect data is photographed by using the positioning optical camera to obtain a current area image; The defect feature recognition is performed on the current area image to obtain a defect center coordinate, and the horizontal offset and the vertical offset of the defect center coordinate and the initial defect coordinates are calculated; The updated defect coordinates are calculated according to the horizontal offset, the vertical offset and the initial defect coordinates, the crosshair is aligned with the updated defect coordinates to obtain crosshair coordinates, and the arrangement type OLED panel corresponding to the panel defect data is photographed to obtain updated defect center coordinates; It is judged whether the updated defect center coordinates are equal to the crosshair coordinates; If the updated defect center coordinates are not equal to the crosshair coordinates, the updated defect center coordinates are taken as the initial defect coordinates, and the step of photographing the arrangement type OLED panel corresponding to the panel defect data by using the positioning optical camera is returned until the updated defect center coordinates are equal to the crosshair coordinates; If the updated defect center coordinates are equal to the crosshair coordinates, the updated defect center coordinates are taken as the accurate defect coordinates; The accurate defect coordinate set is obtained by summarizing the accurate defect coordinates; The synchronous detection of the multi-type OLED panel is completed based on the accurate defect coordinate set. 2.The method of claim 1, wherein, The step of dividing the glass substrate into grids according to the OLED panel set of the type to be tested to obtain a standard grid set includes: Obtain the panel area set and allowable rotation angle based on the OLED panel set of the type to be tested; Obtain the substrate length and substrate width of the glass substrate, and identify the maximum panel area group and the minimum panel area group from the panel area set. The maximum panel area group includes one or more maximum panel areas, and the minimum panel area group includes one or more minimum panel areas. The number of maximum and minimum panel areas is determined based on the maximum and minimum panel area groups. The largest and smallest panel area groups are removed from the panel area set to obtain the remaining panel area set. The average panel area is then calculated based on the remaining panel area set. The grid area threshold is calculated based on the average panel area, the number of panels with the largest area, and the number of panels with the smallest area. The initial number of grid rows and columns are determined based on the grid area threshold, substrate length, and substrate width. A set of candidate mesh partitioning schemes is generated based on the allowed rotation angle, the initial number of mesh rows, and the initial number of mesh columns. The area fit set is then calculated based on the set of candidate mesh partitioning schemes. Obtain the optimal area fit from the area fit set, take the candidate mesh partitioning scheme corresponding to the optimal area fit as the optimal mesh partitioning scheme, and determine the standard mesh set based on the optimal mesh partitioning scheme. 3.The method of claim 2, wherein the detecting the plurality of types of OLED panels in synchronization comprises: detecting the plurality of types of OLED panels in synchronization by using a plurality of types of synchronization signals corresponding to the plurality of types of OLED panels. The process of generating a candidate mesh partitioning scheme set based on the allowed rotation angle, the initial number of mesh rows, and the initial number of mesh columns includes: Determine the row number fine-tuning amount and column number fine-tuning amount, fix the initial grid row number to obtain a fixed grid row number, and adjust the initial grid column number using the column number fine-tuning amount to obtain multiple fine-tuned grid column numbers; The first candidate grid partitioning scheme group is obtained by combining the fixed number of grid rows with each of the multiple fine-tuning grid column numbers; The initial number of grid columns is fixed to obtain a fixed number of grid columns. The initial number of grid rows is adjusted using the row number fine-tuning amount to obtain multiple fine-tuning grid row numbers. The fixed number of grid columns is combined with each of the multiple fine-tuning grid row numbers to obtain the second candidate grid partitioning scheme group; Based on the allowed rotation angle, the number of rows and columns of the initial grid are interchanged to obtain the third candidate grid partitioning scheme; The first candidate mesh partitioning scheme group, the second candidate mesh partitioning scheme group, and the third candidate mesh partitioning scheme are summarized to obtain the candidate mesh partitioning scheme set. 4.The method of claim 3, wherein the detecting the plurality of types of OLED panels in synchronization comprises: detecting the plurality of types of OLED panels in synchronization by using a plurality of types of synchronization signals corresponding to the plurality of types of OLED panels. The calculation of the area fit set based on the candidate grid partitioning scheme set includes: For each candidate mesh partitioning scheme in the candidate mesh partitioning scheme set, perform the following operations: The area of ​​each grid cell is calculated based on the candidate grid partitioning scheme. The area fit is then calculated based on the grid cell area, the maximum panel area, and the minimum panel area. The formula for calculating the area fit is as follows: ， wherein, represents an area fitness, represents taking a minimum value, represents a grid cell area, represents a maximum panel area, represents a preset maximum panel area weight, represents a minimum panel area, represents a preset minimum panel area weight; The area fit scores are aggregated to obtain an area fit score set, which includes multiple area fit scores, and each area fit score corresponds one-to-one with a candidate grid partitioning scheme. 5.The method of claim 4, wherein the method further comprises: determining a type of the OLED panel based on the first and second signals. The method of simultaneously detecting the target segmentation image set using the target panel defect detection model to obtain a detection data set includes: Perform a size scaling operation on each target segmentation image in the target segmentation image set to obtain a standard segmentation image set, wherein the standard segmentation image set includes multiple standard segmentation images with the same image size; Input the standard segmented image set into the target panel defect detection model to obtain the segmented image set to be detected; Multiple computational streams are created, and the set of images to be detected and segmented is assigned to multiple computational streams, resulting in multiple computational streams to be detected, where each computational stream corresponds one-to-one with an image to be detected and segmented. Parallel inference computation is performed on multiple streams to obtain an initial detection data set and the input model order of the target segmentation image set is obtained. The pre-built result aggregation thread reorders the initial detection data set according to the input model order to obtain the detection data set.

6. A system for synchronized detection using the multi-type OLED panel of claim 1, wherein, The system includes: The OLED panel preparation module is used to determine the glass substrate and the set of OLED panels of the type to be tested. The glass substrate is divided into grids according to the set of OLED panels of the type to be tested to obtain a standard grid set. The set of OLED panels of the type to be tested includes multiple different types of OLED panels of the type to be tested. The inspection station preparation module is used to lay out each type of OLED panel in the standard grid set to obtain an array of glass substrates. The array of glass substrates includes: an array of OLED panels of different types; receiving synchronous inspection instructions for multiple types of OLED panels; and using a pre-built conveying device to transport the array of glass substrates to the pre-built inspection station according to the synchronous inspection instructions for multiple types of OLED panels, thereby obtaining the glass substrate to be inspected. The image acquisition and segmentation module is used to capture a global image of the glass substrate to be inspected, and to segment the global panel image to obtain a segmented panel image set. The segmented panel image set includes multiple segmented panel images, and each segmented panel image contains an OLED panel of an arrangement type. The defect precise location module is used to acquire a panel image feature set, determine the panel information parameter set based on the segmented panel image set and the panel image feature set, acquire a detection data set based on the panel information parameter set, acquire a precise defect coordinate set based on the detection data set, and complete the synchronous detection of multiple types of OLED panels based on the precise defect coordinate set.

Images