Methods, devices, computer devices, media, and products for detecting electrostatic shock

Abstract

The application discloses a method, device, computer equipment, medium and product for detecting electrostatic shock, the method comprising: obtaining a distribution map of defective display panels in multiple motherboards; the motherboard comprises multiple display panels, and the display panels are cut from the motherboard; determining a high-incidence position of defective display panels in the motherboard based on the distribution map of defective display panels in multiple motherboards; obtaining a distribution map of defective positions in display panels of the high-incidence position of defective display panels in multiple motherboards; determining whether the defective positions in the display panels are gathered in X coordinates and / or Y coordinates based on the distribution map of defective positions in display panels of the high-incidence position of defective display panels in multiple motherboards; each position of the display panel comprises different X coordinates and / or Y coordinates; the display panel with defective positions gathered in X coordinates and / or Y coordinates is determined as a target panel with the risk of electrostatic shock, and production equipment contacted by multiple target panels is determined as a risk equipment generating electrostatic shock.

Description

Technical Field

[0001] This application belongs to the field of display technology, and specifically relates to a method, apparatus, computer equipment, medium and product for detecting electrostatic discharge damage. Background Technology

[0002] OLED (Organic Light-Emitting Diode) devices are extremely sensitive to ESD (Electrostatic Discharge) and are easily damaged by electrostatic discharge, leading to failure. However, there is currently no method to directly detect ESD defects; identification of ESD risks mainly relies on the experience of business experts, which is inefficient and has a high false negative rate. Summary of the Invention

[0003] This application provides a method, apparatus, computer equipment, medium, and product for detecting electrostatic discharge (ESD) damage, aiming to at least partially solve the problems of low efficiency and high false negative rate in identifying ESD risks relying on the experience of business experts.

[0004] In a first aspect of this application, a method for detecting electrostatic discharge damage is provided, the method comprising: Obtain a distribution diagram of defective display panels in multiple motherboards; the motherboards include multiple display panels, and the display panels are cut from the motherboards; Based on the distribution map of defective display panels in the multiple motherboards, the locations of high-incidence defects in the motherboards are determined; Obtain a distribution map of defective points in the display panel at locations with high defect incidence in the multiple motherboards; Based on the distribution map of defective points in the display panel at high-incidence locations in the multiple motherboards, it is determined whether the defective points in the display panel are clustered in the X and / or Y coordinates; each point in the display panel includes different X and / or Y coordinates. Display panels with defective points clustered on the X and / or Y coordinates are identified as target panels with a risk of electrostatic discharge (ESD) damage, and production equipment that has been in contact with multiple of these target panels is identified as equipment at risk of ESD damage.

[0005] In some embodiments, obtaining a distribution map of defective display panels among multiple motherboards includes: Obtain a distribution map of defective points in each display panel of the motherboard; Based on the distribution map of defective points in each display panel in the master board, a distribution map of defective display panels in the master board is generated; the defective display panels in the master board are the display panels in the master board that have defective points.

[0006] In some embodiments, the distribution map of defective display panels in the motherboard includes the number of defective points in each display panel of the motherboard; determining the high-incidence locations of defects in the motherboard based on the distribution map of defective display panels in the plurality of motherboards includes: Count the number of defective display panels at the same location in the multiple motherboards; The number of defective display panels at each location in the motherboard is input into the aggregation identification model to determine the high-incidence locations of defects in the motherboard; the number of defective display panels at the same high-incidence location in the multiple motherboards is greater than the number of defective display panels at the same non-high-incidence location in the multiple motherboards, and the difference between the two is greater than the number threshold.

[0007] In some embodiments, determining whether defective points in the display panel are clustered in the X and / or Y coordinates based on the distribution map of defective points in the multiple motherboards with high defect incidence locations includes: If the high-frequency defect location in the motherboard is a single display panel, then the sum of the number of defective points at the same point in the display panels at the same location in the multiple motherboards is counted. If the high-frequency defect locations in the motherboard are the same row of the display panels, then the sum of the number of defect locations at the same point in the same row of the display panels in the multiple motherboards is counted. If the high-frequency defect locations in the motherboard are the same column of the display panels, then the sum of the number of defect locations at the same point in the same column of the display panels in the multiple motherboards is counted. Based on the sum of the number of defective points counted at each point in the display panel, it is determined whether the defective points in the display panel are clustered in the X and / or Y coordinates.

[0008] In some embodiments, the display panel is divided into multiple first regions arranged along the X-coordinate direction or multiple second regions arranged along the Y-coordinate direction, wherein the number of points in each first region is the same, and the number of points in each second region is the same; the step of determining whether the defective points in the display panel are clustered in the X and / or Y coordinates based on the sum of the number of defective points counted from each point in the display panel includes: Based on the sum of the number of defective points counted at each point in the display panel, the number of defective points in each first area and the number of defective points in each second area of ​​the display panel are counted. The number of defective points in each of the first areas of the display panel is input into the clustering identification model to select the first area where defective points are clustered; the number of defective points in the first area where defective points are clustered is greater than the number of defective points in the first area where defective points are not clustered, and the difference between the two is greater than the number threshold. If the clustering identification model selects the first area where defective points are clustered, then it is determined that the defective points in the display panel are clustered on the Y coordinate. The number of defective points counted in each of the second areas of the display panel is input into the clustering identification model to select the second area where defective points are clustered; the number of defective points in the second area where defective points are clustered is greater than the number of defective points in the second area where defective points are not clustered, and the difference between the two is greater than the number threshold. If the clustering identification model selects the second region where defective points are clustered, then it is determined that the defective points in the display panel are clustered on the X coordinate.

[0009] In some implementations, the number of points in the first area and the number of points in the second area are negatively correlated with the sum of the number of defective points counted for each point in the display panel.

[0010] In some embodiments, determining whether defective points in the display panel are clustered in the X and / or Y coordinates based on the sum of the number of defective points counted at each point in the display panel further includes: The system displays to the user the number of defective points in the first area where defective points are clustered and the number of defective points in the second area where defective points are clustered. In response to the first and second regions selected by the user, the first and second regions selected by the user are marked as non-aggregated bad points, and the aggregation identification model is optimized.

[0011] In some embodiments, the display panel that clusters defective points on the X and / or Y coordinates is identified as a target panel with a risk of electrostatic discharge (ESD) damage, including: If it is determined that the defective points in the display panel are clustered in the X and / or Y coordinates, then the display panel is identified as the target panel; If it is determined that the defective points in the display panel are not clustered on the X coordinate and not clustered on the Y coordinate, then the display panel is determined to be a non-target panel; Based on the defect distribution map of the points in the target panel and the defect distribution map of the points in the non-target panel, the similarity between the distribution positions of the defective points in the target panel and the distribution positions of the defective points in the non-target panel is determined. The target panels with similarity scores higher than the similarity threshold are reclassified as non-target panels.

[0012] In some embodiments, the method further includes: Obtain a distribution map of the edge-defective display panels among the multiple motherboards; Based on the distribution map of the edge defect display panels in the multiple motherboards, the high-incidence locations of edge defects in the motherboards are determined; Obtain a distribution map of edge defect points in the display panel, corresponding to the locations with high incidence of edge defects among the multiple motherboards; Based on the distribution map of edge defect points in the display panel at high-incidence locations of edge defects in the multiple motherboards, determine whether the edge defect points in the display panel are clustered in the X and / or Y coordinates; The display panel whose edge defects are clustered on the X and / or Y coordinates is determined as the target panel.

[0013] In some embodiments, obtaining a distribution map of edge defect points in the display panel at high-incidence locations among the plurality of motherboards includes: Determine the edge area of ​​the display panel; In the distribution map of defective points in the display panel, only the defective points in the edge area are retained, thus obtaining the distribution map of edge defective points in the display panel.

[0014] In some embodiments, determining the edge region of the display panel includes: Based on the distribution map of defective points in the display panel, the convex hull algorithm is used for edge recognition to obtain the first polygonal outline of the display panel; The first polygonal outline is divided into four parts using cross-shaped cutting lines; The four parts are superimposed in a symmetrical relationship, and the edge recognition is performed using the convex hull algorithm to obtain the second polygonal outline of the display panel; A sliding window is used to slide along the X-coordinate direction on the contour of the second polygon, and the maximum value of the Y-coordinate is continuously updated; If the maximum value of the Y coordinate has not been updated when the sliding window has been slid a set number of times, then the position where the Y coordinate of the sliding window was first not updated is determined as the corner dividing point, and the third polygonal outline of the display panel is obtained. The major axis of the ellipse is determined based on the corner boundary points of the third polygonal contour, and the ellipse is fitted to form the fourth polygonal contour of the display panel. The area extending inward from the fourth polygonal outline is defined as the edge area of ​​the display panel.

[0015] In some embodiments, the display panel that clusters defective points on the X and / or Y coordinates is identified as a target panel with a risk of electrostatic discharge (ESD) damage, including: Based on the defect distribution map of each target panel, the sum of the number of defective points at the same point in all the target panels is calculated. Based on the sum of the number of defective points counted at each point in the target panel, the high-incidence area of ​​defects in the target panel is determined. The target panel whose defect points are located outside the high-incidence defect area is redefined as a non-target panel.

[0016] In some embodiments, the production equipment that has been in contact with all of the target panels is identified as a risk device for electrostatic discharge (ESD) injury, including: Obtain the production equipment that each of the target panels has been manufactured on; Based on the production equipment that each of the target panels has been through, determine the target equipment that each of the target panels has been through; If each of the target panels is in contact with the target device, then the target device is identified as the risk device.

[0017] In some implementations, identifying the target device as the risk device if all of the target panels are in contact with the target device includes: Based on the defect distribution map of the points in the target panel, the distribution area of ​​the defect points in the target panel is determined; If each of the target panels is in contact with the target device, and the contact area coincides with the distribution area of ​​the defective points in the target panel, then the target device is identified as the risk device.

[0018] In a second aspect of this application, an apparatus for detecting electrostatic discharge (ESD) damage is provided, the apparatus comprising: The first acquisition module is configured to acquire a distribution map of defective display panels in multiple motherboards; the motherboards include multiple display panels, and the display panels are cut from the motherboards; The location determination module is configured to determine the high-frequency locations of defects in the motherboards based on the distribution map of defective display panels in the plurality of motherboards; The second acquisition module is configured to acquire a distribution map of defective points in the display panel at locations with high defect incidence in the plurality of motherboards; The clustering judgment module is configured to determine whether the defective points in the display panel are clustered in the X and / or Y coordinates based on the distribution map of defective points in the display panel at high defective locations in the multiple motherboards; each point in the display panel includes different X and / or Y coordinates. The electrostatic discharge (ESD) determination module is configured to identify display panels with defective points clustered on the X and / or Y coordinates as target panels with a risk of ESD damage, and to identify production equipment that has been in contact with multiple target panels as equipment at risk of ESD damage.

[0019] In a third aspect of this application, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method provided in any embodiment of the first aspect.

[0020] In a fourth aspect of this application, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in any embodiment of the first aspect.

[0021] In a fifth aspect of this application, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the steps of the method provided in any embodiment of the first aspect.

[0022] According to one or more embodiments of this application, a method, apparatus, computer device, medium, and product for detecting electrostatic discharge (ESD) damage are first obtained, which shows the distribution of defective display panels in multiple motherboards. Based on this distribution, the high-incidence locations of defects in the motherboards are determined. The motherboards include multiple display panels, which are cut from the motherboards. Then, a distribution map of defective points in the display panels at the high-incidence locations in the motherboards is obtained. Based on this distribution map, it is determined whether the defective points in the display panels cluster on the X and / or Y coordinates. Each point on the display panel includes different... The system automatically identifies the display panels on the X and / or Y coordinates where defective points cluster, and determines the target panels that have been in contact with multiple target panels as the risk equipment for electrostatic discharge (ESD). This system addresses the issue of ESD causing fixed-type clustering of defects, automatically identifying target panels where defective points cluster on the X and / or Y coordinates, and then determining the production equipment that has been in contact with all target panels as the risk equipment for ESD. This allows business personnel to quickly locate the risk location for practical verification, improving the efficiency and high false negative rate of relying on business expert experience to identify ESD risks. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart illustrating a method for detecting electrostatic discharge damage according to one or more embodiments of this application is shown.

[0025] Figure 2 It shows Figure 1 A schematic diagram of the distribution of defective display panels in the motherboard in step S101.

[0026] Figure 3 It shows Figure 1 Step S104 displays a schematic diagram showing the distribution of the number of defective points on each X or Y coordinate in the display panel.

[0027] Figure 4 It shows Figure 1 The flowchart of step S101 is shown.

[0028] Figure 5 It shows Figure 1 The flowchart of step S102 is shown.

[0029] Figure 6 It shows Figure 1 The flowchart of step S104 is shown.

[0030] Figure 7 It shows Figure 6 The flowchart of step S404 is shown.

[0031] Figure 8 It shows Figure 6 Another flowchart for step S404.

[0032] Figure 9 It shows Figure 8 A schematic diagram showing the distribution of the number of defective points in the first or second area of ​​the display panel after step S506 is executed.

[0033] Figure 10 It shows Figure 8 A schematic diagram showing the distribution of the number of defective points in the first or second area of ​​the display panel after step S507 is executed.

[0034] Figure 11 It shows Figure 1 A flowchart of step S105.

[0035] Figure 12 A flowchart illustrating a method for detecting electrostatic discharge damage according to another embodiment of this application is shown.

[0036] Figure 13 It shows Figure 12 The flowchart of step S108 is shown.

[0037] Figure 14 It shows Figure 13 A schematic diagram of the position of the edge area of ​​the display panel in step S701.

[0038] Figure 15 It shows Figure 13 The flowchart of step S701 is shown.

[0039] Figure 16 It shows Figure 15 In step S801, a schematic diagram of the shape of the first polygonal outline of the display panel is shown.

[0040] Figure 17 It shows Figure 15 In step S803, a schematic diagram of the shape of the second polygonal outline of the display panel is shown.

[0041] Figure 18 It shows Figure 15 The schematic diagram of the third polygonal outline of the display panel in step S805.

[0042] Figure 19 It shows Figure 15 The schematic diagram of the fourth polygonal outline of the display panel in step S806.

[0043] Figure 20 It shows Figure 1 Another flowchart of step S105.

[0044] Figure 21 It shows Figure 1 Another flowchart of step S105.

[0045] Figure 22 It shows Figure 21 The flowchart of step S1003 is shown.

[0046] Figure 23 A structural block diagram of an apparatus for detecting electrostatic discharge (ESD) damage according to one or more embodiments of this application is shown.

[0047] Figure 24 A structural block diagram of a computer device according to one or more embodiments of this application is shown. Detailed Implementation

[0048] To enable those skilled in the art to more clearly understand this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0049] OLED devices are extremely sensitive to ESD and are easily damaged by electrostatic discharge (ESD). However, there is currently no method to directly detect ESD defects. But when ESD occurs, it can lead to other types of defects, such as clusters of defective display panels in the motherboard, like bright spots or X-lines, and these defective points in the display panel tend to cluster along the X or Y coordinates. Utilizing this characteristic, the inventors have proposed a method for detecting ESD damage. This method can automatically identify clusters of defective display panels in the motherboard and clusters of defective points in the display panel, thereby locating the equipment involved in the ESD risk during the production process. This allows business personnel to quickly pinpoint the location of the risk and conduct physical verification.

[0050] Figure 1 For a flowchart illustrating the method for detecting electrostatic discharge (ESD) damage in one or more embodiments of this application, please refer to [link / reference]. Figure 1 According to a first aspect of this application, a method for detecting electrostatic discharge damage is provided, the method comprising the following steps S101 to S105.

[0051] Step S101: Obtain the distribution map of defective display panels in multiple motherboards.

[0052] In this embodiment, the motherboard includes multiple display panels, which are cut from the motherboard. Figure 2 for Figure 1 Please refer to the schematic diagram of the distribution of defective display panels in the motherboard in step S101. Figure 2 A motherboard 10 includes multiple display panels 20, which can be arranged in multiple rows and columns within the motherboard 10. The motherboard 10 is cut along the boundary lines between adjacent display panels 20 to obtain the display panels 20. By highlighting the defective display panels 20 in the motherboard 10, a distribution map of the defective display panels in the motherboard can be obtained. For example, the defective display panels in the motherboard can be display panels with defective points in the motherboard.

[0053] by Figure 2For example, the motherboard 10 includes twenty-five display panels 20, which are arranged in five rows and five columns. The display panel 20 in the second row and fourth column and the display panel 20 in the third row and third column are highlighted, and they are both defective display panels with faulty points.

[0054] Step S102: Based on the distribution map of defective display panels in multiple motherboards, determine the high-incidence locations of defects in the motherboards.

[0055] In this embodiment, defective display panels at the same location in multiple motherboards are superimposed, and the location with a significantly larger number of defective display panels is designated as a high-incidence defect location. The number of defective display panels at the same high-incidence defect location is greater than the number of defective display panels at the same non-high-incidence defect location, and the difference between the two is greater than a quantity threshold.

[0056] Or with Figure 2 For example, if the display panel 20 in the second row and fourth column and the display panel 20 in the third row and third column of most of the motherboards 10 are defective display panels, then the second row and fourth column and the third row and third column are the high-frequency defective positions in the motherboards 10.

[0057] Step S103: Obtain the distribution map of defective points in the display panel at high defect locations in multiple motherboards.

[0058] In this embodiment, the display panel includes multiple points, which can be arranged in multiple rows along the Y-axis and multiple columns along the X-axis. Each point includes an X-coordinate and a Y-coordinate, and the X-coordinates and / or Y-coordinates of two points are different. The distribution diagram of defective points in the display panel is similar to the distribution diagram of defective display panels in the motherboard, and the defective points in the display panel can be highlighted.

[0059] A defective display panel in the motherboard is a display panel with defective points on the motherboard. That is, if a display panel has at least one defective point, then that display panel is considered a defective display panel in the motherboard. Therefore, if the distribution map of defective points in a display panel shows at least one defective point, then that display panel will be highlighted as a defective display panel in the motherboard. In practical applications, based on the distribution map of defective points in each display panel of the motherboard, a distribution map of defective display panels in the motherboard can be generated.

[0060] Or with Figure 2 For example, if the second row, fourth column and the third row, third column are the locations with the highest incidence of defects in the motherboard 10, then obtain the distribution map of the defective points in the display panel 20 in the second row, fourth column and the distribution map of the defective points in the display panel 20 in the third row, third column of each motherboard 10.

[0061] Step S104: Based on the distribution map of defective points in the display panel at high defect locations in multiple motherboards, determine whether the defective points in the display panel are clustered in the X and / or Y coordinates.

[0062] In this embodiment, each point on the display panel includes a different X coordinate and / or Y coordinate.

[0063] First, defective points at the same location on each display panel are superimposed. Then, defective points at the same X-coordinate are superimposed. By comparing the number of defective points at each X-coordinate, it is determined whether the defective points are clustered on the X-coordinate. When the number of defective points on the X-coordinate is significantly larger, the defective points are clustered on the X-coordinate. The number of defective points on the X-coordinate where defective points are clustered is greater than the number of defective points on the X-coordinate where defective points are not clustered, and the difference between the two is greater than a quantity threshold.

[0064] First, defective points at the same location on each display panel are superimposed. Then, defective points at the same Y-coordinate are superimposed. By comparing the number of defective points at each Y-coordinate, it is determined whether the defective points are clustered on the Y-coordinate. When the number of defective points on the Y-coordinate is significantly larger, the defective points are clustered on the Y-coordinate. The number of defective points on the Y-coordinate where defective points are clustered is greater than the number of defective points on the Y-coordinate where defective points are not clustered, and the difference between the two is greater than a quantity threshold.

[0065] Figure 3 for Figure 1 In step S104, please refer to the schematic diagram showing the distribution of the number of defective points on each X or Y coordinate in the display panel. Figure 3 The x-axis represents the X or Y coordinate, and the y-axis represents the number of defective points. If the number of defective points at the X or Y coordinate is significantly large, then the defective points are considered to be clustered along that X or Y coordinate. Figure 3 For example, the number of defective points at the first and last X / Y coordinates is significantly larger, indicating that defective points are clustered at the first and last X / Y coordinates.

[0066] Step S105: Display panels with defective points clustered on the X and / or Y coordinates are identified as target panels with a risk of electrostatic discharge (ESD) damage, and production equipment that has been in contact with multiple target panels is identified as equipment at risk of ESD damage.

[0067] In this embodiment, the target panel posing a risk of electrostatic discharge (ESD) damage may include a display panel with defective points clustered on the X and / or Y coordinates. The equipment posing a risk of ESD damage may include production equipment that has been in contact with multiple target panels.

[0068] If defective points on a display panel are clustered along the X and / or Y coordinates, this display panel is identified as the target panel. Among the production equipment that each target panel has been in contact with, equipment that has been in contact with multiple target panels is selected as high-risk equipment.

[0069] The aforementioned method for detecting electrostatic discharge (ESD) damage first obtains a distribution map of defective display panels from multiple motherboards. Based on this distribution map, it identifies high-incidence locations of defects within the motherboards. The motherboards comprise multiple display panels, which are cut from them. This allows for the initial screening of display panels located at high-incidence locations within the motherboards. This reduces the amount of data processed subsequently, improving the efficiency of ESD damage detection. Furthermore, it eliminates interference from other defects, enhancing the accuracy of ESD defect identification. A single motherboard can be cut into multiple display panels, some with ESD defects and others without. If these panels are mixed together, and the defective points on the ESD-affected display panels are relatively few, the clustering of ESD defects may be masked by the defects on other display panels. Here, we first determine the high-frequency locations of defects in the motherboard based on the distribution map of defective display panels in multiple motherboards, such as row and column clusters, single panel clusters, half-panel clusters, and similar situations. We can first filter out the display panels located in the high-frequency locations of defects in the motherboard, and the subsequent clustering judgment on the X and / or Y coordinates will only use these display panels located in the high-frequency locations of defects in the motherboard.

[0070] Then, the distribution map of defective points in the display panel at high-incidence locations of defects in multiple motherboards is obtained. Based on the distribution map of defective points in the display panel at high-incidence locations of defects in multiple motherboards, it is determined whether the defective points in the display panel are clustered in the X and / or Y coordinates. Each point in the display panel includes different X and / or Y coordinates. This can help discover potential ESD risks and provide early warnings, identify potential ESD trends as early as possible, and carry out investigations, thereby improving the efficiency of the entire process.

[0071] If ESD actually occurs, defects will inevitably be concentrated on the production equipment, and there will be contact between the production equipment and the display panel. Therefore, display panels with defect points clustered on the X and / or Y coordinates are identified as target panels with a risk of electrostatic discharge (ESD). Production equipment that has been in contact with multiple target panels is identified as risk equipment for ESD. This allows for further analysis and confirmation of potential ESD risks. This approach addresses the issue of ESD causing fixed-type clustering of defects by automatically identifying target panels with defect points clustered on the X and / or Y coordinates, and then determining the production equipment that has been in contact with all target panels as risk equipment for ESD. This facilitates business personnel in quickly locating the risk for practical verification, improving the efficiency and high false negative rate of relying on business expert experience to identify ESD risks, enabling earlier control and improving yield.

[0072] Figure 4 for Figure 1 Please refer to the flowchart of step S101. Figure 4 In some embodiments, step S101 may include steps S201 to S202.

[0073] Step S201: Obtain the distribution map of defective points in each display panel of the motherboard.

[0074] In this embodiment, the distribution diagrams of defective points in each display panel of the same motherboard are obtained. (Still using...) Figure 2 For example, the distribution diagrams of defective points in the twenty-five display panels 20 of a motherboard 10 are obtained respectively.

[0075] Step S202: Based on the distribution map of defective points in each display panel in the motherboard, generate a distribution map of defective display panels in the motherboard.

[0076] In this embodiment, the defective display panel in the motherboard is the display panel with defective points in the motherboard.

[0077] For example, step S202 may include: determining whether the display panel is a defective display panel based on the distribution map of defective points in the display panel; and generating a distribution map of defective display panels in the motherboard based on the distribution positions of the determined defective display panels in the motherboard.

[0078] For example, determining whether a display panel is a defective display panel based on the distribution map of defective points in the display panel may include: determining whether there are defective points in the display panel based on the distribution map of defective points in the display panel; if there are defective points in the display panel, then the display panel is a defective display panel; if there are no defective points in the display panel, then the display panel is a qualified display panel.

[0079] For each display panel, if a display panel has one or more defective pixels, then that display panel is considered a defective display panel. Let's take... Figure 2 For example, there are defective points in both the second row, fourth column display panel 20 and the third row, third column display panel 20. These are defective display panels in the motherboard 10 and are highlighted in the distribution diagram of defective display panels in the motherboard 10.

[0080] In the above embodiments, the distribution map of defective points in each display panel of the motherboard is first obtained, and then the distribution map of defective display panels in the motherboard is generated based on the distribution map of defective points in each display panel of the motherboard. The distribution map of defective points in the display panels can be used to determine whether there are defective points in the display panels, thereby determining whether there are defective points in each display panel of the motherboard, and thus obtaining the defective display panels in the motherboard, and finally generating the distribution map of defective display panels in the motherboard.

[0081] Figure 5 for Figure 1 Please refer to the flowchart for step S102. Figure 5 Accordingly, step S102 may include steps S301 to S302.

[0082] Step S301: Count the number of defective display panels at the same location in multiple motherboards.

[0083] In this embodiment, defective display panels at the same location in multiple motherboards are superimposed to obtain the number of defective display panels at the same location in multiple motherboards. Still using... Figure 2 For example, if all display panels 20 in the second row and fourth column of each motherboard 10 are defective display panels, then the number of defective display panels in the second row and fourth column of multiple motherboards is equal to the number of motherboards; if all display panels 20 in the third row and third column of half of the motherboards 10 are defective display panels, then the number of defective display panels in the third row and third column of multiple motherboards is equal to half the number of motherboards.

[0084] Step S302: Input the number of defective display panels counted at each location in the motherboard into the aggregation identification model to determine the high-incidence locations of defects in the motherboard.

[0085] In this embodiment, the number of defective display panels at the same high-frequency defect location in multiple motherboards is greater than the number of defective display panels at the same non-high-frequency defect location in multiple motherboards, and the difference between the two is greater than a quantity threshold. For example, the high-frequency defect location in the motherboard can be one location or multiple locations. Let's take... Figure 2 For example, the second row, fourth column and the third row, third column can both be high-risk locations for defects in the motherboard.

[0086] Cluster identification models can identify significantly larger data points from multiple datasets. Here, the number of defective display panels at various locations on the motherboard is input into the cluster identification model. The model can then identify locations with significantly larger numbers of defective display panels, defining these as high-incidence defect locations on the motherboard. The number of defective display panels at these high-incidence locations is greater than the number of defective display panels at other locations (non-high-incidence locations), and the difference between the two is greater than a threshold.

[0087] For example, cluster recognition models can employ Deep Sets. Deep Sets are a neural network architecture specifically designed to handle input data such as sets. The core idea is that the order of input elements does not affect the output result.

[0088] In the above embodiments, the number of defective display panels at the same location in multiple motherboards is first counted, and then the number of defective display panels counted at each location in the motherboard is input into the clustering identification model. The clustering identification model can be used to identify significantly larger values ​​among the number of defective display panels counted at each location, so as to determine the high-incidence location of defects in the motherboard.

[0089] Figure 6 for Figure 1 Please refer to the flowchart for step S104. Figure 6 In some embodiments, step S104 may include steps S401 to S404.

[0090] Step S401: If the high-frequency defect location in the motherboard is a single display panel, then the sum of the number of defective points at the same point in the display panels at the same location in multiple motherboards is counted.

[0091] In this embodiment, if no two display panels in the same row or column are considered high-frequency defect locations, then the high-frequency defect location is a single display panel. In this case, the defective points at the same location on multiple motherboards within the same high-frequency defect location are superimposed to obtain the sum of the number of defective points at the same location. Still using... Figure 2 For example, the second row and fourth column and the third row and third column are both high-frequency defect locations, and they are located in different rows and columns. The defective points in the same position in the display panel of the second row and fourth column of multiple motherboards are superimposed, and the defective points in the same position in the display panel of the third row and third column of multiple motherboards are superimposed.

[0092] Step S402: If the high-frequency defect locations in the motherboard are display panels in the same row, then the sum of the number of defect locations at the same point in the same row of display panels in multiple motherboards is counted.

[0093] In this embodiment, if at least two display panels in the same row are at high defect positions, then the high defect rate position is the display panel in that row. In this case, the defective points at the same position in the display panels of multiple motherboards in that row are superimposed to obtain the sum of the number of defective points at the same position. Assuming that the second column of the second row and the fourth column of the second row are both high defect rate positions, and are located in different columns within the same row, the defective points at the same position in the display panels of the second row of multiple motherboards are superimposed.

[0094] In practical applications, if some high-frequency defect locations are a single display panel, and other high-frequency defect locations are display panels in the same row, then for the case where the high-frequency defect location is a single display panel, the sum of the number of defective points at the same point in the display panels at the same location in multiple motherboards is counted. For the case where the high-frequency defect location is a display panel in the same row, the sum of the number of defective points at the same point in the display panels in the same row in multiple motherboards is counted.

[0095] Step S403: If the high-frequency defect locations in the motherboard are display panels in the same column, then the sum of the number of defect locations at the same point in the same column of display panels in multiple motherboards is counted.

[0096] In this embodiment, if at least two display panels in the same column are at high defect rates, then the high defect rate position is the display panel in that column. In this case, the defective points at the same position in the display panels of multiple motherboards in that column are superimposed to obtain the sum of the number of defective points at the same position. Assuming that the third row, third column and the fourth row, third column are both high defect rate positions, and are located in the same column but different rows, the defective points at the same position in the display panels of the third column of multiple motherboards are superimposed.

[0097] In practical applications, if some high-frequency defect locations are a single display panel, and other high-frequency defect locations are display panels in the same column, then for the case where the high-frequency defect location is a single display panel, the sum of the number of defective points at the same point in the display panels at the same location in multiple motherboards is counted. For the case where the high-frequency defect location is a display panel in the same column, the sum of the number of defective points at the same point in the display panels in the same column in multiple motherboards is counted.

[0098] If some of the high-frequency defect locations are in the same row of the display panels, and another part of the high-frequency defect locations are in the same column of the display panels, then for the case where the high-frequency defect locations are in the same row of the display panels, the sum of the number of defective points at the same location in the display panels in the same row of the multiple motherboards is counted; and for the case where the high-frequency defect locations are in the same column of the display panels, the sum of the number of defective points at the same location in the display panels in the same column of the multiple motherboards is counted.

[0099] If some high-frequency defect locations are a single display panel, another group of high-frequency defect locations are display panels in the same row, and yet another group of high-frequency defect locations are display panels in the same column, then for the case where the high-frequency defect location is a single display panel, the sum of the number of defective points at the same location in the display panels of the same position across multiple motherboards is calculated. For the case where the high-frequency defect location is a display panel in the same row, the sum of the number of defective points at the same location in the display panels of the same row across multiple motherboards is calculated. And for the case where the high-frequency defect location is a display panel in the same column, the sum of the number of defective points at the same location in the display panels of the same column across multiple motherboards is calculated.

[0100] Step S404: Based on the sum of the number of defective points counted at each point on the display panel, determine whether the defective points on the display panel are clustered on the X coordinate and / or Y coordinate.

[0101] In the above embodiments, firstly, based on the different distribution of high-incidence defect locations in the motherboard, the sum of the number of defective points at the same location in different display panels is counted accordingly. If the high-incidence defect location in the motherboard is a single display panel, the sum of the number of defective points at the same location in display panels of multiple motherboards is counted. If the high-incidence defect location in the motherboard is a row of display panels, the sum of the number of defective points at the same location in display panels of multiple motherboards is counted. If the high-incidence defect location in the motherboard is a column of display panels, the sum of the number of defective points at the same location in display panels of multiple motherboards is counted. Then, based on the sum of the number of defective points counted at each location in the display panel, it is determined whether the defective points in the display panel are clustered on the X coordinate and / or Y coordinate. Thus, based on the distribution map of defective points in display panels at high-incidence defect locations in multiple motherboards, it is determined whether the defective points in the display panel are clustered on the X coordinate and / or Y coordinate.

[0102] In some embodiments, the display panel may be divided into a plurality of first regions arranged along the X-coordinate direction or a plurality of second regions arranged along the Y-coordinate direction, wherein the number of points in each first region is the same and the number of points in each second region is the same.

[0103] In this embodiment, the display panel can be divided into multiple first regions arranged along the X-coordinate direction, and the number of points in each first region is the same. Alternatively, the display panel can be divided into multiple second regions arranged along the Y-coordinate direction, and the number of points in each second region is the same.

[0104] The number of defective points on a display panel is typically around 200,000 * 150,000. Direct cluster analysis would result in an extremely complex algorithm. A better approach is to first divide the display panel into multiple first regions arranged along the X-axis or multiple second regions arranged along the Y-axis. Then, cluster analysis is performed using the number of defective points within each of the first or second regions. This method can determine the approximate location of the defective areas, meeting practical needs.

[0105] Figure 7 for Figure 6 Please refer to the flowchart for step S404. Figure 7 Accordingly, step S404 may include steps S501 to S505.

[0106] Step S501: Based on the sum of the number of defective points counted at each point in the display panel, count the number of defective points in each first area and the number of defective points in each second area of ​​the display panel.

[0107] In this embodiment, the sum of the number of defective points counted at each point within the same first area is added together to obtain the total number of defective points within the first area. Figure 3 In the bar chart, each bar can also represent the number of defective points in each first region. That is, each bar arranged sequentially along the coordinate axis can correspond to the number of defective points in each first region arranged along the X coordinate direction.

[0108] The total number of defective points in the second region is obtained by summing the statistical counts of all points within the same second region. Figure 3 In the bar chart, each bar can also represent the number of defective points in each second region. That is, each bar arranged sequentially along the coordinate axis can correspond to the number of defective points in each second region arranged along the Y-coordinate.

[0109] Step S502: Input the number of defective points counted in each first area of ​​the display panel into the clustering identification model to select the first area where defective points are clustered.

[0110] In this embodiment, the number of defective points in the first region where defective points are clustered is greater than the number of defective points in the first region where defective points are not clustered, and the difference between the two is greater than a quantity threshold. For example, the number of first regions where defective points are clustered can be one or more. Figure 3 Taking the column as an example, which represents the number of defective points in the first area, the first and last first areas are the first areas where defective points are concentrated.

[0111] The clustering identification model can identify significantly larger data points from multiple datasets. Here, the number of defective points in each first region of the display panel is input into the clustering identification model. The model can then identify the first region with a significantly larger number of defective points from the statistical counts of each first region, designating it as the first region where defective points are clustered. The number of defective points in the first region where defective points are clustered is greater than the number of defective points in other first regions (i.e., first regions where defective points are not clustered), and the difference between the two is greater than a threshold.

[0112] Step S503: If the clustering identification model selects the first area where defective points are clustered, then it is determined that the defective points on the display panel are clustered on the Y coordinate.

[0113] In this embodiment, the clustering identification model selects a first region where at least one defective point is clustered, and determines that the defective points on the display panel are clustered on the Y coordinate. Conversely, if the clustering identification model does not select a first region where defective points are clustered, it determines that the defective points on the display panel are not clustered on the Y coordinate.

[0114] Step S504: Input the number of defective points counted in each second area of ​​the display panel into the clustering identification model to select the second area where defective points are clustered.

[0115] In this embodiment, the number of defective points in the second region where defective points are clustered is greater than the number of defective points in the second region where defective points are not clustered, and the difference between the two is greater than a quantity threshold. For example, the number of second regions where defective points are clustered can be one or more. Figure 3 Taking the column as an example, which represents the number of defective points in the second region, the first and last second regions are the first regions where defective points are concentrated.

[0116] The clustering identification model can identify significantly larger data points from multiple datasets. Here, the number of defective points counted in each second region of the display panel is input into the clustering identification model. The model can then identify the first region with a significantly larger number of defective points from the counts in each second region, designating it as the second region where defective points cluster. The number of defective points in this clustered second region is greater than the number of defective points in other second regions (i.e., second regions where defective points are not clustered), and the difference between the two is greater than a threshold.

[0117] Step S505: If the clustering identification model selects the second area where defective points are clustered, then it is determined that the defective points in the display panel are clustered on the X coordinate.

[0118] In this embodiment, the clustering identification model selects a second region where at least one defective point is clustered, determining that the defective points on the display panel are clustered on the X-coordinate. Conversely, if the clustering identification model does not select a second region where defective points are clustered, it is determined that the defective points on the display panel are not clustered on the X-coordinate.

[0119] In the above embodiment, firstly, based on the sum of the number of defective points counted at each point on the display panel, the number of defective points in each first area and each second area of ​​the display panel are counted. Then, the number of defective points counted in each first area and each second area of ​​the display panel are respectively input into the clustering recognition model to select the first area and the second area where defective points are clustered. If the clustering recognition model selects the first area where defective points are clustered, it is determined that the defective points in the display panel are clustered on the Y coordinate; if the clustering recognition model selects the second area where defective points are clustered, it is determined that the defective points in the display panel are clustered on the X coordinate, thereby determining whether the defective points in the display panel are clustered on the X and / or Y coordinates.

[0120] For example, the number of points in the first area and the number of points in the second area can be negatively correlated with the sum of the number of defective points counted for each point on the display panel.

[0121] In this embodiment, if the sum of the number of defective points counted at each point on the display panel is large, then the number of points in the first area and the number of points in the second area are both small; conversely, if the sum of the number of defective points counted at each point on the display panel is small, then the number of points in the first area and the number of points in the second area are both large. In practical applications, the range of values ​​for the number of points in the first area and the number of points in the second area can be defined, and then the number of points in the first area and the number of points in the second area can be selected within the defined range based on the sum of the number of defective points at each point on the display surface.

[0122] In the above embodiments, the number of points in the first area and the number of points in the second area are negatively correlated with the sum of the number of defective points counted at each point in the display panel. In this way, if the sum of the number of defective points counted at each point in the display panel is large, the number of points in the first area and the number of points in the second area will be small, which can more accurately locate the cluster of defective points. If the sum of the number of defective points counted at each point in the display panel is small, the number of points in the first area and the number of points in the second area will be large, which is beneficial for identifying the cluster of defective points.

[0123] Figure 8 for Figure 6 For another flowchart of step S404, please refer to [link / reference]. Figure 8For example, step S404 may also include steps S506 to S507.

[0124] Step S506: Show the user the number of defective points in the first area where defective points are clustered and the number of defective points in the second area where defective points are clustered.

[0125] Figure 9 for Figure 8 Please refer to the schematic diagram showing the distribution of the number of defective points in the first or second area where defective points are clustered in the display panel after step S506 is executed. Figure 9 The first first / second region, the last first / second region, and the first / second region in the middle are all first / second regions selected by the cluster recognition model.

[0126] In step S507, in response to the first and second regions selected by the user, the first and second regions selected by the user are marked as non-aggregated bad points, and the aggregation identification model is optimized.

[0127] Figure 10 for Figure 8 Please refer to the schematic diagram showing the distribution of the number of defective points in the first or second area where defective points are clustered in the display panel after step S507 is executed. Figure 10 The user selects a first / second area located in the middle, and this first / second area changes from a first / second area with a high concentration of defective points to a first / second area with no high concentration of defective points.

[0128] In the above embodiments, after the clustering identification model selects the first and second regions where defective points are clustered, the number of defective points in the first region and the number of defective points in the second region are displayed to the user. In response to the first and second regions selected by the user, the first and second regions selected by the user are marked as defective points not clustered. The clustering identification model is optimized to correct the first and second regions where defective points are clustered selected by the clustering identification model. The clustering identification model is gradually optimized according to user preferences.

[0129] Figure 11 for Figure 1 Please refer to a flowchart of step S105. Figure 11 For example, step S105 may include steps S601 to S604.

[0130] Step S601: If it is determined that the defective points in the display panel are clustered in the X coordinate and / or Y coordinate, then the display panel is determined as the target panel.

[0131] In this embodiment, if it is determined that the defective points in the display panel are clustered on the X coordinate, or that the defective points in the display panel are clustered on the Y coordinate, or that the defective points in the display panel are clustered on both the X and Y coordinates, then the display panel is determined as the target panel.

[0132] Step S602: If it is determined that the defective points in the display panel are not clustered on the X coordinate and not clustered on the Y coordinate, then the display panel is determined as a non-target panel.

[0133] In this embodiment, if it is determined that the defective points in the display panel are not clustered on the X coordinate and are not clustered on the Y coordinate, then the display panel is determined to be a non-target panel.

[0134] Step S603: Based on the defect distribution map of the points in the target panel and the defect distribution map of the points in the non-target panel, determine the similarity between the distribution locations of defective points in the target panel and the distribution locations of defective points in the non-target panel.

[0135] Step S604: Target panels with similarity higher than the similarity threshold are reclassified as non-target panels.

[0136] In practical applications, a similarity network can be set up inside the clustering recognition model. The input of the similarity network can be two sets: the target panel and the non-target panel, which are determined by whether the defective points in the display panel are clustered in the X and / or Y coordinates. In this case, the loss function needs to consider the classification accuracy of each element and the similarity between the two sets.

[0137] The pattern of defective points clustering in a display panel with ESD defects should differ significantly from that of a non-ESD defective display panel. If the clustering of defective points in the target panel and the non-target panel is similar, the target panel is likely not ESD defective; if the clustering of defective points in the target panel and the non-target panel is significantly different, the target panel is very likely ESD defective.

[0138] In the above embodiments, the display panel is first divided into target panels and non-target panels based on whether the defective points in the display panel are clustered on the X and / or Y coordinates. The defective points in the target panel are clustered on the X and / or Y coordinates, while the defective points in the non-target panel are not clustered on either the X or Y coordinates. Then, based on the defective distribution map of the points in the target panel and the defective distribution map of the points in the non-target panel, the similarity between the distribution positions of the defective points in the target panel and the distribution positions of the defective points in the non-target panel is determined. Finally, the target panel with a similarity higher than the similarity threshold is re-determined as a non-target panel. This requires the clustering recognition model to determine whether the defective points in the display panel are clustered on the X and / or Y coordinates. The determined target panel and non-target panel are different in clustering position, which reduces false alarms and excludes display panels with non-ESD defects.

[0139] A display panel defect is not necessarily caused by ESD. Only a consistent pattern of clustering is likely an ESD defect. ESD defects are more prevalent at the corners of the display panel, meaning they mainly occur at the edges and corners, while their probability of occurrence in other areas (i.e., the center of the display panel) is relatively low. If corner defects and non-corner defects are mixed together in cluster analysis, the corner defects may be masked by the non-corner defects. For example, the center of the display panel may have a large number of non-ESD defects, while the corners may only have a small number of ESD defects. In this case, the corner defects are relatively few and irrelevant to the clustering judgment. The focus of the clustering analysis is on the center of the display panel, but the center of the display panel does not exhibit a specific clustering pattern, so it will be judged as having no ESD defects. However, if you look at the corners alone, you will find ESD defects.

[0140] To address the above issues, in addition to performing cluster analysis on the entire display panel using the methods described above, cluster analysis can also be performed on the corners of the display panel. Unlike cluster analysis on the entire panel, which uses all defective points, corner cluster analysis only marks defective points at the edges as defective, allowing for separate analysis of display panels with corner defects. This can serve as a supplement for special cases. Both cluster analyses can be performed in parallel, and an alert can be triggered if either one detects a cluster of ESD defects.

[0141] Figure 12 For a flowchart illustrating a method for detecting electrostatic discharge damage in another embodiment of this application, please refer to [link / reference]. Figure 12 In some embodiments, the method may further include steps S106 to S110.

[0142] Step S106: Obtain the distribution map of edge-defective display panels among multiple motherboards.

[0143] In this embodiment, the edge-defect display panel is a display panel with edge defects, meaning that the defective points of the display panel are located at least at the edge of the display panel. That is, all defective points of the display panel can be located at the edge of the display panel, or some defective points of the display panel can be located at the edge of the display panel, and other defective points can be located at non-edge positions of the display panel.

[0144] The distribution diagram of edge defective display panels in the motherboard is similar to the distribution diagram of defective display panels in the motherboard in step S101. The difference is that the defective points of the defective display panels in the distribution diagram of defective display panels in the motherboard can be located at any position on the display panel, while the defective points of the edge defective display panels in the distribution diagram of edge defective display panels in the motherboard are at least located at the edge of the display panel.

[0145] Step S107: Based on the distribution map of edge defect display panels in multiple motherboards, determine the high-incidence locations of edge defects in the motherboards.

[0146] In this embodiment, edge-defect display panels at the same location from multiple motherboards are stacked. The location with a significantly larger number of edge-defect display panels is the high-incidence location for edge defects. The number of edge-defect display panels at the same high-incidence location is greater than the number of edge-defect display panels at the same non-high-incidence location, and the difference between the two is greater than a quantity threshold.

[0147] Step S108: Obtain the distribution map of edge defect points in the display panel at locations with high incidence of edge defects in multiple motherboards.

[0148] In this embodiment, edge defective points are defective points located at the edge of the display panel. The distribution map of edge defective points in the display panel is similar to the distribution map of defective points in the display panel in step S103, except that the defective points in the distribution map of defective points in the display panel can be located at any position on the display panel, while the defective points in the distribution map of edge defective points in the display panel must be located at the edge of the display panel.

[0149] A display panel with edge defects in the motherboard is a display panel with edge defects located on the motherboard. That is, if a display panel has at least one edge defect, then that display panel is considered an edge-defective display panel in the motherboard. Therefore, if the distribution map of edge defects in a display panel shows at least one edge defect, that display panel will be highlighted as an edge-defective display panel in the motherboard. In practical applications, based on the distribution map of edge defects in each display panel in the motherboard, a distribution map of edge-defective display panels in the motherboard can be generated.

[0150] Step S109: Based on the distribution map of edge defect points in the display panel at high-incidence locations of edge defects in multiple motherboards, determine whether the edge defect points in the display panel are clustered in the X and / or Y coordinates.

[0151] In this embodiment, edge defective points at the same location in each display panel are first superimposed, and then edge defective points at the same X-coordinate are superimposed. By comparing the number of edge defective points at each X-coordinate, it is determined whether the edge defective points are clustered on the X-coordinate. When the number of edge defective points on the X-coordinate is significantly larger, the edge defective points are clustered on the X-coordinate. The number of edge defective points on the X-coordinate where edge defective points are clustered is greater than the number of edge defective points on the X-coordinate where edge defective points are not clustered, and the difference between the two is greater than a quantity threshold.

[0152] First, the defective edge points at the same location on each display panel are superimposed. Then, the defective edge points at the same Y-coordinate are superimposed. By comparing the number of defective edge points at each Y-coordinate, it is determined whether the defective edge points are clustered on the Y-coordinate. When the number of defective edge points on the Y-coordinate is significantly larger, the defective edge points are clustered on the Y-coordinate. The number of defective edge points clustered on the Y-coordinate is greater than the number of defective edge points not clustered on the Y-coordinate, and the difference between the two is greater than a quantity threshold.

[0153] Step S110: Determine the display panel whose edge defect points are clustered on the X and / or Y coordinates as the target panel.

[0154] In this embodiment, the target panel with electrostatic discharge risk may further include a display panel whose edge defects are clustered on the X and / or Y coordinates. That is, the target panel can be the union of display panels whose defects are clustered on the X and / or Y coordinates and display panels whose edge defects are clustered on the X and / or Y coordinates.

[0155] In the above embodiments, firstly, a distribution map of edge-defective display panels in multiple motherboards is obtained. Based on the distribution map of edge-defective display panels in multiple motherboards, the high-incidence locations of edge defects in the motherboards are determined. Then, a distribution map of edge-defective points in the display panels at the high-incidence locations of edge defects in multiple motherboards is obtained. Based on the distribution map of edge-defective points in the display panels at the high-incidence locations of edge defects in multiple motherboards, it is determined whether the edge-defective points in the display panels are clustered on the X and / or Y coordinates. Finally, the display panels where the edge-defective points are clustered on the X and / or Y coordinates are also determined as target panels, which is beneficial for accurately identifying target panels with electrostatic discharge risk.

[0156] Figure 13 for Figure 12 Please refer to the flowchart for step S108. Figure 13For example, step S108 may include steps S701 to S702.

[0157] Step S701: Determine the edge area of ​​the display panel.

[0158] In this embodiment, the edge area of ​​the display panel is the area within a predetermined distance from the boundary line of the display panel. For example, the predetermined distance can be 0.5mm to 5mm, such as 0.5mm, 1mm, 1.5mm, 2mm, 2.5mm, 3mm, 3.5mm, 4mm, 4.5mm, 5mm, etc.

[0159] Figure 14 for Figure 13 Please refer to the schematic diagram of the position of the edge area of ​​the display panel in step S701. Figure 14 , Figure 14 The edge area is highlighted in the display panel, which is the area defined by a distance inward from the boundary line of the display panel.

[0160] Step S702: In the distribution map of defective points in the display panel, only the defective points in the edge area are retained to obtain the distribution map of edge defective points in the display panel.

[0161] In this embodiment, for the distribution map of defective points in the display panel, the defective points outside the edge area (i.e., non-edge area) are removed, and only the defective points in the edge area are retained, thus obtaining the distribution map of edge defective points in the display panel.

[0162] In the above embodiments, the edge area of ​​the display panel is first determined, and then only the defective points in the edge area are retained in the distribution map of defective points in the display panel to obtain the distribution map of defective points at the edge of the display panel. Thus, the distribution map of defective points at the edge of the display panel can be obtained from the distribution map of defective points in the display panel.

[0163] Figure 15 for Figure 13 Please refer to the flowchart of step S701. Figure 15 For example, step S701 may include steps S801 to S807.

[0164] Step S801: Based on the distribution map of defective points in the display panel, the convex hull algorithm is used to perform edge recognition to obtain the first polygonal outline of the display panel.

[0165] In this embodiment, the convex hull algorithm is a technique in computational geometry used to find the smallest convex polygon for a set of points. In two-dimensional space, the convex hull can be intuitively understood as one or more closed curves formed by stretching a taut rubber band around a set of nails (representing the set of points). If you imagine a plane with many nails, and you wrap all the nails around a rubber band, the shape formed by the rubber band is the convex hull. Mathematically, the convex hull is the smallest convex polygon that contains all points, ensuring that any line connecting any two points inside the polygon lies inside the polygon.

[0166] Figure 16 for Figure 15 Please refer to the schematic diagram of the first polygonal outline of the display panel in step S801. Figure 16 The first polygonal outline simply shows the outer contour of the display panel, without even showing the approximate positions of the corners.

[0167] Step S802: Use cross-shaped cutting lines to cut the first polygonal outline into four parts.

[0168] In this embodiment, the cross-shaped cutting lines penetrate the first polygonal contour in both the horizontal and vertical directions, and the horizontal and vertical penetration lines intersect within the first polygonal contour, thereby dividing the first polygonal contour into four parts.

[0169] For example, the center of the cross-shaped cutting line can be the geometric center of the first polygonal profile.

[0170] In step S803, the four parts are superimposed according to a symmetrical relationship, and the edge recognition is performed using the convex hull algorithm to obtain the second polygonal outline of the display panel.

[0171] In this embodiment, if the first polygonal outline includes a right upper part, a left upper part, a left lower part, and a right lower part that are connected end to end, that is, the four parts are the right upper part, the left upper part, the left lower part, and the right lower part, respectively, and the right upper part and the left upper part, the left lower part and the right lower part are cut by a vertical penetrating line, and the left upper part and the left lower part, the right lower part and the right upper part are cut by a horizontal penetrating line, then theoretically the right upper part and the left upper part are symmetrical about the vertical penetrating line, the left upper part and the left lower part are symmetrical about the horizontal penetrating line, the left lower part and the right lower part are symmetrical about the vertical penetrating line, and the right lower part and the right upper part are symmetrical about the horizontal penetrating line. This is the symmetrical relationship between the four parts. Therefore, when the upper right and upper left parts are superimposed, the sides of the upper right part that are close to the upper left part and the sides of the upper left part that are close to the upper right part coincide; when the upper left and lower left parts are superimposed, the sides of the upper left part that are close to the lower left part and the sides of the lower left part that are close to the upper left part coincide; when the lower left and lower right parts are superimposed, the sides of the lower left part that are close to the lower right part and the sides of the lower right part that are close to the lower left part coincide; when the lower right and upper right parts are superimposed, the sides of the lower right part that are close to the upper right part and the sides of the upper right part that are close to the lower right part coincide.

[0172] After superimposing the four parts, the corner points of the four parts can be superimposed together. Then, the convex hull algorithm is used for edge recognition to obtain the second polygonal outline of the display panel, and the corners can be kept consistent.

[0173] Figure 17 for Figure 15 Please refer to the schematic diagram of the second polygonal outline of the display panel in step S803. Figure 17 The second polygonal outline is symmetrical about the longitudinal and transverse through lines, respectively.

[0174] In step S804, a sliding window is used to slide along the X-coordinate direction on the second polygonal contour, and the maximum value of the Y-coordinate is continuously updated.

[0175] Step S805: If the maximum value of the Y coordinate has not been updated when the sliding window has been slid a set number of times, then the position where the Y coordinate of the sliding window was first not updated is determined as the corner dividing point, and the third polygonal outline of the display panel is obtained.

[0176] In this embodiment, a sliding window is first used to slide along the X-coordinate direction on the second polygonal outline, and the maximum value of the Y-coordinate is continuously updated until the maximum value of the Y-coordinate has not been updated after the sliding window has slid a set number of times. The position where the sliding window's Y-coordinate has not been updated the earliest can be determined as the corner boundary point, thus obtaining the third polygonal outline of the display panel. Therefore, the third polygonal outline of the display panel can initially determine the corners of the display panel.

[0177] Figure 18 for Figure 15 Please refer to the schematic diagram of the third polygonal outline of the display panel in step S805. Figure 18 The corners have appeared in the third polygonal outline.

[0178] In one possible embodiment, a sliding window can be used to slide along the X-coordinate direction on the second polygonal outline, and the maximum value of the Y-coordinate can be continuously updated. If the maximum value of the Y-coordinate is not updated after the sliding window has slid a set number of times, the position where the sliding window was first not updated in the Y-coordinate is determined as the two corner boundary points of the same X-coordinate. Then, based on the symmetry of the second polygonal outline in the X-coordinate, the two corner boundary points of another X-coordinate are obtained, and thus the third polygonal outline of the display panel is obtained.

[0179] In another possible embodiment, a sliding window can first be used to slide along the X-coordinate direction on the second polygonal outline, and the maximum value of the Y-coordinate can be continuously updated. If the maximum value of the Y-coordinate has not been updated after the sliding window has slid a set number of times, then the position where the sliding window has not been updated in the Y-coordinate is determined as the two corner dividing points of the same X-coordinate. Then, the sliding window can be used to slide along the opposite direction of the X-coordinate direction on the second polygonal outline, and the maximum value of the Y-coordinate can be continuously updated. If the maximum value of the Y-coordinate has not been updated after the sliding window has slid a set number of times, then the position where the sliding window has not been updated in the Y-coordinate is determined as the two corner dividing points of another X-coordinate, thereby obtaining the third polygonal outline of the display panel.

[0180] In practical applications, a sliding window can be used to slide along the Y-axis on the second polygonal outline and continuously update the maximum value of the X-coordinate. If the maximum value of the X-coordinate is not updated after the sliding window has slid a set number of times, the position where the sliding window has not updated the X-coordinate the earliest is determined as the corner dividing point, thus obtaining the third polygonal outline of the display panel.

[0181] Step S806: Determine the major axis of the ellipse based on the corner boundary points of the third polygonal contour, fit the ellipse, and form the fourth polygonal contour of the display panel.

[0182] Figure 19 for Figure 15 Please refer to the schematic diagram of the fourth polygonal outline of the display panel in step S806. Figure 19 The corners of the fourth polygonal outline become more like actual ellipses.

[0183] Step S807: The area with a set distance inward from the fourth polygonal outline is used as the edge area of ​​the display panel.

[0184] If the convex hull algorithm is directly applied to the corner recognition of the display panel, there are two problems: First, the symmetry between the four corners of the display panel is ignored, and separate recognition may result in a large deviation due to the small number of defective points; second, when there are few defective points, the convex hull algorithm is not accurate enough in recognizing the corner boundary points. In the above embodiment, firstly, based on the distribution map of defective points in the display panel, the convex hull algorithm is used for edge recognition to obtain the first polygonal contour of the display panel. Then, a cross-shaped cutting line is used to cut the first polygonal contour into four parts. The four parts are superimposed according to a symmetrical relationship, and the convex hull algorithm is used for edge recognition to obtain the second polygonal contour of the display panel. Then, a sliding window is used to slide along the X-coordinate direction on the second polygonal contour, and the maximum value of the Y-coordinate is continuously updated. If the maximum value of the Y-coordinate has not been updated within a set number of sliding window movements, the position where the sliding window has not been updated in the Y-coordinate at the earliest point is determined as the corner boundary point to obtain the third polygonal contour of the display panel. Next, the major axis of the ellipse is determined based on the corner boundary point of the third polygonal contour, and the ellipse is fitted to form the fourth polygonal contour of the display panel. Finally, the area with a set distance inward from the fourth polygonal contour is taken as the edge area of ​​the display panel. In this way, the defective point data of the four corners can be fully utilized, and the edge area of ​​the display panel can be determined more accurately.

[0185] For example, step S106 may include: obtaining a distribution map of edge defect points in each display panel of the motherboard; and generating a distribution map of edge defect display panels in the motherboard based on the distribution map of edge defect points in each display panel of the motherboard.

[0186] For example, step S107 may include: counting the number of edge defective display panels at the same location in multiple motherboards; inputting the counted number of edge defective display panels at each location in the motherboard into an aggregation identification model to determine the high-incidence locations of edge defects in the motherboard.

[0187] For example, step S109 may include: if the high incidence of edge defects in the motherboard is a single display panel, then the sum of the number of edge defect points at the same point in the display panels at the same location in multiple motherboards is counted; if the high incidence of edge defects in the motherboard is a row of display panels, then the sum of the number of edge defect points at the same point in the display panels at the same row in multiple motherboards is counted; if the high incidence of edge defects in the motherboard is a column of display panels, then the sum of the number of edge defect points at the same point in the display panels at the same column in multiple motherboards is counted; based on the sum of the number of edge defect points counted at each point in the display panel, it is determined whether the edge defect points in the display panel are clustered in the X coordinate and / or Y coordinate.

[0188] For example, the display panel can be divided into multiple first regions arranged along the X-coordinate direction or multiple second regions arranged along the Y-coordinate direction, with the same number of points in each first region and the same number of points in each second region.

[0189] Accordingly, determining whether edge defective points in the display panel are clustered in the X and / or Y coordinates based on the sum of the statistical counts of edge defective points at each point in the display panel can include: counting the number of edge defective points in each first region and each second region of the display panel based on the sum of the statistical counts of edge defective points at each point in the display panel; inputting the statistical counts of edge defective points in each first region of the display panel into the clustering recognition model to select the first region where edge defective points are clustered; if the clustering recognition model selects the first region where edge defective points are clustered, then it is determined that the edge defective points in the display panel are clustered in the Y coordinate; inputting the statistical counts of edge defective points in each second region of the display panel into the clustering recognition model to select the second region where edge defective points are clustered; if the clustering recognition model selects the second region where edge defective points are clustered, then it is determined that the edge defective points in the display panel are clustered in the X coordinate.

[0190] For example, determining whether edge defective points in the display panel are clustered in the X and / or Y coordinates based on the sum of the number of edge defective points counted at each point in the display panel may further include: displaying the number of edge defective points in a first region where edge defective points are clustered and the number of edge defective points in a second region where edge defective points are clustered to the user; in response to the first region and the second region selected by the user, marking the first region and the second region selected by the user as not clustered edge defective points, and optimizing the clustering identification model.

[0191] Figure 20 for Figure 1 For another flowchart of step S105, please refer to [link / reference]. Figure 20 Step S105 may include steps S901 to S903.

[0192] Step S901: Based on the defect distribution map of each target panel, count the sum of the number of defective points at the same location in all target panels.

[0193] In this embodiment, defective points at the same location in each target panel are superimposed to obtain the sum of the number of defective points at the same location.

[0194] Step S902: Based on the sum of the number of defective points counted at each point in the target panel, determine the high-incidence area of ​​defects in the target panel.

[0195] In this embodiment, the high-defect area in the target panel is a cluster of defective points in the target panel where the sum of the number of defective points is significantly large. For example, the high-defect area can be a closed shape, such as a rectangle.

[0196] For example, step S902 may include: inputting the sum of the number of defective points counted at each point in the target panel into the aggregation identification model to obtain the high-incidence points of defects in the target panel; and determining the high-incidence areas of defects in the target panel based on the high-incidence points of defects in the target panel.

[0197] Step S903: The target panel whose defect points are located outside the high-incidence area of ​​defects is redefined as a non-target panel.

[0198] In this embodiment, the target panel whose defective points are located outside the high-incidence area is changed to a non-target panel, while the target panel whose defective points are located within the high-incidence area remains a target panel.

[0199] In the above embodiments, firstly, based on the defect distribution map of each target panel, the sum of the number of defective points at the same point in all target panels is counted. Then, based on the sum of the number of defective points counted for each point in the target panel, the high-incidence area of ​​defects in the target panel is determined. Finally, the target panels whose defective points are located outside the high-incidence area of ​​defects are redefined as non-target panels, which is beneficial for excluding display panels with non-ESD defects.

[0200] Figure 21 for Figure 1 For another flowchart of step S105, please refer to [link / reference]. Figure 21 Step S105 may include steps S1001 to S1003.

[0201] Step S1001: Obtain the production equipment that each target panel has been through.

[0202] Step S1002: Based on the production equipment that each target panel has been through, determine the target equipment that each target panel has been through.

[0203] Step S1003: If there is contact between each target panel and the target device, then the target device is identified as a risk device.

[0204] In the above embodiments, the production equipment that each target panel has been through is first obtained, and then the target equipment that each target panel has been through is determined based on the production equipment that each target panel has been through. Finally, if each target panel has contact with the target equipment, the target equipment is identified as a risk equipment, thereby identifying the production equipment that multiple target panels have been in contact with as a risk equipment that may cause electrostatic discharge.

[0205] The target panels are concentrated on the production equipment, and it cannot be confirmed that the defects were caused by electrostatic discharge (ESD). Since ESD requires contact with the production equipment, the probability of ESD is particularly high if areas with high defect rates have had contact with the equipment. To address this, contact matching was added when locating potentially hazardous equipment to further reduce false alarms.

[0206] There are two scenarios for ESD defects: one where the production equipment simply makes contact with the target panel, and the other where the production equipment must make contact with a high-risk area for defects. These two scenarios address different types of ESD defects. Some ESD defects occur simply when the production equipment comes into contact with the display panel, and the defective point may not be at the point of contact; conversely, some defective points are located at the point of contact. If the contact rate between the production equipment and the high-risk area is very high, it can generally be considered an ESD defect. If the production equipment only has a high contact rate with the display panel, but a low contact rate with the high-risk area, it can be considered for investigation.

[0207] There are three scenarios for contact matching. First, the production equipment and the target panel are not in contact, which can be confirmed as a non-ESD defect. Second, the production equipment is in contact with the target panel, but not with areas prone to defects. In this case, the contact point may differ from the defect point, requiring further confirmation. Third, the production equipment is in contact with areas prone to defects, which is highly likely to be an ESD defect and can also be further verified.

[0208] Figure 22 for Figure 21 Please refer to the flowchart for step S1003. Figure 22 Step S1003 may include steps S1101 to S1102.

[0209] Step S1101: Based on the defect distribution map of the points in the target panel, determine the distribution area of ​​the defect points in the target panel.

[0210] Step S1102: If there is contact between each target panel and the target device, and the contact area coincides with the distribution area of ​​the defective points in the target panel, then the target device is identified as a risk device.

[0211] In the above embodiments, based on the defect distribution map of the points in the target panel, the distribution area of ​​the defect points in the target panel is determined. If each target panel is in contact with the target device and the contact area coincides with the distribution area of ​​the defect points in the target panel, then the target device is identified as a risk device, and production equipment unrelated to ESD defects can be excluded.

[0212] Figure 23 For a structural block diagram of the apparatus for detecting electrostatic discharge in one or more embodiments of this application, please refer to [link / reference]. Figure 23 According to a second aspect of this application, an apparatus for detecting electrostatic discharge damage is provided. The apparatus includes a first acquisition module 1201, a position determination module 1202, a second acquisition module 12003, an aggregation judgment module 1204, and an electrostatic discharge determination module 1205.

[0213] The first acquisition module 1201 is configured to acquire a distribution map of defective display panels among multiple motherboards. The motherboards include multiple display panels, which are cut from the motherboards.

[0214] The location determination module 1202 is configured to determine the high-incidence locations of defects in the motherboard based on the distribution map of defective display panels in multiple motherboards.

[0215] The second acquisition module 1203 is configured to acquire a distribution map of defective points in the display panel of multiple motherboards at locations with high defect incidence.

[0216] The clustering judgment module 1204 is configured to determine whether the defective points in the display panel are clustered in the X and / or Y coordinates based on the distribution map of defective points in high-frequency defect locations across multiple motherboards. Each point on the display panel includes different X and / or Y coordinates.

[0217] The electrostatic discharge determination module 1205 is configured to identify display panels with defective points clustered on the X and / or Y coordinates as target panels with a risk of electrostatic discharge, and to identify production equipment that has been in contact with multiple target panels as equipment at risk of electrostatic discharge.

[0218] The above-mentioned device for detecting electrostatic discharge In some embodiments, the first acquisition module 1201 may include an acquisition unit and a generation unit. The acquisition unit is configured to acquire a distribution map of defective points in each display panel of the motherboard. The generation unit is configured to generate a distribution map of defective display panels in the motherboard based on the distribution map of defective points in each display panel of the motherboard. Defective display panels in the motherboard are display panels in the motherboard that have defective points.

[0219] For example, the location determination module 1202 may include a first statistics unit and a location determination unit. The first statistics unit is configured to count the number of defective display panels at the same location in multiple motherboards. The location determination unit is configured to input the counted number of defective display panels at each location in the motherboards into an aggregation identification model to determine the high-incidence locations of defects in the motherboards. The number of defective display panels at the same high-incidence location in multiple motherboards is greater than the number of defective display panels at the same non-high-incidence location in multiple motherboards, and the difference between the two is greater than a quantity threshold.

[0220] In some embodiments, the aggregation determination module 1204 may include a second statistics unit and an aggregation determination unit. The second statistics unit is configured to: if the high-frequency defect location in the motherboard is a single display panel, then sum the number of defective points at the same location in display panels across multiple motherboards; if the high-frequency defect location in the motherboard is a row of display panels, then sum the number of defective points at the same location in display panels across multiple motherboards; if the high-frequency defect location in the motherboard is a column of display panels, then sum the number of defective points at the same location in display panels across multiple motherboards. The aggregation determination unit is configured to determine whether the defective points in the display panel are clustered in the X and / or Y coordinates based on the sum of the number of defective points counted for each point in the display panel.

[0221] For example, the display panel can be divided into multiple first regions arranged along the X-coordinate direction or multiple second regions arranged along the Y-coordinate direction, with the same number of points in each first region and the same number of points in each second region.

[0222] Accordingly, the clustering judgment unit can be configured to, based on the sum of the number of defective points counted in each point on the display panel, count the number of defective points in each first area and the number of defective points in each second area of ​​the display panel. The count of defective points in each first area of ​​the display panel is input into the clustering recognition model to select the first area where defective points cluster. The number of defective points in the first area where defective points cluster is greater than the number of defective points in the first area where defective points are not clustered, and the difference between the two is greater than a quantity threshold. If the clustering recognition model selects the first area where defective points cluster, it is determined that the defective points in the display panel are clustered on the Y-coordinate. The count of defective points in each second area of ​​the display panel is input into the clustering recognition model to select the second area where defective points cluster. The number of defective points in the second area where defective points cluster is greater than the number of defective points in the second area where defective points are not clustered, and the difference between the two is greater than a quantity threshold. If the clustering recognition model selects the second area where defective points cluster, it is determined that the defective points in the display panel are clustered on the X-coordinate.

[0223] For example, the number of points in the first area and the number of points in the second area can be negatively correlated with the sum of the number of defective points counted for each point on the display panel.

[0224] For example, the clustering judgment unit can also be configured to display to the user the number of defective points in a first region where defective points are clustered and the number of defective points in a second region where defective points are clustered. In response to the first and second regions selected by the user, the first and second regions selected by the user are marked as defective points not clustered, and the input clustering recognition model is optimized.

[0225] For example, the electrostatic determination module 1205 may include a determination unit, a similarity determination unit, and a first determination unit. The determination unit is configured to determine the display panel as a target panel if it determines that defective points in the display panel are clustered in the X and / or Y coordinates; and to determine the display panel as a non-target panel if it determines that defective points in the display panel are not clustered in either the X or Y coordinates. The similarity determination unit is configured to determine the similarity between the distribution positions of defective points in the target panel and the distribution positions of defective points in the non-target panel in the X and / or Y coordinates, based on the defect distribution map of points in the target panel and the defect distribution map of points in the non-target panel. The re-determination unit is configured to re-determine target panels with similarity higher than a similarity threshold as non-target panels.

[0226] In some embodiments, the device may further include a third acquisition module, a high-incidence determination module, a fourth acquisition module, an edge judgment module, and an edge determination module. The third acquisition module is configured to acquire a distribution map of edge-defective display panels among multiple motherboards. The high-incidence determination module is configured to determine high-incidence locations of edge defects in the motherboards based on the distribution map of edge-defective display panels among the multiple motherboards. The fourth acquisition module is configured to acquire a distribution map of edge-defect points in the display panels at high-incidence locations of edge defects among the multiple motherboards. The edge judgment module is configured to determine whether edge-defect points in the display panels are clustered in the X and / or Y coordinates based on the distribution map of edge-defect points in the display panels at high-incidence locations of edge defects among the multiple motherboards. The edge determination module is configured to identify display panels where edge-defect points are clustered in the X and / or Y coordinates as target panels.

[0227] For example, the fourth acquisition module may include an edge determination unit and a distribution map acquisition unit. The edge determination unit is configured to determine the edge region of the display panel. The distribution map acquisition unit is configured to retain only the defective points within the edge region in the distribution map of defective points in the display panel, thereby obtaining a distribution map of edge defective points in the display panel.

[0228] For example, the edge determination unit is configured to perform edge recognition using a convex hull algorithm based on the distribution map of defective points in the display panel, to obtain the first polygonal contour of the display panel; to cut the first polygonal contour into four parts using a cross-shaped cutting line; to superimpose the four parts according to a symmetrical relationship and perform edge recognition using a convex hull algorithm, to obtain the second polygonal contour of the display panel; to slide a sliding window along the X-coordinate direction on the second polygonal contour and continuously update the maximum value of the Y-coordinate; if the maximum value of the Y-coordinate has not been updated for a set number of sliding steps, the position where the sliding window has not been updated in the Y-coordinate is determined as the corner boundary point, to obtain the third polygonal contour of the display panel; to determine the major axis of an ellipse based on the corner boundary point of the third polygonal contour, to fit an ellipse, to form the fourth polygonal contour of the display panel; and to use the area at a set distance inward from the fourth polygonal contour as the edge area of ​​the display panel.

[0229] In some embodiments, the electrostatic discharge determination module 1205 may include a third statistical unit, a region determination unit, and a second determination unit. The third statistical unit is configured to sum the number of defective points at the same location across all target panels based on the defect distribution map of points in each target panel. The region determination unit is configured to determine high-incidence areas of defects in the target panels based on the sum of the number of defective points calculated for each point in the target panels. The second determination unit is configured to re-determine target panels whose defective points are located outside of high-incidence areas as non-target panels.

[0230] In some embodiments, the electrostatic discharge determination module 1205 may include a device acquisition unit, a device determination unit, and a risk determination unit. The device acquisition unit is configured to acquire the production equipment that each target panel has been through. The device determination unit is configured to determine target equipment that all target panels have been through, based on the production equipment that each target panel has been through. The risk determination unit is configured to determine a target equipment as a risk device if there is contact between each target panel and a target equipment.

[0231] For example, the risk determination unit can be configured to determine the distribution area of ​​defective points in the target panel based on the defect distribution map of the points in the target panel; if each target panel is in contact with the target device and the contact area coincides with the distribution area of ​​defective points in the target panel, then the target device is determined as a risk device.

[0232] Each module in the aforementioned device for detecting electrostatic discharge can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0233] A third aspect of this application provides a computer device, which may be a server, and its internal structure diagram may be as follows: Figure 24 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores relevant data during the training of an artificial intelligence model. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements an artificial intelligence model training method.

[0234] Those skilled in the art will understand that Figure 24 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0235] A third aspect of this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0236] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0237] A fifth aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0238] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units, etc., and are not limited to these.

[0239] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0240] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

[0241] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature being directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature being directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0242] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", and "counterclockwise" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0243] In this application, unless otherwise expressly specified and limited, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can mean a fixed connection, a detachable connection, or an integral part; it can mean a mechanical connection or an electrical connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0244] Furthermore, the use of terms such as "first" and "second" in this application is for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0245] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for detecting electrostatic discharge damage, characterized in that, The method includes: Obtain a distribution diagram of defective display panels in multiple motherboards; the motherboards include multiple display panels, and the display panels are cut from the motherboards; Based on the distribution map of defective display panels in the multiple motherboards, the locations of high-incidence defects in the motherboards are determined; Obtain a distribution map of defective points in the display panel at locations with high defect incidence in the multiple motherboards; Based on the distribution map of defective points in the display panel at high-incidence locations in the multiple motherboards, it is determined whether the defective points in the display panel are clustered in the X and / or Y coordinates; each point in the display panel includes different X and / or Y coordinates. Display panels with defective points clustered on the X and / or Y coordinates are identified as target panels with a risk of electrostatic discharge (ESD) damage, and production equipment that has been in contact with multiple of these target panels is identified as equipment at risk of ESD damage.

2. The method according to claim 1, characterized in that, The step of obtaining a distribution map of defective display panels in multiple master boards includes: Obtain a distribution map of defective points in each display panel of the motherboard; Based on the distribution map of defective points in each display panel in the master board, a distribution map of defective display panels in the master board is generated; the defective display panels in the master board are the display panels in the master board that have defective points.

3. The method according to claim 2, characterized in that, The step of determining the high-frequency locations of defects in the motherboards based on the distribution map of defective display panels in the multiple motherboards includes: Count the number of defective display panels at the same location in the multiple motherboards; The number of defective display panels at each location in the motherboard is input into the aggregation identification model to determine the high-incidence locations of defects in the motherboard; the number of defective display panels at the same high-incidence location in the multiple motherboards is greater than the number of defective display panels at the same non-high-incidence location in the multiple motherboards, and the difference between the two is greater than the number threshold.

4. The method according to claim 1, characterized in that, The determination of whether defective points in the display panel are clustered in the X and / or Y coordinates, based on the distribution map of defective locations in the multiple motherboards and the high-incidence locations of defects, includes: If the high-frequency defect location in the motherboard is a single display panel, then the sum of the number of defect points at the same point in the display panels at the same location in the multiple motherboards is counted. If the high-frequency defect locations in the motherboard are the same row of the display panels, then the sum of the number of defect locations at the same point in the same row of the display panels in the multiple motherboards is counted. If the high-frequency defect locations in the motherboard are the same column of the display panels, then the sum of the number of defect locations at the same point in the same column of the display panels in the multiple motherboards is counted. Based on the sum of the number of defective points counted at each point in the display panel, it is determined whether the defective points in the display panel are clustered in the X and / or Y coordinates.

5. The method according to claim 4, characterized in that, The display panel is divided into multiple first regions arranged along the X-coordinate direction or multiple second regions arranged along the Y-coordinate direction. The number of points in each first region is the same, and the number of points in each second region is the same. The step of determining whether defective points in the display panel are clustered along the X and / or Y coordinates based on the sum of the number of defective points counted from each point in the display panel includes: Based on the sum of the number of defective points counted at each point in the display panel, the number of defective points in each first area and the number of defective points in each second area of ​​the display panel are counted. The number of defective points in each of the first areas of the display panel is input into the clustering identification model to select the first area where defective points are clustered; the number of defective points in the first area where defective points are clustered is greater than the number of defective points in the first area where defective points are not clustered, and the difference between the two is greater than the number threshold. If the clustering identification model selects the first area where defective points are clustered, then it is determined that the defective points in the display panel are clustered on the Y coordinate. The number of defective points counted in each of the second areas of the display panel is input into the clustering identification model to select the second area where defective points are clustered; the number of defective points in the second area where defective points are clustered is greater than the number of defective points in the second area where defective points are not clustered, and the difference between the two is greater than the number threshold. If the clustering identification model selects the second region where defective points are clustered, then it is determined that the defective points in the display panel are clustered on the X coordinate.

6. The method according to claim 5, characterized in that, The number of points in the first area and the number of points in the second area are negatively correlated with the sum of the number of defective points counted for each point in the display panel.

7. The method according to claim 5, characterized in that, The step of determining whether defective points in the display panel are clustered in the X and / or Y coordinates based on the sum of the number of defective points counted from each point in the display panel further includes: The system displays to the user the number of defective points in the first area where defective points are clustered and the number of defective points in the second area where defective points are clustered. In response to the first and second regions selected by the user, the first and second regions selected by the user are marked as non-aggregated bad points, and the aggregation identification model is optimized.

8. The method according to claim 5, characterized in that, The method of identifying display panels whose defective points are clustered on the X and / or Y coordinates as target panels with a risk of electrostatic discharge includes: If it is determined that the defective points in the display panel are clustered in the X and / or Y coordinates, then the display panel is identified as the target panel; If it is determined that the defective points in the display panel are not clustered on the X coordinate and not clustered on the Y coordinate, then the display panel is determined to be a non-target panel; Based on the defect distribution map of the points in the target panel and the defect distribution map of the points in the non-target panel, the similarity between the distribution positions of the defective points in the target panel and the distribution positions of the defective points in the non-target panel is determined. The target panels with similarity scores higher than the similarity threshold are reclassified as non-target panels.

9. The method according to any one of claims 1-8, characterized in that, The method further includes: Obtain a distribution map of the edge-defective display panels among the multiple motherboards; Based on the distribution map of the edge defect display panels in the multiple motherboards, the high-incidence locations of edge defects in the motherboards are determined; Obtain a distribution map of edge defect points in the display panel, corresponding to the locations with high incidence of edge defects among the multiple motherboards; Based on the distribution map of edge defect points in the display panel at high-incidence locations of edge defects in the multiple motherboards, determine whether the edge defect points in the display panel are clustered in the X and / or Y coordinates; The display panel whose edge defects are clustered on the X and / or Y coordinates is determined as the target panel.

10. The method according to claim 9, characterized in that, The step of obtaining a distribution map of edge defect points in the display panel, based on the high-incidence locations of defects in the multiple motherboards, includes: Determine the edge area of ​​the display panel; In the distribution map of defective points in the display panel, only the defective points in the edge area are retained, thus obtaining the distribution map of edge defective points in the display panel.

11. The method according to claim 10, characterized in that, Determining the edge region of the display panel includes: Based on the distribution map of defective points in the display panel, the convex hull algorithm is used for edge recognition to obtain the first polygonal outline of the display panel; The first polygonal outline is divided into four parts using cross-shaped cutting lines; The four parts are superimposed in a symmetrical relationship, and the edge recognition is performed using the convex hull algorithm to obtain the second polygonal outline of the display panel; A sliding window is used to slide along the X-coordinate direction on the contour of the second polygon, and the maximum value of the Y-coordinate is continuously updated; If the maximum value of the Y coordinate has not been updated when the sliding window has been slid a set number of times, then the position where the Y coordinate of the sliding window was first not updated is determined as the corner dividing point, and the third polygonal outline of the display panel is obtained. The major axis of the ellipse is determined based on the corner boundary points of the third polygonal contour, and the ellipse is fitted to form the fourth polygonal contour of the display panel. The area extending inward from the fourth polygonal outline is defined as the edge area of ​​the display panel.

12. The method according to any one of claims 1-8, characterized in that, The method of identifying display panels whose defective points are clustered on the X and / or Y coordinates as target panels with a risk of electrostatic discharge includes: Based on the defect distribution map of each target panel, the sum of the number of defective points at the same point in all the target panels is calculated. Based on the sum of the number of defective points counted at each point in the target panel, the high-incidence area of ​​defects in the target panel is determined. The target panel whose defect points are located outside the high-incidence defect area is redefined as a non-target panel.

13. The method according to any one of claims 1-8, characterized in that, The determination that production equipment that has been in contact with multiple target panels is a risk device for electrostatic discharge includes: Obtain the production equipment that each of the target panels has been manufactured on; Based on the production equipment that each of the target panels has been through, determine the target equipment that each of the target panels has been through; If each of the target panels is in contact with the target device, then the target device is identified as the risk device.

14. The method according to claim 13, characterized in that, If each of the target panels is in contact with the target device, then the target device is identified as the risk device, including: Based on the defect distribution map of the points in the target panel, the distribution area of ​​the defect points in the target panel is determined; If each of the target panels is in contact with the target device, and the contact area coincides with the distribution area of ​​the defective points in the target panel, then the target device is identified as the risk device.

15. A device for detecting electrostatic discharge damage, characterized in that, The device includes: The first acquisition module is configured to acquire a distribution map of defective display panels in multiple motherboards; the motherboards include multiple display panels, and the display panels are cut from the motherboards; The location determination module is configured to determine the high-frequency locations of defects in the motherboards based on the distribution map of defective display panels in the plurality of motherboards; The second acquisition module is configured to acquire a distribution map of defective points in the display panel at locations with high defect incidence in the plurality of motherboards; The clustering judgment module is configured to determine whether the defective points in the display panel are clustered in the X and / or Y coordinates based on the distribution map of defective points in the display panel at high defective locations in the multiple motherboards; each point in the display panel includes different X and / or Y coordinates. The electrostatic discharge (ESD) determination module is configured to identify display panels with defective points clustered on the X and / or Y coordinates as target panels with a risk of ESD damage, and to identify production equipment that has been in contact with multiple target panels as equipment at risk of ESD damage.

16. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-15.

17. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-15.

18. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-15.

Images