Crystal bonding method, device, equipment and storage medium

By obtaining the solid crystal surface image of the base, identifying the bracket image, calculating the edge shrinkage coefficient, and adjusting the mounting trajectory parameters, the chip mounting position deviation caused by the deformation of the large base is solved, and the electrical connection stability of the LED chip is improved and optical performance consistency.

CN118712296BActive Publication Date: 2024-11-08SHENZHEN JINGRUI PHOTOELECTRIC CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411198794.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-29
Publication Date
2024-11-08
Estimated Expiration
2044-08-29

AI Technical Summary

Technical Problem

In the large-scale production of LED chips, large bases are prone to deformation due to environmental factors during the manufacturing process, resulting in chip mounting position deviation, affecting the stability of electrical connection and consistency of optical performance.

Method used

By obtaining the solid crystal surface image of the base, identifying the bracket image, dividing it into multiple detection areas, using the edge detection network to extract edge features, calculate the edge shrinkage coefficient, and adjusting the mounting trajectory parameters to achieve accurate mounting and heating of the crystal.

Benefits of technology

Improves the electrical connection stability and optical performance consistency between the chip and the base, and enhances production efficiency and product quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118712296B_ABST
    Figure CN118712296B_ABST
Patent Text Reader

Abstract

The present application provides a die bonding method, device, equipment and storage medium for fixing a chip on a base, the method comprising: obtaining a bracket image of a die bonding surface image of a base; segmenting the die bonding surface image into a plurality of detection areas, and then dividing the detection area into a first sub-area and a second sub-area according to a plurality of preset second segmentation boundaries; extracting edge features of the first sub-area and the second sub-area, determining the maximum edge expansion and contraction coefficients of the first sub-area and the second sub-area, thereby determining a target segmentation boundary, and dividing the detection area into a first target sub-area and a second target sub-area according to the target segmentation boundary; determining a first correction parameter of the first target sub-area and a second correction parameter of the second target sub-area according to the maximum edge expansion and contraction coefficient; adjusting a preset trajectory parameter according to the first correction parameter or the second correction parameter to obtain a target parameter, and grabbing the chip to the base according to the target parameter for heating and die bonding.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of LED die bonding technology, and in particular to a die bonding method, device, equipment and storage medium. Background Art

[0002] In the modern optoelectronics industry, light-emitting diodes (LEDs) are core components that play a vital role in display technology, lighting systems, signal indication and other fields. The manufacturing process of LEDs involves multiple precision steps, among which chip mounting and die bonding processes are particularly critical and directly determine the optical performance and reliability of the product. LED chip mounting refers to the precise placement of tiny LED chips at designated positions on the base, followed by a die bonding process to securely fix them. This process requires extremely high precision, as any slight positional deviation will significantly affect the luminous efficiency and directionality of the LED. In the mass production of LED chips, using a large base for chip mounting and then dividing them into individual devices has the advantages of cost-effectiveness and production efficiency. However, the large base may be affected by environmental factors such as temperature changes, mechanical stress, material properties, etc. during the manufacturing process, and the base is prone to deformation. The base deformation will cause a deviation between the preset chip mounting position and the actual position, which will in turn affect the electrical connection stability and optical performance consistency between the chip and the base. Summary of the invention

[0003] The present application provides a die bonding method, device, equipment and storage medium for improving the electrical connection stability and optical performance consistency between a chip and a base during the die bonding process.

[0004] In a first aspect, an embodiment of the present application provides a die bonding method for fixing a chip on a base, the method comprising:

[0005] Acquire a crystal-fixing surface image of the base, and identify the support image on the crystal-fixing surface image;

[0006] Dividing the solid crystal surface image into a plurality of detection regions according to a preset first segmentation boundary, and dividing the detection region into a first sub-region and a second sub-region according to a plurality of preset second segmentation boundaries;

[0007] Extract edge features from the stent image in the first sub-region and the stent image in the second sub-region respectively through a preset edge detection network to obtain a first edge feature set and a second edge feature set;

[0008] Determine an edge expansion and contraction coefficient according to the first edge feature set and the second edge feature set, set the second segmentation boundary corresponding to the largest edge expansion and contraction coefficient as a target segmentation boundary, and divide the detection area into a first target sub-area and a second target sub-area according to the target segmentation boundary;

[0009] Determine a first correction parameter of the first target sub-region and a second correction parameter of the second target sub-region respectively according to the largest edge expansion and contraction coefficient;

[0010] The preset trajectory parameters for mounting the chip are adjusted according to the first correction parameter or the second correction parameter to obtain target parameters, and the chip is grabbed onto the base according to the target parameters to perform heating and die bonding.

[0011] In a second aspect, an embodiment of the present application provides a die-bonding device for fixing a chip on a base, the device comprising:

[0012] An image acquisition module, used to acquire an image of a crystal-fixing surface of the base and identify the image of the bracket on the crystal-fixing surface image;

[0013] Dividing the solid crystal surface image into a plurality of detection regions according to a preset first segmentation boundary, and dividing the detection region into a first sub-region and a second sub-region according to a plurality of preset second segmentation boundaries;

[0014] Extract edge features from the stent image in the first sub-region and the stent image in the second sub-region respectively through a preset edge detection network to obtain a first edge feature set and a second edge feature set;

[0015] Determine an edge expansion and contraction coefficient according to the first edge feature set and the second edge feature set, set the second segmentation boundary corresponding to the largest edge expansion and contraction coefficient as a target segmentation boundary, and divide the detection area into a first target sub-area and a second target sub-area according to the target segmentation boundary;

[0016] Determine a first correction parameter of the first target sub-region and a second correction parameter of the second target sub-region respectively according to the largest edge expansion and contraction coefficient;

[0017] The preset trajectory parameters for mounting the chip are adjusted according to the first correction parameter or the second correction parameter to obtain target parameters, and the chip is grabbed onto the base according to the target parameters to perform heating and die bonding.

[0018] In a third aspect, an embodiment of the present application provides an electronic device, the electronic device comprising a memory and a processor;

[0019] The memory is used to store computer programs;

[0020] The processor is used to execute the computer program and implement the die bonding method as described in any one of the embodiments of the present application when executing the computer program.

[0021] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor enables the processor to implement a die bonding method as described in any one of the embodiments of the present application.

[0022] The embodiment of the present application provides a die bonding method for fixing a chip on a base, the method comprising: acquiring a die bonding surface image of the base, identifying a support image on the die bonding surface image; dividing the die bonding surface image into a plurality of detection areas according to a preset first segmentation boundary, and dividing the detection area into a first sub-area and a second sub-area according to a plurality of preset second segmentation boundaries; performing edge feature extraction on the support image in the first sub-area and the support image in the second sub-area through a preset edge detection network, respectively, to obtain a first edge feature set and a second edge feature set; determining an edge expansion and contraction coefficient according to the first edge feature set and the second edge feature set, setting a second segmentation boundary corresponding to the maximum edge expansion and contraction coefficient as a target segmentation boundary, and dividing the detection area into a first target sub-area and a second target sub-area according to the target segmentation boundary; determining a first correction parameter for the first target sub-area and a second correction parameter for the second target sub-area according to the maximum edge expansion and contraction coefficient; adjusting a preset trajectory parameter for mounting the chip according to the first correction parameter or the second correction parameter to obtain a target parameter, grabbing the chip to the base according to the target parameter, and performing heating and die bonding. Through the above method, the solid crystal surface image is divided into multiple detection areas, and calibration is performed for each detection area, which is beneficial to improving the calibration accuracy. During calibration, the deformation direction of the base is unknown, so it is necessary to divide the detection area into a first sub-area and a second sub-area through multiple preset second segmentation boundaries, and the edge features of the bracket image in the first sub-area and the bracket image in the second sub-area are extracted through a preset edge detection network. After comparison, the edge expansion and contraction coefficient is obtained, and each second segmentation boundary corresponds to an edge expansion and contraction coefficient. The second segmentation boundary corresponding to the largest edge expansion and contraction coefficient is the target segmentation boundary, corresponding to the first target sub-area and the second target sub-area, and then the first correction parameter of the first target sub-area and the second correction parameter of the second target sub-area are determined according to the largest edge expansion and contraction coefficient. The preset trajectory parameters for mounting the chip are adjusted according to the first correction parameter or the second correction parameter to obtain the target parameters. After accurate mounting is achieved by performing separate trajectory correction on the sub-areas, heating and solid crystal bonding are performed to improve the electrical connection stability and optical performance consistency between the chip and the base. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0024] Figure 1 A schematic flow chart of a die bonding method provided in an embodiment of the present application;

[0025] Figure 2 A schematic diagram of segmentation of a detection area provided in an embodiment of the present application;

[0026] Figure 3 A schematic block diagram of a die bonding device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0029] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0030] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0031] It should be noted that in the embodiment of the present application, the chip is an LED chip, the base is the substrate material of the LED chip, and the base includes: a bracket, a pad, a heat sink, and a coordinate reference point. The bracket is the core part of the LED base, usually made of metal, such as copper, copper alloy or aluminum. The bracket provides a platform on which the LED chip can be fixed and provides an electrical connection point.

[0032] See also Figure 1 , Figure 1 The embodiment of the present application also provides a schematic flow chart of a die bonding method. Figure 1 As shown, the specific steps of the die bonding method include: S101-S106.

[0033] S101, acquiring a crystal-fixing surface image of a base, and identifying a support image on the crystal-fixing surface image.

[0034] Exemplarily, the crystal fixing surface of the base is a plane for mounting LED chips. In the process of detecting the deformation of the base, the bracket image of the bracket is obtained, and the edge features of the bracket image are compared and analyzed to detect whether there are differences between different brackets, thereby determining whether the base is deformed.

[0035] S102 , dividing the crystal-bonding surface image into a plurality of detection regions according to a preset first segmentation boundary, and dividing the detection region into a first sub-region and a second sub-region according to a plurality of preset second segmentation boundaries.

[0036] For example, since different regions have different degrees of deformation, the track offset cannot be uniformly set during the mounting process, otherwise, mounting errors may occur in non-deformed regions. The solid crystal surface image can be divided into multiple detection regions as needed, for example, the detection region is a 2cm×2cm square region.

[0037] Before knowing the deformation direction of the base, it is necessary to measure the deformation direction of the base. Usually, quality inspectors need to screen with horizontal plane tools and scrap the bases with excessive deformation, but this wastes manpower and material resources and increases production costs. The first sub-area and the second sub-area are divided by different schemes, and then the optimal division method is tested. The optimal division method can reflect the deformation of the base.

[0038] See also Figure 2 , Figure 2 A schematic diagram of segmenting a detection area provided in an embodiment of the present application is shown. Figure 2 As shown, in each segmentation example, the detection area is divided into a first sub-area and a second sub-area by the second segmentation boundary. It should be noted that Figure 2 The segmentation method provided in the embodiment is only used to illustrate the segmentation process, and is not used to limit the present application. In actual use, more segmentation examples can be set as needed.

[0039] S103, extracting edge features from the stent image in the first sub-region and the stent image in the second sub-region respectively through a preset edge detection network to obtain a first edge feature set and a second edge feature set.

[0040] Exemplarily, the edge detection network in the embodiment of the present application is any one of a fully convolutional network, a DeepEdge network or a RefineNet network. In the embodiment of the present application, the edge corner points of the bracket image are extracted by a trained edge detection network, and the size parameters of the edge corner points are obtained to generate an edge feature set. Each edge feature set generally includes a series of feature points, edge lines and contour information, which can be used to further analyze the shape, size and position of the bracket. For example, the feature points may correspond to the key points of the bracket, the edge lines represent the outer contour of the bracket, and the contour depicts the overall shape of the bracket. By extracting edge features from the bracket image, the position and posture of the bracket can be more accurately located, which helps to improve the accuracy of subsequent solid crystal operations. In addition, the edge feature set can also be used for quality control of the bracket, such as checking for defects or deformations.

[0041] S104, determining an edge expansion and contraction coefficient according to the first edge feature set and the second edge feature set, setting a second segmentation boundary corresponding to the maximum edge expansion and contraction coefficient as a target segmentation boundary, and dividing the detection area into a first target sub-area and a second target sub-area according to the target segmentation boundary.

[0042] In the embodiment of the present application, multiple segmentation methods (second segmentation boundaries) divide the detection area into a first sub-area and a second sub-area, and then compare the edge expansion and contraction coefficients detected in different segmentation methods. From the multiple second segmentation boundaries, the target segmentation boundary that best reflects the dividing line between the deformed area and the non-deformed area is determined, thereby further improving the correction accuracy for a detection area.

[0043] In each segmented first sub-region and second sub-region, edge detection technology is applied to calculate the edge expansion coefficient. The edge expansion coefficient reflects the degree of deviation of the image edge relative to the ideal state and can quantify the degree of deformation. By comparing the edge expansion coefficients detected under different segmentation methods, the segmentation method that can best distinguish the deformed area from the non-deformed area is evaluated as the basis for subsequent precise correction. This optimization method based on multiple segmentation methods can significantly improve the accuracy and efficiency of correction, especially when dealing with complex or irregular deformations.

[0044] S105 , determining a first correction parameter of the first target sub-region and a second correction parameter of the second target sub-region respectively according to the maximum edge expansion / contraction coefficient.

[0045] Exemplarily, the first correction parameter is the track offset and the placement force variable in the first target sub-area. For example, there is a placement left side in the first target sub-area, and the corresponding placement track is generated according to the placement coordinates. Only using the placement track for placement will produce a placement offset, but after correcting the placement track based on the track offset, the placement offset will be corrected. Similarly, the second correction parameter is the track offset and the placement force variable in the second target sub-area.

[0046] In some embodiments, the deformation of the base is unidirectional, for example, the left side does not deform. The right side deforms, and a deformation boundary appears. If a detection area includes the deformation boundary, in the most ideal case, the target segmentation boundary and the deformation boundary coincide, the first sub-area does not deform, and the second sub-area is a deformation area. Therefore, it is only necessary to determine the second correction parameter of the second target sub-area based on the maximum edge expansion and contraction coefficient.

[0047] In other embodiments, the first target sub-region and the second target sub-region are both deformation regions, and it is necessary to determine the first correction parameter of the first target sub-region and the second correction parameter of the second target sub-region respectively according to the maximum edge expansion and contraction coefficient.

[0048] By determining the correction parameters based on the maximum edge expansion and contraction coefficient, the placement parameters can be corrected in a targeted manner for the area of ​​the base where the deformation is most significant, thereby improving the placement accuracy.

[0049] S106, adjusting the preset trajectory parameters for chip mounting according to the first correction parameter or the second correction parameter to obtain target parameters, grabbing the chip to the base according to the target parameters, and performing heating and die bonding.

[0050] Exemplarily, the preset trajectory parameters are formulated based on the ideal conditions of the undeformed base, including the chip grabbing position, moving path, placement position, etc. In actual operation, these parameters need to be adjusted according to the deformation. The preset trajectory parameters related to the first target sub-area are adjusted according to the first correction parameters to ensure that the chip can be accurately placed in the correct position after deformation. Similarly, the preset trajectory parameters related to the second target sub-area are adjusted according to the second correction parameters to ensure the placement accuracy of all areas.

[0051] Through the above process, even when the base is deformed, the preset trajectory parameters can be adjusted to ensure accurate chip placement, thereby improving production efficiency and product quality.

[0052] The embodiment of the present application provides a die bonding method for fixing a chip on a base, the method comprising: acquiring a die bonding surface image of the base, identifying a support image on the die bonding surface image; dividing the die bonding surface image into a plurality of detection areas according to a preset first segmentation boundary, and dividing the detection area into a first sub-area and a second sub-area according to a plurality of preset second segmentation boundaries; performing edge feature extraction on the support image in the first sub-area and the support image in the second sub-area through a preset edge detection network, respectively, to obtain a first edge feature set and a second edge feature set; determining an edge expansion and contraction coefficient according to the first edge feature set and the second edge feature set, setting a second segmentation boundary corresponding to the maximum edge expansion and contraction coefficient as a target segmentation boundary, and dividing the detection area into a first target sub-area and a second target sub-area according to the target segmentation boundary; determining a first correction parameter for the first target sub-area and a second correction parameter for the second target sub-area according to the maximum edge expansion and contraction coefficient; adjusting a preset trajectory parameter for mounting the chip according to the first correction parameter or the second correction parameter to obtain a target parameter, grabbing the chip to the base according to the target parameter, and performing heating and die bonding. Through the above method, the solid crystal surface image is divided into multiple detection areas, and calibration is performed for each detection area, which is beneficial to improving the calibration accuracy. During calibration, the deformation direction of the base is unknown, so it is necessary to divide the detection area into a first sub-area and a second sub-area through multiple preset second segmentation boundaries, and the edge features of the bracket image in the first sub-area and the bracket image in the second sub-area are extracted through a preset edge detection network. After comparison, the edge expansion and contraction coefficient is obtained, and each second segmentation boundary corresponds to an edge expansion and contraction coefficient. The second segmentation boundary corresponding to the largest edge expansion and contraction coefficient is the target segmentation boundary, corresponding to the first target sub-area and the second target sub-area, and then the first correction parameter of the first target sub-area and the second correction parameter of the second target sub-area are determined according to the largest edge expansion and contraction coefficient. The preset trajectory parameters for mounting the chip are adjusted according to the first correction parameter or the second correction parameter to obtain the target parameters. After accurate mounting is achieved by performing separate trajectory correction on the sub-areas, heating and solid crystal bonding are performed to improve the electrical connection stability and optical performance consistency between the chip and the base.

[0053] In order to more clearly introduce the technical solution of the present application, the technical solution of the present application will be introduced through specific embodiments below. It should be noted that the specific embodiments are used to expand the technical solution of the present application, but are not intended to limit the present application.

[0054] In some embodiments, edge features of the support image in the first sub-region and the support image in the second sub-region are respectively extracted through a preset edge detection network to obtain a first edge feature set and a second edge feature set, including: extracting features of the support image in the first sub-region through the edge detection network, determining the first edge corner point, and obtaining the first size parameter of the first edge corner point, and generating the first edge feature set according to the first size parameter; extracting features of the support image in the second sub-region through the edge detection network, determining the second edge corner point, and obtaining the second size parameter of the second edge corner point, and generating the second edge feature set according to the second size parameter.

[0055] Exemplarily, specific steps for processing the first sub-region: S1031-S1032.

[0056] S1031. Feature extraction.

[0057] The support image in the first sub-region is input to the edge detection network, which outputs a feature map containing edge information, where high response values ​​correspond to significant edges in the image.

[0058] S1032: Determine a first edge corner point.

[0059] The edge detection network analyzes the feature map and locates the points with the highest local intensity, which are the edge corners. Corners are usually the intersection of multiple edges and are very important for shape recognition.

[0060] S1033. Obtain a first size parameter.

[0061] The edge detection network measures the edge length, angle or other size-related properties near each first edge corner point, and these measurement results constitute the first size parameter.

[0062] S1034: Generate a first edge feature set.

[0063] The edge detection network integrates the position coordinates of all first edge corner points and the corresponding first size parameters to form a first edge feature set, which fully describes the edge features of the stent image in the first sub-region.

[0064] Similarly, the second sub-region is processed in the same manner.

[0065] Through the above process, the key edges and corners in the bracket image are identified, and the size parameters describing these features can be quantified. These edge feature sets are used for subsequent deformation analysis, pose estimation and precise placement.

[0066] In some embodiments, the edge expansion and shrinkage coefficient is determined according to the first edge feature set and the second edge feature set, including: performing standard deviation calculation on the first size parameter to obtain the first expansion and shrinkage standard deviation, performing standard deviation calculation on the second size parameter to obtain the second expansion and shrinkage standard deviation; if the first expansion and shrinkage standard deviation is greater than the second expansion and shrinkage standard deviation, setting the area center of the first sub-area as the direction starting point, setting the area center of the second sub-area as the direction end point, and determining the expansion and shrinkage direction according to the direction starting point and the direction end point; based on the expansion and shrinkage direction, determining multiple groups of target size parameters from the first size parameter and the second size parameter, the target size parameters being two size parameters adjacent in the expansion and shrinkage direction; based on the expansion and shrinkage direction, performing difference calculation on the target size parameters to obtain the inch-size expansion and shrinkage coefficient; calculating the average value of the size expansion and shrinkage coefficient to obtain the edge expansion and shrinkage coefficient.

[0067] For example, the expansion and shrinkage standard deviation reflects the overall deformation degree in a sub-region. If there are deformed regions and non-deformed regions in a detection area, the sub-region with a small expansion and shrinkage standard deviation is the non-deformed region, and the sub-region with a large expansion and shrinkage standard deviation is the deformed region, which can determine the expansion and shrinkage direction (deformation direction). In the area where deformation occurs, following the expansion and shrinkage direction, the expansion and shrinkage size between each group of target size parameters should be roughly the same, so the average value of the size expansion and shrinkage coefficient is calculated to obtain the inch-size expansion and shrinkage coefficient.

[0068] It should be noted that if the inch-size expansion and contraction coefficient is less than the preset expansion and contraction threshold, it will be set to 0, which helps to distinguish the deformation area from the non-deformation area. The inch-size expansion and contraction coefficient set to 0 does not participate in the average value calculation.

[0069] In some embodiments, a first correction parameter of the first target sub-area and a second correction parameter of the second target sub-area are respectively determined according to the maximum edge expansion and contraction coefficient, including: generating a grab offset parameter according to the edge expansion and contraction coefficient and a preset expansion and contraction offset function; using the grab offset parameter as the second correction parameter; if the first expansion and contraction standard deviation is less than a preset first standard threshold, removing the first correction parameter; if the first expansion and contraction standard deviation is greater than or equal to the preset standard threshold, using the grab offset parameter as the first correction parameter.

[0070] Exemplarily, the grasping offset parameter is calculated based on the edge expansion coefficient and the preset expansion offset function, and is used to indicate how the robot grasper should adjust its position and posture relative to the initially calculated grasping point. It can be a vector that includes the offset in the X and Y directions and the rotation angle.

[0071] In the embodiment of the present application, the second target sub-region is a region with deformation, and therefore, it is necessary to grab the offset parameter for correction, that is, the grab offset parameter is used as the second correction parameter.

[0072] In one embodiment, if the first expansion and contraction standard deviation is less than the preset first standard threshold, there is no deformation in the first target sub-region, and a large correction may not be required. In this case, the first correction parameter will be removed, and the grasping operation will be mainly based on the original grasping point position. If the first expansion and contraction standard deviation is greater than or equal to the preset first standard threshold, it means that there is deformation in the first target sub-region, and the grasping offset parameter will be used as the first correction parameter to significantly adjust the grasping point position to ensure that the robot can grasp the object more accurately.

[0073] Through such a mechanism, the robotic system can dynamically adjust its grasping strategy under different image quality and environmental conditions, thereby improving the robustness and efficiency of the operation.

[0074] In some embodiments, the base includes a coordinate reference point. Before adjusting the preset trajectory parameters for mounting the chip according to the first correction parameter or the second correction parameter to obtain the target parameter, the method also includes: generating mounting coordinates according to the preset base layout information and the coordinate reference point; generating initial trajectory parameters according to the coordinate reference point and the mounting coordinates; if the mounting coordinates are in the first target sub-area, correcting the initial trajectory parameters according to the first correction parameter to obtain the preset trajectory parameters; if the mounting coordinates are in the second target sub-area, correcting the initial trajectory parameters according to the second correction parameter to obtain the preset trajectory parameters.

[0075] In some embodiments, obtaining a crystal-fixing surface image of a base and identifying a support image on the crystal-fixing surface image include: obtaining a crystal-fixing surface image of a base, performing support identification on the crystal-fixing surface image, and obtaining a support image in the crystal-fixing surface image.

[0076] In some embodiments, the preset edge detection network includes: a fully convolutional network, a DeepEdge network, and a RefineNet network.

[0077] See also Figure 3 , Figure 3 The embodiment of the present application further provides a schematic block diagram of a die-bonding device, wherein the die-bonding device 300 is used to perform the above-mentioned die-bonding method. The die-bonding device 300 can be configured in a server.

[0078] Among them, the server can be an independent server or a server cluster, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN) and big data and artificial intelligence platforms.

[0079] like Figure 3As shown, the die bonding device 300 includes: an image acquisition module 301 , a region segmentation module 302 , an edge extraction module 303 , a region determination module 304 , a parameter generation module 305 and a mounting management module 306 .

[0080] The image acquisition module 301 is used to acquire the crystal-fixing surface image of the base and identify the support image on the crystal-fixing surface image.

[0081] In some embodiments, when the image acquisition module 301 is used to acquire the solid crystal surface image of the base and identify the support image on the solid crystal surface image, it is specifically used to achieve: acquire the solid crystal surface image of the base, perform support identification on the solid crystal surface image, and obtain the support image in the solid crystal surface image.

[0082] The region segmentation module 302 segments the solid crystal surface image into a plurality of detection regions according to a preset first segmentation boundary, and divides the detection region into a first sub-region and a second sub-region according to a plurality of preset second segmentation boundaries.

[0083] The edge extraction module 303 extracts edge features from the stent image in the first sub-region and the stent image in the second sub-region respectively through a preset edge detection network to obtain a first edge feature set and a second edge feature set.

[0084] In some embodiments, when the edge extraction module 303 is used to implement edge feature extraction of the support image in the first sub-region and the support image in the second sub-region through a preset edge detection network to obtain a first edge feature set and a second edge feature set, it is specifically used to implement: feature extraction of the support image in the first sub-region through the edge detection network, determining the first edge corner point, and obtaining the first size parameter of the first edge corner point, and generating a first edge feature set according to the first size parameter; feature extraction of the support image in the second sub-region through the edge detection network, determining the second edge corner point, and obtaining the second size parameter of the second edge corner point, and generating a second edge feature set according to the second size parameter.

[0085] The region determination module 304 determines the edge expansion and contraction coefficient according to the first edge feature set and the second edge feature set, sets the second segmentation boundary corresponding to the maximum edge expansion and contraction coefficient as the target segmentation boundary, and divides the detection region into the first target sub-region and the second target sub-region according to the target segmentation boundary.

[0086] In some embodiments, when the area determination module 304 is used to determine the edge expansion and shrinkage coefficient according to the first edge feature set and the second edge feature set, it is specifically used to implement: performing standard deviation calculation on the first size parameter to obtain the first expansion and shrinkage standard deviation, performing standard deviation calculation on the second size parameter to obtain the second expansion and shrinkage standard deviation; if the first expansion and shrinkage standard deviation is greater than the second expansion and shrinkage standard deviation, setting the area center of the first sub-area as the direction starting point, and setting the area center of the second sub-area as the direction end point, and determining the expansion and shrinkage direction according to the direction starting point and the direction end point; based on the expansion and shrinkage direction, determining multiple groups of target size parameters from the first size parameter and the second size parameter, the target size parameters being two size parameters adjacent in the expansion and shrinkage direction; based on the expansion and shrinkage direction, performing difference calculation on the target size parameters to obtain the inch expansion and shrinkage coefficient; calculating the average value of the size expansion and shrinkage coefficient to obtain the edge expansion and shrinkage coefficient.

[0087] The parameter generation module 305 determines a first correction parameter of the first target sub-region and a second correction parameter of the second target sub-region according to the maximum edge expansion and contraction coefficient.

[0088] In some embodiments, when the parameter generation module 305 is used to determine the first correction parameter of the first target sub-area and the second correction parameter of the second target sub-area according to the maximum edge expansion and contraction coefficient, it is specifically used to achieve: generating a grab offset parameter according to the edge expansion and contraction coefficient and a preset expansion and contraction offset function; using the grab offset parameter as the second correction parameter; if the first expansion and contraction standard deviation is less than the preset first standard threshold, removing the first correction parameter; if the first expansion and contraction standard deviation is greater than or equal to the preset standard threshold, using the grab offset parameter as the first correction parameter.

[0089] The mounting management module 306 adjusts the preset trajectory parameters for mounting the chip according to the first correction parameter or the second correction parameter to obtain the target parameter, grabs the chip to the base according to the target parameter, and performs heating and die bonding.

[0090] In some embodiments, the base includes a coordinate reference point. Before the placement management module 306 is used to adjust the preset trajectory parameters for mounting the chip according to the first correction parameter or the second correction parameter to obtain the target parameter, it is also specifically used to achieve: generating placement coordinates according to the preset base layout information and the coordinate reference point; generating initial trajectory parameters according to the coordinate reference point and the placement coordinates; if the placement coordinates are in the first target sub-area, correcting the initial trajectory parameters according to the first correction parameter to obtain the preset trajectory parameters; if the placement coordinates are in the second target sub-area, correcting the initial trajectory parameters according to the second correction parameter to obtain the preset trajectory parameters.

[0091] An embodiment of the present application provides an electronic device, which includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and implement any one of the bonding methods in the embodiments of the present application when executing the computer program.

[0092] An embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor implements any one of the die bonding methods in the embodiments of the present application.

[0093] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for bonding a die, It is characterized in that For fixing a chip on a base, the method comprises: Acquire a crystal-fixing surface image of the base, and identify a support image on the crystal-fixing surface image; Dividing the solid crystal surface image into a plurality of detection regions according to a preset first segmentation boundary, and dividing the detection region into a first sub-region and a second sub-region according to a plurality of preset second segmentation boundaries; By means of a preset edge detection network, feature extraction is performed on the support image in the first sub-region, a first edge corner point is determined, and a first size parameter of the first edge corner point is obtained, and a first edge feature set is generated according to the first size parameter; by means of a preset edge detection network, feature extraction is performed on the support image in the second sub-region, a second edge corner point is determined, and a second size parameter of the second edge corner point is obtained, and a second edge feature set is generated according to the second size parameter; Performing a standard deviation calculation on the first size parameter to obtain a first expansion and contraction standard deviation, and performing a standard deviation calculation on the second size parameter to obtain a second expansion and contraction standard deviation; if the first expansion and contraction standard deviation is greater than the second expansion and contraction standard deviation, setting the area center of the first sub-area as a direction starting point, setting the area center of the second sub-area as a direction end point, and determining the expansion and contraction direction according to the direction starting point and the direction end point; based on the expansion and contraction direction, determining multiple groups of target size parameters from the first size parameter and the second size parameter, the target size parameters being two size parameters adjacent to each other in the expansion and contraction direction; performing a difference calculation on the target size parameters based on the expansion and contraction direction to obtain a size expansion and contraction coefficient; calculating an average value of the size expansion and contraction coefficient to obtain an edge expansion and contraction coefficient, setting the second segmentation boundary corresponding to the largest edge expansion and contraction coefficient as a target segmentation boundary, and dividing the detection area into a first target sub-area and a second target sub-area according to the target segmentation boundary; Determine a first correction parameter of the first target sub-region and a second correction parameter of the second target sub-region respectively according to the largest edge expansion and contraction coefficient; The preset trajectory parameters for mounting the chip are adjusted according to the first correction parameter or the second correction parameter to obtain target parameters, and the chip is grabbed onto the base according to the target parameters to perform heating and die bonding.

2. The die bonding method according to claim 1, It is characterized in that The determining of the first correction parameter of the first target sub-region and the second correction parameter of the second target sub-region respectively according to the maximum edge expansion and contraction coefficient includes: Generate a grabbing offset parameter according to the edge expansion / contraction coefficient and a preset expansion / contraction offset function; Using the grab offset parameter as the second correction parameter; If the first expansion / contraction standard deviation is less than a preset first standard threshold, removing the first correction parameter; If the first expansion and contraction standard deviation is greater than or equal to a preset standard threshold, the grabbing offset parameter is used as the first correction parameter.

3. The die bonding method according to claim 2, It is characterized in that The base includes a coordinate reference point. Before adjusting the preset trajectory parameters for mounting the chip according to the first correction parameter or the second correction parameter to obtain the target parameter, the method further includes: Generate mounting coordinates according to preset base layout information and the coordinate reference points; Generate initial trajectory parameters according to the coordinate reference point and the mounting coordinates; If the placement coordinates are in the first target sub-area, correcting the initial trajectory parameters according to the first correction parameters to obtain the preset trajectory parameters; If the placement coordinates are in the second target sub-area, the initial trajectory parameters are corrected according to the second correction parameters to obtain the preset trajectory parameters.

4. The die bonding method according to claim 1, It is characterized in that The step of acquiring the crystal-fixing surface image of the base and identifying the support image on the crystal-fixing surface image includes: The crystal-fixing surface image of the base is acquired, and the support is identified on the crystal-fixing surface image to obtain the support image in the crystal-fixing surface image.

5. The die bonding method according to claim 1, It is characterized in that The preset edge detection network includes: a fully convolutional network, a DeepEdge network and a RefineNet network.

6. A crystal bonding device, It is characterized in that Used to implement the die bonding method according to any one of claims 1 to 5, used to fix the chip on the base, the die bonding device comprises: An image acquisition module, used to acquire an image of a crystal-fixing surface of the base and identify an image of a bracket on the image of the crystal-fixing surface; A region segmentation module, used to segment the solid crystal surface image into a plurality of detection regions according to a preset first segmentation boundary, and to divide the detection region into a first sub-region and a second sub-region according to a plurality of preset second segmentation boundaries; An edge extraction module is used to perform feature extraction on the support image in the first sub-region through a preset edge detection network, determine a first edge corner point, and obtain a first size parameter of the first edge corner point, and generate a first edge feature set according to the first size parameter; perform feature extraction on the support image in the second sub-region through a preset edge detection network, determine a second edge corner point, and obtain a second size parameter of the second edge corner point, and generate a second edge feature set according to the second size parameter; an area determination module, configured to perform standard deviation calculation on the first size parameter to obtain a first expansion and shrinkage standard deviation, and perform standard deviation calculation on the second size parameter to obtain a second expansion and shrinkage standard deviation; if the first expansion and shrinkage standard deviation is greater than the second expansion and shrinkage standard deviation, the area center of the first sub-area is set as a direction starting point, and the area center of the second sub-area is set as a direction end point, and the expansion and shrinkage direction is determined according to the direction starting point and the direction end point; based on the expansion and shrinkage direction, multiple groups of target size parameters are determined from the first size parameter and the second size parameter, and the target size parameters are two size parameters adjacent to each other in the expansion and shrinkage direction; based on the expansion and shrinkage direction, the target size parameters are calculated by difference calculation to obtain a size expansion and shrinkage coefficient; the average value of the size expansion and shrinkage coefficient is calculated to obtain an edge expansion and shrinkage coefficient, and the second segmentation boundary corresponding to the largest edge expansion and shrinkage coefficient is set as a target segmentation boundary, and the detection area is divided into a first target sub-area and a second target sub-area according to the target segmentation boundary; A parameter generation module, used to determine a first correction parameter of the first target sub-region and a second correction parameter of the second target sub-region respectively according to the maximum edge expansion and contraction coefficient; The mounting management module adjusts the preset trajectory parameters for mounting the chip according to the first correction parameter or the second correction parameter to obtain the target parameter, and grabs the chip to the base according to the target parameter to perform heating and die bonding.

7. An electronic device, It is characterized in that The electronic device comprises a memory and a processor; The memory is used to store computer programs; The processor is used to execute the computer program and implement the die bonding method according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the die bonding method according to any one of claims 1 to 5.