Method, device and equipment for detecting dense gold threads of packaged chip, medium and product

Abstract

The invention discloses a method, a device and equipment for detecting dense gold threads of a packaged chip, a medium and a product. The method comprises the following steps: carrying out detection area labeling on a packaged chip dense gold thread image to generate a target chip dense gold thread image with a target detection area, and determining the number of reference gold threads and the width of the reference gold threads of dense gold threads contained in the target detection area; performing gold thread skeleton trajectory feature extraction on a target detection area of the target chip dense gold thread image to obtain a gold thread skeleton trajectory feature map corresponding to the target detection area; according to the reference gold thread width, performing region division on the target detection region to obtain a plurality of detection sub-regions included in the target detection region; according to the gold thread skeleton trajectory feature map, based on the number of reference gold threads, determining region identifiers corresponding to the detection sub-regions; and generating a dense gold thread detection result of the target detection region of the dense gold thread image of the packaged chip according to the region identifier corresponding to each detection sub-region.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of semiconductor package detection, and in particular to a package chip dense gold wire detection method, device, equipment, medium and product. BACKGROUND

[0002] In the integrated circuit packaging process, dense gold wire is the core development direction of integrated circuit packaging technology, such as high-density interconnection and miniaturization packaging scene, and its core value lies in improving the chip interconnection efficiency and packaging integration, but the dense arrangement caused by space limitation causes three core defect risks: adhesion, deformation and fracture. The defect may directly cause chip electrical performance failure, therefore, dense gold wire quality detection is a key link of packaging quality control.

[0003] The existing dense gold wire detection technology has the problems of singleness and limitation of feature detection in defect detection, resulting in a large number of missed detection and false detection in the gold wire detection process, and being greatly affected by image edge contrast, and having poor anti-interference ability, especially in the detection accuracy of complex dense gold wire detection scene.

[0004] Therefore, it is urgent to propose a package chip dense gold wire detection method, which can have certain detection accuracy in complex dense gold wire detection scene, and avoid the occurrence of missed detection and false detection in the gold wire detection process. SUMMARY

[0005] The present application provides a package chip dense gold wire detection method, device, equipment, medium and product, to improve the dense gold wire detection accuracy in complex scene, and reduce the probability of missed detection and false detection in the gold wire detection process.

[0006] According to an aspect of the present application, a package chip dense gold wire detection method is provided, the method comprising:

[0007] Obtaining a package chip dense gold wire image, detecting and labeling the region of the package chip dense gold wire image, generating a target chip dense gold wire image with a target detection region, and determining the reference gold wire number and reference gold wire width of the dense gold wire contained in the target detection region;

[0008] Extracting the gold wire skeleton trajectory features of the target detection region of the target chip dense gold wire image to obtain the gold wire skeleton trajectory feature map corresponding to the target detection region;

[0009] According to the reference gold wire width, the target detection region is divided into several detection sub-regions contained in the target detection region;

[0010] According to the gold wire skeleton trajectory feature map, a region identifier corresponding to each detection sub-region is determined based on the reference gold wire quantity; the region identifier is a region detection pass identifier or a region detection fail identifier;

[0011] According to the region identifier corresponding to each detection sub-region, a dense gold wire detection result of the target detection region of the package chip dense gold wire image is generated.

[0012] According to another aspect of the present application, a package chip dense gold wire detection device is provided, and the device comprises:

[0013] A gold wire image acquisition module is configured to acquire a package chip dense gold wire image, perform target detection region annotation on the package chip dense gold wire image, generate a target chip dense gold wire image with a target detection region, and determine a reference gold wire quantity and a reference gold wire width of the dense gold wire contained in the target detection region;

[0014] A feature extraction module is configured to perform gold wire skeleton trajectory feature extraction on the target detection region of the target chip dense gold wire image, and obtain a gold wire skeleton trajectory feature map corresponding to the target detection region;

[0015] A region division module is configured to perform region division on the target detection region according to the reference gold wire width, and obtain a plurality of detection sub-regions contained in the target detection region;

[0016] A region identifier determination module is configured to determine a region identifier corresponding to each detection sub-region based on the gold wire skeleton trajectory feature map and the reference gold wire quantity; the region identifier is a region detection pass identifier or a region detection fail identifier;

[0017] A detection result generation module is configured to generate a dense gold wire detection result of the target detection region of the package chip dense gold wire image according to the region identifier corresponding to each detection sub-region.

[0018] According to another aspect of the present application, an electronic device is provided, and the electronic device comprises:

[0019] at least one processor; and

[0020] a memory in communication with the at least one processor; wherein

[0021] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the package chip dense gold wire detection method according to any one of the embodiments of the present application.

[0022] According to another aspect of the present application, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement the package chip dense gold line detection method according to any of the embodiments of the present application when executed by the processor.

[0023] According to another aspect of the present application, there is provided a computer program product comprising a computer program for implementing the package chip dense gold line detection method according to any of the embodiments of the present application when executed by a processor.

[0024] The technical scheme of the embodiment of the present application detects the target chip dense gold line image of the package chip dense gold line image to generate the target chip dense gold line image with the target detection region, determines the reference gold line number and the reference gold line width of the dense gold line contained in the target detection region, extracts the gold line skeleton track features of the target detection region of the target chip dense gold line image to obtain the gold line skeleton track feature map corresponding to the target detection region, divides the target detection region according to the reference gold line width to obtain a plurality of detection sub-regions contained in the target detection region, determines the region identifier corresponding to each detection sub-region based on the reference gold line number according to the gold line skeleton track feature map, and generates the dense gold line detection result of the target detection region of the package chip dense gold line image according to the region identifier corresponding to each detection sub-region. In the process of detecting the dense gold line of the package chip dense gold line image, the above technical scheme extracts the gold line skeleton features of the detection region, and detects the dense gold line in the detection region dimension with finer granularity by means of the detection sub-region division, thereby improving the dense gold line detection accuracy in the complex scene, improving the anti-interference ability to a certain extent, reducing the probability of missed detection and false detection in the gold line detection process, and improving the dense gold line detection efficiency while taking into account the dense gold line detection accuracy.

[0025] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0027] Figure 1A is a flowchart of a package chip dense gold line detection method according to an embodiment of the present application;

[0028] Figure 1B is a target chip dense gold line image schematic diagram marked with a target detection region according to an embodiment of the present application;

[0029] Figure 1C is a gold line skeleton trajectory feature map schematic diagram obtained by performing gold line skeleton trajectory feature extraction on the target detection region according to an embodiment of the present application;

[0030] Figure 1D is a detection sub-region schematic diagram after region division of the target detection region according to an embodiment of the present application;

[0031] Figure 1E is a trajectory feature sub-map schematic diagram of the detection sub-region according to an embodiment of the present application;

[0032] Figure 1F is a detection sub-region schematic diagram containing multiple gold line skeletons according to an embodiment of the present application;

[0033] Figure 2A is a flowchart of a packaged chip dense gold line detection method according to an embodiment of the present application;

[0034] Figure 2B is a gold line region feature map schematic diagram obtained by performing gold line region feature extraction on the target detection region according to an embodiment of the present application;

[0035] Figure 2C is a region feature sub-map schematic diagram of the detection sub-region according to an embodiment of the present application;

[0036] Figure 3 is a flowchart of a packaged chip dense gold line detection method according to an embodiment of the present application;

[0037] Figure 4 is a structural schematic diagram of a packaged chip dense gold line detection device according to an embodiment of the present application;

[0038] Figure 5 is a structural schematic diagram of an electronic device implementing a packaged chip dense gold line detection method according to an embodiment of the present application. DETAILED DESCRIPTION

[0039] In the following, the technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort should belong to the protection scope of the present application.

[0040] It should be noted that the terms "first", "second" and the like in the description and claims of the present application and the above drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in other than the order illustrated or described herein. In addition, the terms "comprise" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a list of steps or units does not necessarily limit to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to such processes, methods, products or devices.

[0041] Embodiment one

[0042] Figure 1A A flowchart of a packaged chip dense gold wire detection method is provided for the first embodiment of the present application. The present embodiment can be applicable to the quality detection of dense gold wires of packaged chips in a semiconductor scene. The method can be executed by a packaged chip dense gold wire detection method device, which can be realized in the form of hardware and / or software, and can be configured in an electronic device. As shown in the figure, the method comprises: Figure 1A

[0043] S110, obtaining a packaged chip dense gold wire image, performing detection region labeling on the packaged chip dense gold wire image, generating a target chip dense gold wire image with a target detection region, and determining a reference gold wire quantity and a reference gold wire width of the dense gold wire contained in the target detection region.

[0044] S120, performing gold wire skeleton trajectory feature extraction on the target detection region of the target chip dense gold wire image to obtain a gold wire skeleton trajectory feature map corresponding to the target detection region.

[0045] S130, performing region division on the target detection region according to the reference gold wire width to obtain a plurality of detection sub-regions contained in the target detection region.

[0046] ​S140, determining, according to the gold wire skeleton track feature map, a region identifier corresponding to each detection sub-region based on the reference gold wire quantity; the region identifier is a region detection pass identifier or a region detection fail identifier.

[0047] S150, generating a dense gold wire detection result of a target detection region of the packaged chip dense gold wire image according to the region identifier corresponding to each detection sub-region.

[0048] The packaged chip dense gold wire image can be an image focused on the densely arranged gold wire interconnection structure in the chip packaging region collected by a professional optical equipment. The image subject is the dense gold wire, which is continuous in gold wire line, clear in edge, uniform in gray scale, and has no overlap between adjacent gold wires in an ideal state. In a complex scene, there can be gold wire edge overlap or adhesion in a defect scene.

[0049] Before labeling the detection region of the packaged chip dense gold wire image, the packaged chip dense gold wire image can also be pre-processed to further improve the image quality. The pre-processing methods can include but are not limited to noise reduction, image enhancement, edge sharpening, and gray scale adjustment.

[0050] The region labeling of the detection region of the pre-processed packaged chip dense gold wire image is performed, wherein the labeling method can be manual framing or algorithm pre-positioning, and a target chip dense gold wire image with a target detection region is generated. The target detection region can also be referred to as a ROI (Region of Interest). Figure 1B As shown in a target chip dense gold wire image with a target detection region shown in a target chip dense gold wire image with a target detection region. Figure 1B The black rectangular region in the target chip dense gold wire image with a target detection region is the target detection region obtained by division.

[0051] The reference gold wire quantity and the reference gold wire width of the dense gold wire contained in the target detection region are determined. The reference gold wire quantity and the reference gold wire width can be manually labeled by a related technical personnel, and the reference gold wire quantity is the gold wire quantity contained in the target detection region; the reference gold wire width is the line width of any gold wire in the target detection region. If an automatic method is used to identify the reference gold wire quantity and the reference gold wire width of the target detection region, a gold wire parameter prediction model can be used for gold wire quantity and width prediction. Specifically, the target detection region of the target chip dense gold wire image can be regionally imaged to generate a region image corresponding to the target detection region, and the region image is input into a pre-trained gold wire parameter prediction model to obtain a prediction result output by the model, wherein the prediction result includes the predicted reference gold wire quantity and the reference gold wire width.

[0052] The embodiment also provides a model training manner of the gold line parameter prediction model. A historical chip dense gold line image with a target detection region in a historical time period is obtained, and a region image of the target detection region in the historical chip dense gold line image is intercepted to obtain a historical region image. A position of a gold line in the historical region image is labeled to obtain sample labels of a standard gold line quantity and a standard gold line width. The historical region image with the sample labels is input into a preselected network model to obtain a predicted gold line quantity and a predicted gold line width output by the model. The network model is trained based on the standard gold line quantity and the standard gold line width according to the predicted gold line quantity and the predicted gold line width until a preset model training end condition is met, and the gold line parameter prediction model is obtained. The model training end condition can be that a current loss value tends to be stable, or the current loss value reaches a set loss threshold, or a current iteration number reaches a set iteration number threshold, and the embodiment does not limit this.

[0053] If the reference gold line quantity and the reference gold line width are determined manually, the reference gold line quantity and the reference gold line width can be calibrated by related personnel based on a front-end interface of the parameter calibration software.

[0054] The Gaussian line model can be used to extract the gold line skeleton track features of the target detection region of the target chip dense gold line image to obtain the gold line skeleton track feature map corresponding to the target detection region. Specifically, a two-dimensional Gaussian function, that is, a Gaussian mask or a Gaussian kernel, is preselected. To ensure that the image gray value range after subsequent convolution remains unchanged, the Gaussian kernel is preprocessed to obtain a normalized Gaussian kernel. The region gray image corresponding to the target detection region is two-dimensionally convolved with the normalized Gaussian kernel to obtain a smoothed region image of the target detection region, which is referred to as a target smoothed region image hereinafter.

[0055] The Sobel operator is used to calculate the first-order partial derivatives of each pixel point in the target smoothed region image in the x and y directions to obtain the first-order partial derivative of each pixel point in the x direction and the first-order partial derivative of each pixel point in the y direction . The Laplace difference operator is used to calculate the second-order partial derivatives to obtain the quadratic term coefficients of the Taylor quadratic polynomial, including the term partial derivative, term partial derivative, and xy cross term partial derivative. According to the term partial derivative, term partial derivative, and xy cross term partial derivative, an approximate expression of the Taylor quadratic polynomial is generated. The line direction of the gold line is the vertical direction of the gray gradient, that is, the extension direction of the gold line. Therefore, the direction with the smallest gray change, that is, the line extension direction, can be found through the extreme value analysis of the Taylor quadratic polynomial, and the direction is perpendicular to the maximum change direction of the gray gradient.

[0056] The slope of the line direction is obtained by solving the unit vector of the line direction through the extreme value condition of the Taylor polynomial, and further converting the slope of the line direction to the angle of the line direction. It should be noted that the gray scale of the gold line changes slowly in the extension direction, i.e., the line direction, and is distributed in a "peak" shape in the vertical direction, i.e., the center gray scale is high and the edge is low. The gray scale maximum value point in the vertical line direction is screened through Non-Maximum Suppression (NMS), and then the sub-pixel accuracy is improved through quadratic fitting, so as to finally connect into a continuous contour. The candidate line points are screened through Low (low threshold) and High (high threshold) double thresholds, the high confidence points are reserved, the medium confidence points are connected, and finally the complete gold line skeleton track is formed, i.e., the gold line skeleton track feature map corresponding to the target detection region is obtained.

[0057] As shown in a gold line skeleton track feature map obtained by extracting the gold line skeleton track feature of the target detection region. Figure 1C As shown in a gold line skeleton track feature map obtained by extracting the gold line skeleton track feature of the target detection region.

[0058] According to the reference gold line width, the target detection region is divided into several detection sub-regions. The target detection region can be manually divided by a related technical personnel, and the target detection region can also be automatically divided according to the reference gold line width.

[0059] As shown in a gold line skeleton track feature map obtained by extracting the gold line skeleton track feature of the target detection region. Figure 1D As shown in a gold line skeleton track feature map obtained by extracting the gold line skeleton track feature of the target detection region. The determination method of the region width of the detection sub-region can be as follows:

[0060]

[0061] Wherein, n is a pre-set multiple, for example, n can be set to 2-5 times; is the reference gold line width.

[0062] According to the gold line skeleton track feature map, based on the reference gold line number, the region identifier corresponding to each detection sub-region is determined; wherein, the region identifier is a region detection pass identifier or a region detection fail identifier.

[0063] In an optional embodiment, according to the gold wire skeleton track feature map, based on the reference gold wire number, the region identifier corresponding to each detection sub-region is determined, including:

[0064] Step a1, according to the gold wire skeleton track feature map, the track feature sub-graph corresponding to each detection sub-region is determined.

[0065] For any detection sub-region, the intersection part of the gold wire skeleton track feature map and the region of the detection sub-region is determined as the track feature sub-graph of the detection sub-region. As shown in a track feature sub-graph diagram of a detection sub-region. Figure 1E

[0066] Step a2, according to the coincidence feature between the line end of the gold wire skeleton track in the track feature sub-graph of the corresponding detection sub-region and the axial edge of the sub-region, the effective gold wire number corresponding to each detection sub-region is determined.

[0067] For any detection sub-region, it is judged whether the line end of the gold wire skeleton track in the track feature sub-graph corresponding to the detection sub-region exists coincident with the axial edge of the sub-region. As shown in a detection sub-region diagram containing multiple gold wire skeleton tracks. Figure 1F

[0068] Step a3, according to the effective gold wire number of the corresponding detection sub-region and the reference gold wire number, it is judged whether the corresponding detection sub-region satisfies the preset effective gold wire number judgment condition.

[0069] Wherein, the effective gold wire number judgment condition can be set by the relevant technical personnel, specifically, the effective gold wire number judgment condition can be whether the effective gold wire number of the corresponding detection sub-region is not less than the reference gold wire number, if yes, it is determined that the corresponding detection sub-region satisfies the effective gold wire number judgment condition; if not, it is determined that the corresponding detection sub-region does not satisfy the effective gold wire number judgment condition.

[0070] Step a4, if yes, the region identifier of the corresponding detection sub-region is determined as the region detection passing identifier.

[0071] ​​Specifically, if the corresponding detection sub-region satisfies the preset effective gold line number judgment condition, that is, the number of effective gold lines in the corresponding detection sub-region is not less than the reference gold line number, the region identifier of the corresponding detection sub-region is determined as a region detection passing identifier.

[0072] If the corresponding detection sub-region does not satisfy the preset effective gold line number judgment condition, that is, the number of effective gold lines in the corresponding detection sub-region is less than the reference gold line number, the region identifier of the corresponding detection sub-region is determined as a region detection failure identifier.

[0073] The above technical solution determines the trajectory feature subgraph corresponding to each detection sub-region based on the gold line skeleton trajectory feature map, and determines the number of effective gold lines corresponding to each detection sub-region according to the overlapping feature between the line end of the gold line skeleton trajectory in the trajectory feature subgraph of the corresponding detection sub-region and the axial edge of the sub-region. When the number of effective gold lines of the corresponding detection sub-region is not less than the reference gold line number, the region identifier of the corresponding detection sub-region is determined as a region detection passing identifier, which realizes accurate determination of the region identifier of the corresponding detection sub-region. The dense gold line detection is performed by extracting the gold line trajectory features of the detection sub-region, which improves the accuracy of dense gold line detection. Especially in complex scenes such as poor gold line imaging quality or adjacent line edge overlap, whether the corresponding detection sub-region of the gold line trajectory feature passes the detection is determined, which improves the accuracy of determining the region identifier of the detection sub-region.

[0074] According to the region identifier corresponding to each detection sub-region, a dense gold line detection result of the target detection region of the package chip dense gold line image is generated. For example, a gold line detection passing number threshold can be set according to the total number of detection sub-regions. When the number of detection sub-regions whose region identifier is a region detection passing identifier is greater than the gold line detection passing number threshold, it is determined that the dense gold line detection result of the target detection region of the package chip dense gold line image is a detection passing, that is, it is determined that the current group of dense gold lines does not exist abnormally. Conversely, when the number of detection sub-regions whose region identifier is a region detection passing identifier is not greater than the gold line detection passing number threshold, it is determined that the dense gold line detection result of the target detection region of the package chip dense gold line image is a detection failure, that is, it is determined that the current group of dense gold lines exists abnormally.

[0075] The technical scheme of the embodiment of the present application detects the region label of the packaged chip dense gold line image to generate a target chip dense gold line image with a target detection region, determines the reference gold line quantity and the reference gold line width of the dense gold line contained in the target detection region, extracts the gold line skeleton track feature of the target detection region of the target chip dense gold line image to obtain the gold line skeleton track feature map corresponding to the target detection region, divides the target detection region according to the reference gold line width to obtain a plurality of detection sub-regions contained in the target detection region, determines the region identifier corresponding to each detection sub-region based on the reference gold line quantity according to the gold line skeleton track feature map, and generates the dense gold line detection result of the target detection region of the packaged chip dense gold line image according to the region identifier corresponding to each detection sub-region. In the above technical scheme, the gold line skeleton feature of the detection region is extracted, and the detection sub-region division is used to detect the dense gold line in the detection region dimension with finer granularity by combining the gold line skeleton feature, so that the dense gold line detection accuracy in a complex scene is improved, the anti-interference ability is improved to a certain extent, the missed detection and false detection and other situations in the gold line detection process are reduced, the dense gold line detection efficiency is improved, and the dense gold line detection accuracy is considered.

[0076] Embodiment two

[0077] Figure 2A A flowchart of a packaged chip dense gold line detection method provided by the second embodiment of the present application is provided, and the embodiment is optimized and improved on the basis of the above technical schemes.

[0078] Further, the step of “determining the region identifier corresponding to each detection sub-region based on the reference gold line quantity according to the gold line skeleton track feature map” is refined as “determining the track feature sub-map corresponding to each detection sub-region according to the gold line skeleton track feature map; determining the effective gold line quantity corresponding to each detection sub-region according to the coincidence feature between the line end of the gold line skeleton track in the track feature sub-map of the corresponding detection sub-region and the axial edge of the sub-region; judging whether the corresponding detection sub-region meets the preset effective gold line quantity judgment condition according to the effective gold line quantity of the corresponding detection sub-region and the reference gold line quantity; and if yes, determining that the region identifier of the corresponding detection sub-region is a region detection passing identifier”.

[0079] Correspondingly, after the step of judging whether the corresponding detection sub-region satisfies the preset effective gold line number judgment condition, a step of performing gold line region feature extraction on the target detection region of the target chip dense gold line image to obtain a gold line region feature map corresponding to the target detection region is added; a region feature sub-map corresponding to the corresponding detection sub-region is determined according to the gold line region feature map; if there is at least one gold line cluster region in the region feature sub-map of the corresponding detection sub-region, a line cluster prediction number corresponding to the corresponding detection sub-region is determined; whether the corresponding detection sub-region satisfies a preset equivalent gold line number judgment condition is judged according to the line cluster prediction number and the reference gold line number; if yes, the region identifier of the corresponding detection sub-region is determined as a region detection passing identifier. Thus, the determination manner of the region identifier of the corresponding detection sub-region is perfected.

[0080] It should be noted that the parts not described in detail in the embodiments of the present application can refer to the descriptions of other embodiments. For example, Figure 2A As shown in the figure, the method comprises the following specific steps:

[0081] S201, a packaged chip dense gold line image is acquired, a target detection region is labeled on the packaged chip dense gold line image, a target chip dense gold line image with a target detection region is generated, and a reference gold line number and a reference gold line width of the dense gold line contained in the target detection region are determined.

[0082] S202, gold line skeleton track feature extraction is performed on the target detection region of the target chip dense gold line image to obtain a gold line skeleton track feature map corresponding to the target detection region.

[0083] S203, according to the reference gold line width, the target detection region is regionally divided to obtain a plurality of detection sub-regions contained in the target detection region.

[0084] S204, according to the gold line skeleton track feature map, a track feature sub-map corresponding to each detection sub-region is determined.

[0085] S205, according to the coincidence feature between the line end of the gold line skeleton track in the track feature sub-map of the corresponding detection sub-region and the axial edge of the sub-region, an effective gold line number corresponding to each detection sub-region is determined.

[0086] S206, whether the corresponding detection sub-region satisfies a preset effective gold line number judgment condition is judged according to the effective gold line number and the reference gold line number of the corresponding detection sub-region; if yes, S207 is executed; if not, S208-S212 are executed.

[0087] S207, the region identifier of the corresponding detection sub-region is determined as a region detection passing identifier.

[0088] S208, gold wire region feature extraction is performed on the target detection region of the target chip dense gold wire image to obtain a gold wire region feature map corresponding to the target detection region.

[0089] For example, the gold wire region feature in the target detection region can be extracted by threshold segmentation, and a closed operation morphological preprocessing in the specified gold wire direction is performed to obtain the gold wire region feature map corresponding to the target detection region.

[0090] Specifically, the OTSU (automatic image segmentation technology) adaptive threshold segmentation can be used to segment the region image corresponding to the target detection region into a gold wire region (foreground region) and a background region. A structural element such as a rectangular region part is constructed along the gold wire axis direction, and a closed operation is performed to fill the gold wire region holes and connect the broken edges. The connected domain analysis such as the base eight connected rule is performed on the image after the closed operation to extract the pixel set, area, and circumscribed rectangle of all gold wire region clusters, and form the gold wire region feature map of the target detection region.

[0091] As shown in a gold wire region feature map obtained by performing gold wire region feature extraction on a target detection region. Figure 2B As shown in a gold wire region feature map obtained by performing gold wire region feature extraction on a target detection region. The red region part is the extracted dense gold wire region.

[0092] S209, according to the gold wire region feature map, a region feature sub-map corresponding to the detection sub-region is determined.

[0093] For any detection sub-region, the region intersection part of the gold wire region feature map and the detection sub-region is determined as the region feature sub-map of the detection sub-region. As shown in a region feature sub-map of a detection sub-region. Figure 2C As shown in a region feature sub-map of a detection sub-region. The red marked part is the gold wire region of the detection sub-region.

[0094] S210, if there is at least one gold wire cluster region in the region feature sub-map of the corresponding detection sub-region, the number of wire cluster predictions corresponding to the detection sub-region is determined.

[0095] The gold wire cluster region is the gold wire region, which can refer to the region feature sub-map of the detection sub-region shown in Figure 2C As shown in a region feature sub-map of a detection sub-region. The red part of the region is the gold wire region, also known as the gold wire cluster region. Figure 2C As shown in a detection sub-region, there are two gold wire cluster regions.

[0096] In an optional embodiment, if there is at least one gold line block region in the region feature subgraph of the corresponding detection subregion, the line cluster prediction number corresponding to the detection subregion is determined as follows: if there is at least one gold line block region in the region feature subgraph of the corresponding detection subregion, the region line block area corresponding to each gold line block region is determined; the standard line block area is determined according to the region width of the corresponding detection subregion based on the reference gold line width; the line block area ratio is determined according to the region line block area corresponding to each region of the corresponding detection subregion and the standard cluster area; and the line cluster prediction number corresponding to the corresponding detection subregion is determined according to the line cluster area ratio.

[0097] Specifically, for any gold line block region, the region line block area corresponding to the gold line block region is determined according to the pixel number of the gold line block region. The standard line block area is determined according to the region width of the corresponding detection subregion based on the reference gold line width. Since the region widths of the detection subregions are the same, the region width of any detection subregion is denoted as W, and the reference gold line width is denoted as Wg, then the standard cluster area is determined as follows:

[0098]

[0099] wherein, a denotes the calibration coefficient, which can be pre-set by a person skilled in the art, for example, a = 0.7 or a = 0.8.

[0100] The line block area ratio is determined according to the region line block area corresponding to each region of the corresponding detection subregion and the standard cluster area. Specifically, the region line block areas of the corresponding detection subregion are added up, for example, if the detection subregion A includes three gold line block regions, the region line block areas corresponding to the three gold line block regions are A1, A2 and A3 respectively, and the detection subregion A is taken as an example, the region cluster areas of the three gold line block regions of the detection subregion A are summed up as follows: The line cluster prediction number corresponding to the detection subregion is determined according to the line cluster area ratio. For example, if the line cluster area ratio is A, the line cluster prediction number corresponding to the detection subregion is determined as follows:

[0101]

[0102] ​​​​​​​​​​​is an integer, the integer value corresponding to the linear cluster area ratio can be determined as the predicted number of linear clusters corresponding to the detection sub-region. If the linear cluster area ratio is a decimal number, rounding principles or rounding up and down principles can be used for integer conversion, which can be pre-set by a person skilled in the art according to actual needs. The converted integer value is determined as the predicted number of linear clusters corresponding to the detection sub-region.

[0103] S211, determining whether the corresponding detection sub-region meets the preset equivalent gold line number judgment condition according to the predicted number of linear clusters and the reference gold line number.

[0104] The equivalent gold line number judgment condition can be pre-set by a person skilled in the art. Specifically, the equivalent gold line number judgment condition can be whether the predicted number of linear clusters of the corresponding detection sub-region is not less than the reference gold line number. For example, if the predicted number of linear clusters is not less than the reference gold line number, it is determined that the corresponding detection sub-region meets the preset equivalent gold line number judgment condition. If the predicted number of linear clusters is less than the reference gold line number, it is determined that the corresponding detection sub-region does not meet the preset equivalent gold line number judgment condition.

[0105] S212, if yes, determining that the region identifier of the corresponding detection sub-region is a region detection passing identifier.

[0106] If the corresponding detection sub-region meets the preset equivalent gold line number judgment condition, the region identifier of the corresponding detection sub-region is determined as a region detection passing identifier. If the corresponding detection sub-region does not meet the preset equivalent gold line number judgment condition, the region identifier of the corresponding detection sub-region is determined as a region detection failure identifier.

[0107] S213, generating a dense gold line detection result of the target detection region of the packaged chip dense gold line image according to the region identifiers respectively corresponding to each detection sub-region.

[0108] In an optional embodiment, generating a dense gold line detection result of the target detection region of the packaged chip dense gold line image according to the region identifiers respectively corresponding to each detection sub-region includes: determining the passing sub-region number of the detection sub-region whose region identifier is a region detection passing identifier in each detection sub-region; determining the total number of sub-regions corresponding to each detection sub-region; determining the sub-region number ratio between the passing sub-region number and the total number of sub-regions; and generating a dense gold line detection result of the target detection region of the packaged chip dense gold line image based on the preset reference ratio threshold according to the sub-region number ratio.

[0109] For example, if the passing sub-region number of the detection sub-region whose region identifier is a region detection passing identifier in each detection sub-region is ​, the total number of sub-regions corresponding to each detection sub-region is , the determination of the sub-region number ratio between the sub-region number and the total number of sub-regions is as follows:

[0110]

[0111] It can be understood that the sub-region number ratio between the sub-region number and the total number of sub-regions is used to represent the detection pass rate. If the sub-region number ratio is greater than a preset reference ratio threshold, the dense gold line detection result of the target detection region of the packaged chip dense gold line image is the current group of dense gold lines normal; if the sub-region number ratio is not greater than the preset reference ratio threshold, the dense gold line detection result of the target detection region of the packaged chip dense gold line image is the current group of dense gold lines abnormal. The reference ratio threshold can be preset by a related technical person, for example, the reference ratio threshold can be set to 90%.

[0112] The technical scheme of the embodiment determines the gold line region feature map corresponding to the target detection region by extracting the gold line region feature of the target chip dense gold line image when it is determined that the corresponding detection sub-region does not meet the effective gold line number judgment condition, determines the region feature subgraph corresponding to the corresponding detection sub-region according to the gold line region feature map, determines the line cluster prediction number corresponding to the corresponding detection sub-region if there is at least one gold line cluster region in the region feature subgraph of the corresponding detection sub-region, judges whether the corresponding detection sub-region meets the preset equivalent gold line number judgment condition according to the line cluster prediction number and the reference gold line number, and if so, determines the region identifier of the corresponding detection sub-region as a region detection pass identifier. In the judgment process of whether the region detection passes, dual feature fusion judgment is adopted, that is, the gold line skeleton track feature and the line region shape feature are introduced, and the detection accuracy in complex scenes such as gold line gray level jump and line overlap is improved.

[0113] Embodiment three

[0114] Figure 3 A flow structure schematic diagram of a packaged chip dense gold line detection method provided by the third embodiment of the application. The present embodiment is based on the above-mentioned embodiments and provides a preferred example.

[0115] As shown in Figure 3 , the method comprises the following steps:

[0116] S1, obtaining a packaged chip dense gold line image, labeling a detection region of the packaged chip dense gold line image, generating a target chip dense gold line image with a target detection region, and determining a reference gold line number and a reference gold line width of the dense gold line contained in the target detection region.

[0117] ​S21, gold line skeleton track feature extraction is performed on the target detection region of the target chip dense gold line image to obtain a gold line skeleton track feature map corresponding to the target detection region.

[0118] S22, gold line region feature extraction is performed on the target detection region of the target chip dense gold line image to obtain a gold line region feature map corresponding to the target detection region.

[0119] S3, region division is performed on the target detection region according to the reference gold line width to obtain a plurality of detection sub-regions contained in the target detection region.

[0120] S41, a track feature sub-map corresponding to each detection sub-region is determined according to the gold line skeleton track feature map.

[0121] S42, a region feature sub-map corresponding to each detection sub-region is determined according to the gold line region feature map.

[0122] S5, the number of effective gold lines corresponding to each detection sub-region is determined according to the coincidence feature between the line end of the gold line skeleton track in the track feature sub-map of the corresponding detection sub-region and the axial edge of the sub-region.

[0123] S6, whether the corresponding detection sub-region meets the preset effective gold line number judgment condition is judged according to the number of effective gold lines of the corresponding detection sub-region and the reference gold line number; if yes, S7 is executed; if no, S8-S10 are executed.

[0124] S7, the region identifier of the corresponding detection sub-region is determined as a region detection pass identifier.

[0125] S8, if there is at least one gold line cluster region in the region feature sub-map of the corresponding detection sub-region, the number of line cluster predictions corresponding to the corresponding detection sub-region is determined.

[0126] Specifically, the area of each gold line cluster region is determined respectively; according to the region width of the corresponding detection sub-region, the standard line cluster area is determined based on the reference gold line width; according to the area of each region line cluster area and the standard cluster area corresponding to the corresponding detection sub-region, the line cluster area ratio is determined; according to the line cluster area ratio, the number of line cluster predictions corresponding to the corresponding detection sub-region is determined.

[0127] S9, whether the corresponding detection sub-region meets the preset equivalent gold line number judgment condition is judged according to the number of line cluster predictions and the reference gold line number; if yes, S7 is executed; if no, S10 is executed;

[0128] S10, the region identifier of the corresponding detection sub-region is determined as a region detection fail identifier.

[0129] S11, determine the passing sub-region number of each detection sub-region in which the region identifier is a region detection passing identifier, and determine the total number of sub-regions corresponding to each detection sub-region.

[0130] S12, determine the sub-region number ratio between the passing sub-region number and the total number of sub-regions.

[0131] S13A, if the sub-region number ratio is greater than a preset reference ratio threshold, the dense gold line detection result of the target detection region of the packaged chip dense gold line image is the current group of dense gold line normal.

[0132] S13B, if the sub-region number ratio is not greater than the preset reference ratio threshold, the dense gold line detection result of the target detection region of the packaged chip dense gold line image is the current group of dense gold line abnormal.

[0133] The technical scheme of the embodiment combines the gold line skeleton track feature and the line region shape feature, and can be applied to complex scenes such as gold line gray level jump and line overlap through double feature fusion judgment. The effective line segment number condition of the gold line track is verified first, and then the line number is estimated through the line cluster area ratio. The double conditions are easy first and difficult later. For most sub-regions with obvious line features, simple detection processing is performed. For potential defect sub-regions, rejudgment is performed, which improves the detection efficiency to a certain extent and ensures the accurate judgment of defects. Global line complex interference noise is simplified after dividing sub-regions. Each sub-region only focuses on whether it is abnormal, and all sub-region abnormal flags are combined for comprehensive judgment, which can effectively improve the anti-interference ability. In addition, by setting a sub-region success rate user parameter, the gold line detection missed detection and false alarm can be adjusted.

[0134] Embodiment four

[0135] Figure 4 A structure diagram of a packaged chip dense gold line detection device provided by the embodiment four of the application. The packaged chip dense gold line detection device provided by the embodiment of the application can be applied to the quality detection of the dense gold line of the packaged chip in the semiconductor scene. The packaged chip dense gold line detection device can be realized in the form of hardware and / or software, as shown in the figure, which comprises a gold line image acquisition module 401, a feature extraction module 402, a region division module 403, a region identifier determination module 404 and a detection result generation module 405. Among them, Figure 4

[0136] The gold line image acquisition module 401 is used for acquiring the packaged chip dense gold line image, labeling the detection region of the packaged chip dense gold line image, generating the target chip dense gold line image with the target detection region, and determining the reference gold line number and the reference gold line width of the dense gold line contained in the target detection region. ​

[0137] The feature extraction module 402 is configured to perform gold wire skeleton track feature extraction on the target detection region of the target chip dense gold wire image, to obtain a gold wire skeleton track feature map corresponding to the target detection region.

[0138] The region division module 403 is configured to perform region division on the target detection region according to the reference gold wire width, to obtain a plurality of detection sub-regions contained in the target detection region.

[0139] The region identifier determination module 404 is configured to determine, according to the gold wire skeleton track feature map and based on the reference gold wire quantity, a region identifier corresponding to each detection sub-region respectively; the region identifier is a region detection pass identifier or a region detection fail identifier.

[0140] The detection result generation module 405 is configured to generate a dense gold wire detection result for the target detection region of the packaged chip dense gold wire image according to the region identifier corresponding to each detection sub-region respectively.

[0141] The technical scheme of the embodiment of the application generates a target chip dense gold wire image with a target detection region by performing detection region annotation on a packaged chip dense gold wire image, determines a reference gold wire quantity and a reference gold wire width of the dense gold wire contained in the target detection region, performs gold wire skeleton track feature extraction on the target detection region of the target chip dense gold wire image, to obtain a gold wire skeleton track feature map corresponding to the target detection region, performs region division on the target detection region according to the reference gold wire width, to obtain a plurality of detection sub-regions contained in the target detection region, determines, according to the gold wire skeleton track feature map and based on the reference gold wire quantity, a region identifier corresponding to each detection sub-region respectively, and generates a dense gold wire detection result for the target detection region of the packaged chip dense gold wire image according to the region identifier corresponding to each detection sub-region respectively. In the dense gold wire detection process of the packaged chip dense gold wire image, the above technical scheme extracts the gold wire skeleton feature of the detection region, and performs dense gold wire detection in the detection sub-region division manner in the detection region dimension with finer granularity, thereby improving the dense gold wire detection accuracy in a complex scene, improving the anti-interference capability to a certain extent, reducing the probability of missed detection and false detection in the gold wire detection process, and improving the dense gold wire detection efficiency while taking into account the dense gold wire detection accuracy.

[0142] Optionally, the region identifier determination module 404 comprises:

[0143] The track feature sub-map determination unit is configured to determine, according to the gold wire skeleton track feature map, a track feature sub-map corresponding to each detection sub-region respectively.

[0144] The effective gold wire number determination unit is configured to determine the number of effective gold wires corresponding to each detection sub-region according to the coincidence feature between the line end of the gold wire skeleton track in the track feature sub-graph of the corresponding detection sub-region and the axial edge of the sub-region.

[0145] The effective gold wire number judgment unit is configured to judge whether the corresponding detection sub-region meets the preset effective gold wire number judgment condition according to the number of effective gold wires and the reference gold wire number of the corresponding detection sub-region.

[0146] The first identifier determination unit is configured to determine the region identifier of the corresponding detection sub-region as a region detection pass identifier if the corresponding detection sub-region meets the preset effective gold wire number judgment condition.

[0147] Optionally, the region identifier determination module 404 further includes:

[0148] The gold wire region feature map determination unit is configured to perform gold wire region feature extraction on the target detection region of the target chip dense gold wire image to obtain the gold wire region feature map corresponding to the target detection region if the corresponding detection sub-region does not meet the preset effective gold wire number judgment condition.

[0149] The region feature sub-graph determination unit is configured to determine the region feature sub-graph corresponding to the corresponding detection sub-region according to the gold wire region feature map.

[0150] The line cluster prediction number determination unit is configured to determine the line cluster prediction number corresponding to the corresponding detection sub-region if there is at least one gold wire cluster region in the region feature sub-graph of the corresponding detection sub-region.

[0151] The equivalent gold wire number judgment unit is configured to judge whether the corresponding detection sub-region meets the preset equivalent gold wire number judgment condition according to the line cluster prediction number and the reference gold wire number.

[0152] The second identifier determination unit is configured to determine the region identifier of the corresponding detection sub-region as a region detection pass identifier if the corresponding detection sub-region meets the preset equivalent gold wire number judgment condition.

[0153] Optionally, the line cluster prediction number determination unit is specifically configured to:

[0154] If there is at least one gold wire cluster region in the region feature sub-graph of the corresponding detection sub-region, the region line cluster area corresponding to each gold wire cluster region is determined.

[0155] The standard line cluster area is determined according to the region width of the corresponding detection sub-region and based on the reference gold wire width.

[0156] According to the area of each of the region line cluster corresponding to the corresponding detection sub-region and the standard cluster area, a line cluster area ratio is determined.

[0157] According to the line cluster area ratio, a line cluster prediction quantity corresponding to the corresponding detection sub-region is determined.

[0158] Optionally, the region identifier determination module 404 further comprises:

[0159] The third identifier determination unit is configured to determine the region identifier of the corresponding detection sub-region as a region detection failure identifier if the corresponding detection sub-region does not satisfy the preset equivalent gold line quantity judgment condition.

[0160] Optionally, the detection result generation module 405 is specifically configured to:

[0161] determine a passing sub-region quantity of the detection sub-region whose region identifier is the region detection passing identifier in each of the detection sub-regions; and

[0162] determine a total sub-region quantity corresponding to each of the detection sub-regions;

[0163] determine a sub-region quantity ratio between the passing sub-region quantity and the total sub-region quantity;

[0164] According to the sub-region quantity ratio, a dense gold line detection result of a target detection region of the packaged chip dense gold line image is generated based on a preset reference ratio threshold.

[0165] The packaged chip dense gold line detection device provided in the embodiments of the present application can execute the packaged chip dense gold line detection method provided in any of the embodiments of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0166] Embodiment five

[0167] Figure 5 A structural schematic diagram of an electronic device 50 that can be used to implement embodiments of the present application is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smart phones, wearable devices (e.g., headsets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit the implementations of the present application described and / or claimed in this document.

[0168] As Figure 5As shown, the electronic device 50 includes at least one processor 51, and a memory, such as a read-only memory (ROM) 52, a random access memory (RAM) 53, etc., communicatively connected to the at least one processor 51, where the memory stores computer programs executable by the at least one processor. The processor 51 can perform various appropriate actions and processes according to the computer programs stored in the read-only memory (ROM) 52 or loaded from the storage unit 58 into the random access memory (RAM) 53. Various programs and data required for the operation of the electronic device 50 can also be stored in the RAM 53. The processor 51, the ROM 52, and the RAM 53 are connected to each other through a bus 54. An input / output (I / O) interface 55 is also connected to the bus 54.

[0169] Various components in the electronic device 50 are connected to the I / O interface 55, including an input unit 56, such as a keyboard, a mouse, etc., an output unit 57, such as various types of displays, a speaker, etc., a storage unit 58, such as a magnetic disk, an optical disk, etc., and a communication unit 59, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 59 allows the electronic device 50 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0170] The processor 51 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the processor 51 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 51 performs various methods and processes described above, such as the package chip dense gold wire detection method.

[0171] In some embodiments, the package chip dense gold wire detection method can be implemented as a computer program tangibly embodied in a computer readable storage medium, such as the storage unit 58. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 50 via the ROM 52 and / or the communication unit 59. When the computer program is loaded onto the RAM 53 and executed by the processor 51, one or more steps of the package chip dense gold wire detection method described above can be performed. Alternatively, in other embodiments, the processor 51 can be configured to perform the package chip dense gold wire detection method by any other appropriate means, such as by means of firmware.

[0172] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a load programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0173] Computer programs used to implement the processes of the application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer program

[0174] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store computer programs for use by or in connection with an instruction execution system, apparatus, or device. Computer-readable storage media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium will include one or more lines of a program of instructions in a transitory signal, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0175] To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0176] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0177] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. A server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, to solve the defects of large management difficulty and weak business scalability in traditional physical host and VPS service.

[0178] It should be understood that the various forms of flow shown above can be re-ordered, added to, or deleted from without departing from the scope of the present disclosure. For example, the steps recited in the present disclosure can be executed in parallel, executed in sequence, or executed in a different order, as long as the desired results of the present disclosure are achieved, and the present disclosure is not limited herein.

[0179] The specific embodiments described above are not intended to be limiting, and persons skilled in the art will appreciate that various modifications, combinations, sub-combinations and alternatives can be made to the specific embodiments without departing from the spirit and principles of the disclosure. Accordingly, the disclosure is not limited to the specific embodiments described above, but only by the scope of the appended claims.

Claims

1. A method for detecting dense gold wires in packaged chips, characterized in that, include: A dense gold wire image of a packaged chip is acquired, a detection area is marked on the dense gold wire image of the packaged chip, a target chip dense gold wire image with a target detection area is generated, and the number and width of reference gold wires of the dense gold wires contained in the target detection area are determined. Gold wire skeleton trajectory features are extracted from the target detection region of the dense gold wire image of the target chip to obtain the gold wire skeleton trajectory feature map corresponding to the target detection region. Based on the reference gold line width, the target detection area is divided into several detection sub-regions. Based on the gold wire skeleton trajectory feature map and the number of reference gold wires, a region identifier is determined for each of the detection sub-regions; the region identifier is either a region detection pass identifier or a region detection fail identifier. Based on the region identifiers corresponding to each of the detection sub-regions, a dense gold line detection result for the target detection region of the dense gold line image of the packaged chip is generated.

2. The method according to claim 1, characterized in that, The step of determining the region identifier corresponding to each of the detection sub-regions based on the gold wire skeleton trajectory feature map and the number of reference gold wires includes: Based on the gold wire skeleton trajectory feature map, determine the trajectory feature sub-map corresponding to each of the detection sub-regions; Based on the overlap characteristics between the line ends of the gold wire skeleton trajectory and the axial edge of the sub-region in the trajectory feature sub-image of the corresponding detection sub-region, the number of effective gold wires corresponding to each detection sub-region is determined. Based on the number of valid gold lines and the number of reference gold lines in the corresponding detection sub-region, determine whether the corresponding detection sub-region meets the preset judgment condition for the number of valid gold lines. If so, then the region identifier of the corresponding detection sub-region is determined as the region detection pass identifier.

3. The method according to claim 2, characterized in that, The method further includes: If the corresponding detection sub-region does not meet the preset valid gold wire quantity judgment condition, then the gold wire region feature is extracted for the target detection region of the dense gold wire image of the target chip to obtain the gold wire region feature map corresponding to the target detection region. Based on the gold line region feature map, determine the region feature sub-map corresponding to the corresponding detection sub-region; If there is at least one gold thread ball region in the regional feature sub-image of the corresponding detection sub-region, then the predicted number of thread balls corresponding to the corresponding detection sub-region is determined. Based on the predicted number of gold threads and the reference number of gold threads, determine whether the corresponding detection sub-region meets the preset equivalent gold thread number judgment condition; If so, then the region identifier of the corresponding detection sub-region is determined as the region detection pass identifier.

4. The method according to claim 3, characterized in that, If at least one gold thread ball region exists in the regional feature sub-image of the corresponding detection sub-region, then the predicted number of thread balls corresponding to the corresponding detection sub-region is determined, including: If at least one gold thread ball region exists in the regional feature sub-image of the corresponding detection sub-region, then the area of ​​the gold thread ball region corresponding to each gold thread ball region is determined. Based on the width of the corresponding detection sub-region, the area of ​​the standard line cluster is determined according to the width of the reference gold line. The ratio of the area of ​​the thread ball is determined based on the area of ​​each thread ball corresponding to the corresponding detection sub-region and the area of ​​the standard thread ball. Based on the ratio of the coil area, the predicted number of coils corresponding to the corresponding detection sub-region is determined.

5. The method according to claim 3, characterized in that, The method further includes: If the corresponding detection sub-region does not meet the preset condition for the number of equivalent gold lines, then the region identifier of the corresponding detection sub-region is determined to be the region detection failure identifier.

6. The method according to claim 1, characterized in that, The step of generating dense gold line detection results for the target detection region of the dense gold line image of the packaged chip based on the region identifiers corresponding to each of the detection sub-regions includes: The number of passed sub-regions in each of the aforementioned detection sub-regions whose region identifier is the region detection pass identifier; and, Determine the total number of sub-regions corresponding to each of the aforementioned detection sub-regions; Determine the ratio of the number of sub-regions to the total number of sub-regions; Based on the ratio of the number of sub-regions and a preset reference ratio threshold, a dense gold line detection result is generated for the target detection region of the dense gold line image of the packaged chip.

7. A device for detecting dense gold wire bonding in packaged chips, characterized in that, include: The gold wire image acquisition module is used to acquire a dense gold wire image of the packaged chip, mark the detection area of ​​the dense gold wire image of the packaged chip, generate a target chip dense gold wire image with the target detection area, and determine the reference gold wire number and reference gold wire width of the dense gold wires contained in the target detection area. The feature extraction module is used to extract the gold wire skeleton trajectory features of the target detection area of ​​the dense gold wire image of the target chip, and obtain the gold wire skeleton trajectory feature map corresponding to the target detection area. The region division module is used to divide the target detection region according to the width of the reference gold line, so as to obtain several detection sub-regions contained in the target detection region; The region identifier determination module is used to determine the region identifier corresponding to each of the detection sub-regions based on the gold wire skeleton trajectory feature map and the number of reference gold wires; the region identifier is a region detection passed identifier or a region detection failed identifier. The detection result generation module is used to generate dense gold line detection results for the target detection region of the dense gold line image of the packaged chip based on the region identifiers corresponding to each of the detection sub-regions.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the packaged chip dense gold wire inspection method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that are used to cause a processor to execute the method for dense gold wire detection of packaged chips according to any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method for dense gold wire detection of packaged chips according to any one of claims 1-6.

Images