Printed circuit board copper embedded quality analysis method and system

By performing semantic vector analysis of the monitoring image data collected during the copper embedding process of printed circuit boards, the problem of low reliability of copper embedding quality analysis in the prior art is solved, and a more accurate copper embedding quality evaluation is achieved .

CN118583888BActive Publication Date: 2024-11-19SICHUAN HUASON ELECTRONICS TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411046456.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-19
Estimated Expiration
2044-08-01

AI Technical Summary

Technical Problem

In the prior art, the reliability of the printed circuit board embedded copper quality analysis is not high, especially the detection efficiency of concealed or small defects is low, resulting in the inaccurate quality analysis of the printed circuit board during the copper embedding process.

Method used

A printed circuit board embedded copper quality analysis method is adopted. By collecting copper embedded monitoring image data, the circuit board semantic vector and copper embedded monitoring image semantic vector are determined. Based on these vectors, the copper embedded quality analysis data is output to improve the analysis. reliability.

Benefits of technology

The reliability of copper embedded quality analysis of printed circuit boards can more accurately reflect the degree of damage during copper embedded process, and improve the inefficiency and low reliability problems in the prior art.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118583888B_ABST
    Figure CN118583888B_ABST
Patent Text Reader

Abstract

The printed circuit board copper embedding quality analysis method and system provided in the present application relate to the copper embedding quality analysis technical field. In the present application, first, the target copper embedding monitoring image data of the target printed circuit board is collected; secondly, the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link are determined; then, based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link is compared and analyzed, and the copper embedding quality analysis data of the target printed circuit board is output. Based on the above content, the problem of relatively low reliability of copper embedding quality analysis of printed circuit boards existing in the prior art can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of embedded copper quality analysis, and in particular to a method and system for analyzing the quality of embedded copper on a printed circuit board. Background Art

[0002] At present, in order to enhance the heat dissipation performance of printed circuit boards, copper blocks are embedded in the printed circuit boards. The specific copper embedding process generally includes opening square holes, embedding copper blocks in the square holes, injecting resin, or directly pressing the copper blocks into the square holes of the printed circuit boards. This method is more efficient, but in order to ensure the stability of the copper blocks, the size of the copper blocks used is larger than the size of the square holes, which is easy to cause damage to the printed circuit boards. Therefore, it is necessary to analyze and evaluate the quality of the copper-embedded printed circuit boards. However, since the existing technology generally performs the corresponding inspection manually, there is not only the problem of low inspection efficiency, but also some hidden defects or small defects are easily overlooked, resulting in the problem of low reliability of the quality analysis of the copper-embedded printed circuit boards. Summary of the invention

[0003] In view of this, the purpose of the present application is to provide a method and system for analyzing the quality of embedded copper in a printed circuit board, so as to improve the problem of relatively low reliability of the embedded copper quality analysis of a printed circuit board existing in the prior art.

[0004] To achieve the above objectives, this application adopts the following technical solutions:

[0005] A method for analyzing the quality of copper embedded in a printed circuit board, comprising:

[0006] The target copper embedding monitoring image data of the target printed circuit board is collected, wherein the target copper embedding monitoring image data includes a copper embedding monitoring image set of the target printed circuit board in a target copper embedding link, the target copper embedding link refers to a copper embedding link other than the last copper embedding link among multiple copper embedding links included in the copper embedding process of the target printed circuit board, and the copper embedding link is formed by segmenting the copper embedding process of the target printed circuit board according to a preset time granularity in a time dimension or according to a preset space granularity in a space dimension;

[0007] Determine a circuit board semantic vector of the target printed circuit board and a copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, wherein the circuit board semantic vector is used to reflect the semantic information of the circuit board included in the target printed circuit board before copper embedding, and the copper embedding monitoring image semantic vector is used to reflect the semantic information of the image of the target printed circuit board during the copper embedding process, and the semantic information includes the semantic information of the circuit board and the semantic information of the embedded copper block;

[0008] Based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link is compared and analyzed, and the copper embedding quality analysis data of the target printed circuit board is output, wherein the copper embedding quality analysis data is used to reflect the quality of the target printed circuit board, and there is a negatively correlated corresponding relationship between the quality and the degree of damage formed on the target printed circuit board during the copper embedding process.

[0009] In a preferred selection of the present application, in the above-mentioned copper embedding quality analysis method for a printed circuit board, the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link is compared and analyzed with the copper embedding monitoring image of the next copper embedding link of the target copper embedding link, and the step of outputting the copper embedding quality analysis data of the target printed circuit board includes:

[0010] Mining the copper embedding link semantic vector of each copper embedding monitoring image according to the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of the copper embedding monitoring image in the copper embedding monitoring image set, wherein the set distribution relationship is used to reflect the collection order information of the corresponding copper embedding monitoring image;

[0011] Mining the copper embedding link depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding link semantic vector;

[0012] Mining the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set according to the circuit board semantic vector corresponding to the target printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set;

[0013] Mining the copper embedding-related depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding-related semantic vector;

[0014] According to the copper embedding link depth vector and the copper embedding related depth vector of the target printed circuit board, a copper embedding monitoring image of a subsequent copper embedding link of the target copper embedding link is compared and analyzed, and copper embedding quality analysis data of the target printed circuit board is output.

[0015] In a preferred embodiment of the present application, in the above-mentioned copper embedded quality analysis method for printed circuit boards, the step of mining the copper embedded link semantic vector of each copper embedded monitoring image based on the copper embedded monitoring image semantic vector of each copper embedded monitoring image and the set distribution relationship of the copper embedded monitoring image in the copper embedded monitoring image set includes:

[0016] Performing vector space mapping processing according to the set distribution relationship of each of the copper-embedded monitoring images in the copper-embedded monitoring image set, and outputting a set distribution semantic vector of each copper-embedded monitoring image;

[0017] Aggregate the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link and the set distribution semantic vector corresponding to each copper embedding monitoring image, and output the copper embedding link semantic vector of each copper embedding monitoring image.

[0018] In a preferred embodiment of the present application, in the above-mentioned copper embedding quality analysis method for printed circuit boards, the step of mining the copper embedding link depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding link semantic vector includes:

[0019] mining a copper-embedded first vector associated with each copper-embedded monitoring image in the copper-embedded monitoring image set according to the copper-embedded link semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set and the second image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set, wherein the second image association information and the first image association information are association relationships between the copper-embedded monitoring images determined based on different dimensions;

[0020] mining out a copper-embedded second vector of each embedded copper monitoring image in the copper-embedded monitoring image set based on the copper-embedded first vector of the last copper-embedded monitoring image in the copper-embedded monitoring image set, the copper-embedded link semantic vector of each embedded copper monitoring image in the copper-embedded monitoring image set, and the second image association information between the embedded copper monitoring images in the copper-embedded monitoring image set;

[0021] Aggregate the copper embedding-related first vector of the last copper embedding monitoring image in the copper embedding monitoring image set and the copper embedding-related second vector of the last copper embedding monitoring image in the copper embedding monitoring image set, and output the copper embedding link depth vector of the target printed circuit board under the target copper embedding link.

[0022] In a preferred embodiment of the present application, in the above-mentioned copper embedded quality analysis method for printed circuit boards, the step of mining the copper embedded first vector associated with each copper embedded monitoring image in the copper embedded monitoring image set based on the copper embedded link semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set and the second image association information between the copper embedded monitoring images in the copper embedded monitoring image set includes:

[0023] Analyze the first image association information between the copper embedded monitoring images in the copper embedded monitoring image set according to the copper embedded link semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set, so as to output the first image association vector corresponding to each copper embedded monitoring image, wherein the first image association information is used to reflect the association relationship between the copper embedded link semantic vectors of the copper embedded monitoring images, and the first image association information is used as a kind of second image association information;

[0024] The first image correlation vector of each copper-embedded monitoring image is respectively focused, and the first copper-embedded correlation vector of each copper-embedded monitoring image in the copper-embedded monitoring image set is output.

[0025] In a preferred embodiment of the present application, in the above-mentioned copper embedded quality analysis method for printed circuit boards, the step of mining the copper embedded second vector associated with each copper embedded monitoring image in the copper embedded monitoring image set based on the copper embedded first vector associated with the last copper embedded monitoring image in the copper embedded monitoring image set, the copper embedded link semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set, and the second image association information between the copper embedded monitoring images in the copper embedded monitoring image set includes:

[0026] Analyze the image second association information between the embedded copper monitoring images in the embedded copper monitoring image set according to the copper-embedded first association vector of the last copper-embedded monitoring image in the copper-embedded monitoring image set and the copper-embedded link semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set, so as to output the image second association vector corresponding to each copper-embedded monitoring image, wherein the image second association information is a kind of second image association information;

[0027] The second image correlation vector of each copper-embedded monitoring image is respectively focused, and the copper-embedded correlation second vector of each copper-embedded monitoring image in the copper-embedded monitoring image set is output.

[0028] In a preferred selection of the present application, in the above-mentioned printed circuit board copper embedded quality analysis method, the step of mining the copper embedded related semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set according to the circuit board semantic vector corresponding to the target printed circuit board, the copper embedded monitoring image semantic vector of each copper embedded monitoring image and the first image association information between the copper embedded monitoring images in the copper embedded monitoring image set includes:

[0029] For each copper embedded monitoring image in the copper embedded monitoring image set, based on the circuit board semantic vector corresponding to the target printed circuit board, cross-focus processing is performed on the copper embedded monitoring image semantic vector of the copper embedded monitoring image, and the copper embedded cross-focus vector of the copper embedded monitoring image is output;

[0030] determining first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set based on an association relationship between copper-embedded cross-focus vectors of the copper-embedded monitoring images in the copper-embedded monitoring image set;

[0031] For each copper-embedded monitoring image in the copper-embedded monitoring image set, based on first image association information between the copper-embedded monitoring image and other copper-embedded monitoring images previously collected in the copper-embedded monitoring image set, weighted sum calculation is performed on copper-embedded cross-focus vectors of other copper-embedded monitoring images previously collected in the copper-embedded monitoring image set to obtain a corresponding weighted copper-embedded cross-focus vector;

[0032] Based on the weighted copper-embedded cross-focusing vector, cross-focusing processing is performed on the copper-embedded monitoring image semantic vector or the copper-embedded cross-focusing vector to output a copper-embedded related semantic vector.

[0033] In a preferred embodiment of the present application, in the above-mentioned copper embedding quality analysis method for printed circuit boards, the step of comparing and analyzing the copper embedding monitoring image of the next copper embedding link of the target copper embedding link based on the copper embedding link depth vector and the copper embedding related depth vector of the target printed circuit board, and outputting the copper embedding quality analysis data of the target printed circuit board includes:

[0034] Aggregating the circuit board semantic vector, the copper embedding link depth vector, the copper embedding related depth vector and the copper embedding monitoring image semantic vector of the last copper embedding monitoring image of the target printed circuit board, and outputting the target copper embedding link semantic vector of the target printed circuit board;

[0035] According to the target copper embedding link semantic vector of the target printed circuit board, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the target printed circuit board is reconstructed through a target image reconstruction network, and the copper embedding reconstructed image of the next copper embedding link of the target copper embedding link is output, and the copper embedding reconstructed image of the next copper embedding link of the target copper embedding link and the copper embedding monitoring image of the next copper embedding link of the target copper embedding link are compared and analyzed, and the copper embedding quality analysis data of the target printed circuit board is output; or, the copper embedding monitoring image semantic vector of the copper embedding monitoring image of the next copper embedding link of the target copper embedding link is determined, and the copper embedding quality analysis processing is performed on the target copper embedding link semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of the copper embedding monitoring image of the next copper embedding link of the target copper embedding link through a target copper embedding quality analysis network, and the copper embedding quality analysis data of the target printed circuit board is output, wherein the target image reconstruction network and the target copper embedding quality analysis network are both trained neural networks.

[0036] In a preferred embodiment of the present application, in the above-mentioned printed circuit board embedded copper quality analysis method, the printed circuit board embedded copper quality analysis method further comprises:

[0037] Determine a circuit board semantic vector of a training printed circuit board and a copper embedding monitoring image semantic vector of each copper embedding monitoring image in a copper embedding monitoring image set of the training printed circuit board in a target copper embedding link;

[0038] Mining out the copper embedding link semantic vector of each copper embedding monitoring image based on the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of each copper embedding monitoring image in the copper embedding monitoring image set, and mining out the copper embedding link depth vector of the training printed circuit board under the target copper embedding link based on the copper embedding link semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set;

[0039] Mining out the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set based on the circuit board semantic vector corresponding to the training printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set; and mining out the copper-embedded related depth vector of the training printed circuit board under the target copper-embedded link based on the copper-embedded related semantic vector of each copper-embedded monitoring image;

[0040] According to the copper embedding link depth vector and the copper embedding related depth vector of the training printed circuit board, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the training printed circuit board is compared and analyzed, and the copper embedding quality analysis data of the training printed circuit board is output;

[0041] Based on the acquired copper embedding quality label data and the copper embedding quality analysis data of the training printed circuit board, the network parameters of the candidate image processing network are updated to form a corresponding target image processing network, wherein the target image processing network belongs to the target image reconstruction network or the target copper embedding quality analysis network.

[0042] The present application also provides a printed circuit board embedded copper quality analysis system, comprising:

[0043] An image data acquisition module, used for acquiring target copper embedding monitoring image data of a target printed circuit board, wherein the target copper embedding monitoring image data includes a copper embedding monitoring image set of the target printed circuit board in a target copper embedding link, wherein the target copper embedding link refers to a copper embedding link other than the last copper embedding link among multiple copper embedding links included in the copper embedding process of the target printed circuit board, and the copper embedding link is formed by segmenting the copper embedding process of the target printed circuit board according to a preset time granularity in a time dimension or according to a preset space granularity in a space dimension;

[0044] A semantic vector mining module, used to determine the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, wherein the circuit board semantic vector is used to reflect the semantic information of the circuit board included in the target printed circuit board before copper embedding, and the copper embedding monitoring image semantic vector is used to reflect the semantic information of the image of the target printed circuit board during the copper embedding process, and the semantic information includes the semantic information of the circuit board and the semantic information of the embedded copper block;

[0045] The copper embedding quality analysis module is used to compare and analyze the copper embedding monitoring image of the next copper embedding link of the target copper embedding link based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, and output the copper embedding quality analysis data of the target printed circuit board, wherein the copper embedding quality analysis data is used to reflect the quality of the target printed circuit board, and there is a negatively correlated correspondence between the quality and the degree of damage formed on the target printed circuit board during the copper embedding process.

[0046] The printed circuit board copper embedding quality analysis method and system provided by the present application firstly collect target copper embedding monitoring image data of a target printed circuit board; secondly, determine the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link; then, based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, compare and analyze the copper embedding monitoring image of the next copper embedding link of the target copper embedding link, and output the copper embedding quality analysis data of the target printed circuit board. Based on the above content, when performing comparative analysis, not only the circuit board semantic vector of the target printed circuit board will be used as a basis, but also the embedded copper monitoring image semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set of the target copper embedded link will be used as a basis, so that the basis for comparative analysis can be richer, and since the analysis object is the next copper embedded link of the target copper embedded link, it has a strong correlation in the copper embedded link. Therefore, it is reliable to use the embedded copper monitoring image semantic vector of the copper embedded monitoring image of the target copper embedded link as a basis. Therefore, the reliability of the analysis can be improved to a certain extent, thereby improving the problem of relatively low reliability of printed circuit board copper embedded quality analysis existing in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings.

[0048] Figure 1 A structural block diagram of an electronic device provided in an embodiment of the present application.

[0049] Figure 2 A schematic diagram of the process flow of a printed circuit board embedded copper quality analysis method provided in an embodiment of the present application.

[0050] Figure 3 A flowchart of a method for determining a copper-embedded related semantic vector provided in an embodiment of the present application.

[0051] Figure 4 A block diagram of a printed circuit board copper embedded quality analysis system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0053] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0054] like Figure 1 As shown, an embodiment of the present application provides an electronic device, wherein the electronic device may include a memory, a processor, and a printed circuit board embedded copper quality analysis system.

[0055] In detail, the memory and the processor are electrically connected directly or indirectly to realize data transmission or interaction. For example, the memory and the processor can be electrically connected through one or more communication buses or signal lines. The printed circuit board copper embedded quality analysis system includes at least one software function module stored in the memory in the form of software or firmware. The processor is used to execute an executable computer program stored in the memory, for example, the software function module and computer program included in the printed circuit board copper embedded quality analysis system, so as to realize the printed circuit board copper embedded quality analysis method provided in the embodiment of the present application.

[0056] Optionally, the memory may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc.

[0057] Optionally, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a system on chip (SoC), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0058] Understandably, Figure 1 The structure shown is for illustration only, and the electronic device may also include Figure 1 More or fewer components as shown, or with Figure 1 The different configurations shown, for example, may also include a communication unit for exchanging information with other devices (such as an image acquisition device).

[0059] Combination Figure 2 The embodiment of the present application also provides a printed circuit board copper embedding quality analysis method applicable to the above electronic device. The method steps defined in the process related to the printed circuit board copper embedding quality analysis method can be implemented by the electronic device.

[0060] The following will Figure 2 The specific process shown is explained in detail.

[0061] Step S110, collecting target copper embedded monitoring image data of the target printed circuit board.

[0062] In an embodiment of the present application, the electronic device can collect target copper embedding monitoring image data of the target printed circuit board, for example, the image of the target printed circuit board in the copper embedding process can be collected through a connected image acquisition device (such as a camera, etc.). The target copper embedding monitoring image data includes a copper embedding monitoring image set of the target printed circuit board in the target copper embedding link, and the target copper embedding link refers to a copper embedding link other than the last copper embedding link among the multiple copper embedding links included in the copper embedding process of the target printed circuit board (for example, when 18 copper embedding links are included, it can be any one of 1-17), and the copper embedding link is formed by dividing the copper embedding process of the target printed circuit board according to a preset time granularity in the time dimension or according to a preset space granularity in the space dimension. For example, if the duration of the entire copper embedding process is 3 minutes, the preset time granularity can be 10 seconds. In this way, in the time dimension, the entire copper embedding process can be divided into 18 copper embedding links, each copper embedding link is 10 seconds, and a corresponding copper embedding monitoring image set includes all images collected within 10 seconds. For example, if the embedding depth of the entire copper embedding process is 1.2mm, the preset spatial granularity can be 0.1mm. In this way, in the spatial dimension, the entire copper embedding process can be divided into 12 copper embedding links, and the embedding depth of each copper embedding link is 0.1mm. The corresponding copper embedding monitoring image set includes all images to reflect the embedding process of 0.1mm. It should be noted that the above parameters are only an example, and other parameters can also be used in other examples. For example, based on the different thickness of the circuit board, the embedding depth can be different, based on the different copper embedding equipment, the corresponding embedding time can also be different, and the corresponding time and space granularity can also be different.

[0063] Step S120, determining a circuit board semantic vector of the target printed circuit board and a copper embedding monitoring image semantic vector of each copper embedding monitoring image in a copper embedding monitoring image set of the target copper embedding link.

[0064] In an embodiment of the present application, after obtaining the target copper embedding monitoring image data, the electronic device can determine the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link. Among them, the circuit board semantic vector is used to reflect the semantic information of the circuit board included in the target printed circuit board before copper embedding, and the copper embedding monitoring image semantic vector is used to reflect the semantic information of the image of the target printed circuit board during the copper embedding process, and the semantic information includes the semantic information of the circuit board and the semantic information of the embedded copper block, that is, it carries the appearance information of the circuit board and the copper block during the continuous embedding process of the copper block. For example, the electronic device can process the image of the circuit board included in the target printed circuit board before copper embedding (such as convolution, pooling and full connection, etc.) through a convolutional neural network to obtain the circuit board semantic vector of the target printed circuit board. Similarly, each copper embedding monitoring image can be processed separately through a convolutional neural network to obtain the corresponding copper embedding monitoring image semantic vector.

[0065] Step S130, based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link is compared and analyzed, and the copper embedding quality analysis data of the target printed circuit board is output.

[0066] In an embodiment of the present application, after obtaining the semantic vector of the circuit board and the semantic vector of the copper embedding monitoring image, the electronic device can compare and analyze the copper embedding monitoring image of the next copper embedding link of the target copper embedding link based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, and output the copper embedding quality analysis data of the target printed circuit board. Among them, the copper embedding quality analysis data is used to reflect the quality of the target printed circuit board, and there is a negative correlation between the quality and the degree of damage formed on the target printed circuit board during the copper embedding process. That is to say, the higher the degree of damage (such as deformation degree, crack size, etc.) formed on the target printed circuit board during the copper embedding process, the lower the corresponding quality, and vice versa, the lower the degree of damage formed on the target printed circuit board during the copper embedding process, the higher the corresponding quality. In addition, the copper embedding quality analysis data can be a continuous parameter, such as 0-100 points, or a segmented parameter, such as extremely high quality, high quality, relatively high quality, general quality, poor quality, poor quality, extremely poor quality, etc.

[0067] Based on the above content, when performing comparative analysis, not only the circuit board semantic vector of the target printed circuit board will be used as a basis, but also the embedded copper monitoring image semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set of the target copper embedded link will be used as a basis, so that the basis for comparative analysis can be richer, and since the analysis object is the next copper embedded link of the target copper embedded link, it has a strong correlation in the copper embedded link. Therefore, it is reliable to use the embedded copper monitoring image semantic vector of the copper embedded monitoring image of the target copper embedded link as a basis. Therefore, the reliability of the analysis can be improved to a certain extent, thereby improving the problem of relatively low reliability of printed circuit board copper embedded quality analysis existing in the prior art.

[0068] It should be noted that for step S130 , the specific method of performing comparative analysis on the copper-embedded monitoring image based on the corresponding semantic vector is not limited and can be selected according to actual needs.

[0069] For example, in an alternative embodiment, in order to improve the reliability of comparative analysis of embedded copper monitoring images, the above-mentioned step S130 may further include step S131, step S132, step S133, step S134 and step S135, and the specific content of each step is described as follows.

[0070] Step S131 , mining the copper embedding link semantic vector of each copper embedding monitoring image according to the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of the copper embedding monitoring image in the copper embedding monitoring image set.

[0071] In the embodiment of the present application, in the first dimension, the copper embedding link semantic vector of each copper embedding monitoring image can be mined based on the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of the copper embedding monitoring image in the copper embedding monitoring image set. The set distribution relationship is used to reflect the acquisition order information of the corresponding copper embedding monitoring image, such as the order of the first frame, the second frame, the third frame, etc. in the copper embedding monitoring image set. In other words, the acquisition order information of the copper embedding monitoring image needs to be integrated into the copper embedding monitoring image semantic vector of the copper embedding monitoring image.

[0072] Step S132: mining the copper embedding link depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding link semantic vector.

[0073] In an embodiment of the present application, after obtaining the copper embedding link semantic vector, the copper embedding link depth vector of the target printed circuit board under the target copper embedding link can be mined based on the copper embedding link semantic vector. In other words, the copper embedding link semantic vector can be deeply mined so that the semantic information of the obtained copper embedding link depth vector can be richer.

[0074] Step S133, mining the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set based on the circuit board semantic vector corresponding to the target printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set.

[0075] In the embodiment of the present application, in the second dimension, the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set can be mined based on the circuit board semantic vector corresponding to the target printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set. In this way, on the basis of the semantic fusion of the acquisition sequence information in the first dimension, the semantic information of the circuit board itself (circuit board semantic vector), the semantic information in the embedding process (copper-embedded monitoring image semantic vector) and the related information in the embedding process (first image association information) are semantically fused in the second dimension, so that the copper-embedded related semantic vector with a different emphasis from the copper-embedded link semantic vector can be obtained.

[0076] Step S134: mining the copper embedding-related depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding-related semantic vector.

[0077] In the implementation of the present application, after obtaining the copper-embedded related semantic information, the copper-embedded related depth vector of the target printed circuit board under the target copper-embedded link can be mined based on the copper-embedded related semantic vector. In other words, the copper-embedded related semantic vector can be deeply mined so that the semantic information of the obtained copper-embedded related depth vector can be richer.

[0078] Step S135 , based on the copper embedding link depth vector and the copper embedding related depth vector of the target printed circuit board, a comparative analysis is performed on the copper embedding monitoring image of the next copper embedding link of the target copper embedding link, and copper embedding quality analysis data of the target printed circuit board is output.

[0079] In an embodiment of the present application, after fusion and deep mining are performed in two dimensions to obtain the corresponding copper embedding link depth vector and copper embedding related depth vector, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link can be compared and analyzed based on the copper embedding link depth vector and copper embedding related depth vector of the target printed circuit board, and the copper embedding quality analysis data of the target printed circuit board can be output. For example, if the target copper embedding link is the first copper embedding link, the copper embedding monitoring image of the second copper embedding link is compared and analyzed, thereby outputting the copper embedding quality analysis data of the target printed circuit board (i.e., obtaining local copper embedding quality analysis data). It should be noted that after obtaining multiple copper embedding quality analysis data, they can be integrated, such as weighted summation.

[0080] It can be understood that, in the above step S131, the specific manner of fusing the image acquisition order information into the embedded copper monitoring image semantic vector is not limited. For example, in an alternative implementation, in order to improve the reliability of fusion, the above step S131 may further include:

[0081] First, vector space mapping processing can be performed according to the set distribution relationship of each of the embedded copper monitoring images in the embedded copper monitoring image set, and the set distribution semantic vector of each embedded copper monitoring image can be output. That is to say, the information describing the set distribution relationship can be mapped to the corresponding vector space to obtain the corresponding set distribution semantic vector. For example, the set distribution semantic vector corresponding to the embedded copper monitoring image A is "first frame image", and the set distribution semantic vector corresponding to the embedded copper monitoring image B is "second frame image". In this way, word embedding processing can be performed on the "first frame image" and the "second frame image" (specifically, it can be implemented through a trained word embedding model) to obtain the corresponding word embedding vector. Then, the word embedding vector can be used as the corresponding set distribution semantic vector. For example, word segmentation processing can be performed on the "first frame image" to obtain "first", "frame", and "image", and then they are mapped to the vector space respectively to obtain the word embedding vector corresponding to "first": [0.418, 0.249, -0.412, ...]; the word embedding vector corresponding to "frame": [0.327, 0.123, -0.678, ...]; the word embedding vector corresponding to "image": [0.123, -0.456, 0.789, ...]; then, the three word embedding vectors can be connected together (such as splicing) to obtain the corresponding set distribution semantic vector: {[0.418, 0.249, -0.412, ...]; [0.327, 0.123, -0.678, ...]; [0.123, -0.456, 0.789, ...]};

[0082] Secondly, the copper embedded monitoring image semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set of the target copper embedded link and the set distribution semantic vector corresponding to each copper embedded monitoring image can be aggregated to output the copper embedded link semantic vector of each copper embedded monitoring image; for example, the copper embedded monitoring image semantic vector and the corresponding set distribution semantic vector can be directly connected (such as splicing) to obtain the corresponding copper embedded link semantic vector; or the set distribution semantic vector can be firstly processed by convolution, pooling, etc., and then the processed vector can be connected with the copper embedded monitoring image semantic vector to obtain the corresponding copper embedded link semantic vector.

[0083] It can be understood that, in the above step S132, the specific method of mining the copper embedding link depth vector under the target copper embedding link is not limited and can be selected according to actual needs.

[0084] For example, in an alternative embodiment, in order to further improve the characterization capability of the mined copper embedded link depth vector, the above-mentioned step S132 may further include step S132a, step S132b and step S132c, and the specific contents of each step are described as follows.

[0085] Step S132a, mining the copper-embedded associated first vector of each embedded copper monitoring image in the copper-embedded monitoring image set according to the copper-embedded link semantic vector of each embedded copper monitoring image in the copper-embedded monitoring image set and the second image association information between the embedded copper monitoring images in the copper-embedded monitoring image set.

[0086] In an embodiment of the present application, in a first aspect, based on the copper embedding link semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set and the second image association information between the copper embedding monitoring images in the copper embedding monitoring image set, the copper embedding association first vector of each copper embedding monitoring image in the copper embedding monitoring image set can be mined. Wherein, the second image association information and the first image association information are association relationships between copper embedding monitoring images determined based on different dimensions, and the specific dimensions can refer to the specific description of the corresponding association relationship in the following text.

[0087] Step S132b, mining the embedded copper-associated second vector of each embedded copper monitoring image in the embedded copper monitoring image set based on the embedded copper-associated first vector of the last embedded copper monitoring image in the embedded copper monitoring image set, the embedded copper link semantic vector of each embedded copper monitoring image in the embedded copper monitoring image set, and the second image association information between the embedded copper monitoring images in the embedded copper monitoring image set.

[0088] In an embodiment of the present application, in a second aspect, the embedded copper-associated second vector of each embedded copper monitoring image in the embedded copper monitoring image set can be mined based on the embedded copper-associated first vector of the last embedded copper monitoring image in the embedded copper monitoring image set, the embedded copper link semantic vector of each embedded copper monitoring image in the embedded copper monitoring image set, and the second image association information between the embedded copper monitoring images in the embedded copper monitoring image set.

[0089] Step S132c, aggregating the copper embedding-related first vector of the last copper embedding monitoring image in the copper embedding monitoring image set and the copper embedding-related second vector of the last copper embedding monitoring image in the copper embedding monitoring image set, and outputting the copper embedding link depth vector of the target printed circuit board under the target copper embedding link.

[0090] In an embodiment of the present application, the copper-embedded first vector associated with the last copper-embedded monitoring image in the copper-embedded monitoring image set and the copper-embedded second vector associated with the last copper-embedded monitoring image in the copper-embedded monitoring image set can be aggregated to output the copper-embedded link depth vector of the target printed circuit board under the target copper-embedded link; illustratively, the copper-embedded first vector and the copper-embedded second vector can be connected together (such as splicing) or superimposed, etc., to obtain the copper-embedded link depth vector of the target printed circuit board under the target copper-embedded link.

[0091] It can be understood that, in the above step S132a, the specific manner of performing association mining based on the second image association information is not limited. For example, in an alternative implementation, in order to improve the reliability of association mining so that the information representation accuracy of the obtained copper-embedded association first vector can be higher, the above step S132a can further include the following contents:

[0092] First, according to the copper-embedded link semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set, the first image association information (a second image association information) between the embedded copper monitoring images in the copper-embedded monitoring image set can be analyzed to output the first image association vector corresponding to each copper-embedded monitoring image, wherein the first image association information is used to reflect the association relationship between the copper-embedded link semantic vectors of the embedded copper monitoring images; illustratively, the cosine similarity between the copper-embedded link semantic vectors can be calculated as the corresponding first image association information, and then, based on each cosine similarity, the corresponding copper-embedded link semantic vectors can be weighted summed, and finally, the corresponding first image association vector is determined based on the weighted summation result. For example, for the copper-embedded monitoring image A, the cosine similarity between the copper-embedded monitoring image A and each other copper-embedded monitoring image can be calculated respectively, and then, based on each cosine similarity, the copper-embedded link semantic vectors of each other copper-embedded monitoring image can be weighted summed, and then, the vector obtained by the weighted summation is connected with the copper-embedded link semantic vector of the copper-embedded monitoring image A to obtain the first image association vector corresponding to the copper-embedded monitoring image A;

[0093] Secondly, the image first association vector of each copper-embedded monitoring image can be focused respectively, and the copper-embedded association first vector of each copper-embedded monitoring image in the copper-embedded monitoring image set can be output. By way of example, for copper-embedded monitoring image A, the product of the image first association vector of copper-embedded monitoring image A and the transposed vector of the image first association vector of copper-embedded monitoring image A can be calculated to obtain a corresponding similarity parameter, and then, based on the similarity parameter, the image first association vector of copper-embedded monitoring image A can be weighted to obtain a corresponding copper-embedded association first vector.

[0094] It can be understood that, in the above step S132b, the specific manner of performing association mining based on the second image association information is not limited. For example, in an alternative implementation, in order to improve the reliability of association mining so that the information representation accuracy of the obtained copper-embedded association second vector can be higher, the above step S132b can further include the following contents:

[0095] First, according to the copper-embedded first vector of the last copper-embedded monitoring image in the copper-embedded monitoring image set and the copper-embedded link semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set, the image second association information (a second image association information) between the embedded copper monitoring images in the copper-embedded monitoring image set can be analyzed to output the image second association vector corresponding to each embedded copper monitoring image; exemplarily, for the embedded copper monitoring image A, the copper-embedded link semantic vector corresponding to the embedded copper monitoring image A can be superimposed with the copper-embedded first vector of the last copper-embedded monitoring image in the copper-embedded monitoring image set to obtain a fusion vector corresponding to the embedded copper monitoring image A, based on which, a fusion vector corresponding to each embedded copper monitoring image can be obtained, and then, the cosine similarity between the fusion vector corresponding to the embedded copper monitoring image A and the fusion vectors corresponding to each other embedded copper monitoring image A can be calculated, and then, based on each cosine similarity, the fusion vectors of each other embedded copper monitoring image can be weighted summed, and then, the vector obtained by the weighted sum is connected with the fusion vector of the embedded copper monitoring image A to obtain the image second association vector corresponding to the embedded copper monitoring image A;

[0096] Secondly, the second image correlation vector of each copper-embedded monitoring image is focused respectively, and the copper-embedded correlation second vector of each copper-embedded monitoring image in the copper-embedded monitoring image set is output; the specific focusing processing method can refer to the processing method of the copper-embedded correlation first vector.

[0097] It can be understood that, in the above step S133, the specific method of mining the copper-embedded related semantic vector based on the first image association information is not limited and can be selected according to actual needs.

[0098] For example, in an alternative implementation, in order to improve the reliability of the fusion mining of relevant information so that the information representation capability of the embedded copper related semantic vector is better, the above step S133 may further include the following contents:

[0099] First, for each copper-embedded monitoring image in the copper-embedded monitoring image set, based on the circuit board semantic vector corresponding to the target printed circuit board, the copper-embedded monitoring image semantic vector of the copper-embedded monitoring image can be cross-focused, and the copper-embedded cross-focus vector of the copper-embedded monitoring image can be output; illustratively, the product between the copper-embedded monitoring image semantic vector of the copper-embedded monitoring image and the transposed vector of the circuit board semantic vector corresponding to the target printed circuit board can be calculated to obtain the corresponding similarity parameter, and then, based on the similarity parameter, the circuit board semantic vector corresponding to the target printed circuit board is weighted to obtain the copper-embedded cross-focus vector of the copper-embedded monitoring image;

[0100] Secondly, based on the correlation relationship between the copper-embedded cross-focus vectors of the copper-embedded monitoring images in the copper-embedded monitoring image set, the first image correlation information between the copper-embedded monitoring images in the copper-embedded monitoring image set can be determined; illustratively, the cosine similarity between the copper-embedded cross-focus vectors can be calculated as the first image correlation information between the corresponding copper-embedded monitoring images;

[0101] Then, for each copper-embedded monitoring image in the copper-embedded monitoring image set, based on the first image association information between the copper-embedded monitoring image and other copper-embedded monitoring images previously collected in the copper-embedded monitoring image set, a weighted sum calculation is performed on the copper-embedded cross-focus vectors of other copper-embedded monitoring images previously collected in the copper-embedded monitoring image set to obtain a corresponding weighted copper-embedded cross-focus vector; illustratively, for the first frame of copper-embedded monitoring image, the copper-embedded cross-focus vector of the first frame of copper-embedded monitoring image can be used as the corresponding weighted copper-embedded cross-focus vector; for each frame of copper-embedded monitoring image other than the first frame of copper-embedded monitoring image, such as the third frame of copper-embedded monitoring image (combined with Figure 3 ), based on the cosine similarity between the third copper-embedded monitoring image and the first copper-embedded monitoring image and the second copper-embedded monitoring image, a weighted sum calculation can be performed on the copper-embedded cross-focus vector of the first copper-embedded monitoring image and the copper-embedded cross-focus vector of the second copper-embedded monitoring image, so that a weighted copper-embedded cross-focus vector corresponding to the third copper-embedded monitoring image can be obtained;

[0102] Finally, based on the weighted copper-embedded cross-focusing vector, the copper-embedded monitoring image semantic vector or the copper-embedded cross-focusing vector (or, the superposition result of the copper-embedded monitoring image semantic vector and the copper-embedded cross-focusing vector) can be cross-focused to output an embedded copper-related semantic vector; in this way, the copper-embedded monitoring image's copper-related semantic vector can be obtained; it can be understood that, in other embodiments, the weighted copper-embedded cross-focusing vector, the copper-embedded monitoring image semantic vector and the copper-embedded cross-focusing vector can also be directly superimposed to obtain the corresponding copper-embedded related semantic vector.

[0103] It can be understood that, in the above step S134, the specific method of mining the copper embedded related depth vector is not limited and can be selected according to actual needs.

[0104] For example, in an alternative implementation, the copper embedding-related semantic vectors of each copper embedding monitoring image may be spliced ​​or superimposed, so that the copper embedding-related depth vector of the target printed circuit board at the target copper embedding link may be obtained.

[0105] For example, in another alternative implementation, the copper-embedded related semantic vectors of each copper-embedded monitoring image can be spliced ​​to obtain a corresponding global spliced ​​semantic vector, and then the global spliced ​​semantic vector can be focused. In this way, the copper-embedded related depth vector of the target printed circuit board under the target copper-embedded link can be obtained.

[0106] It can be understood that, in the above step S135, the specific method of performing comparative analysis on the copper embedding monitoring image of the next copper embedding step is not limited and can be selected according to actual needs.

[0107] For example, in an alternative implementation, in order to reduce the complexity of the neural network to a certain extent, the above step S135 may include the following contents:

[0108] First, the circuit board semantic vector, the copper embedding link depth vector, the copper embedding related depth vector and the copper embedding monitoring image semantic vector of the last copper embedding monitoring image of the target printed circuit board can be aggregated to output the target copper embedding link semantic vector of the target printed circuit board. Exemplarily, the circuit board semantic vector, the copper embedding link depth vector, the copper embedding related depth vector and the copper embedding monitoring image semantic vector of the last copper embedding monitoring image of the target printed circuit board can be superimposed to obtain the corresponding target copper embedding link semantic vector; or, the circuit board semantic vector, the copper embedding link depth vector, the copper embedding related depth vector and the copper embedding monitoring image semantic vector of the last copper embedding monitoring image of the target printed circuit board can be spliced, and then, the spliced ​​result can be convolved, pooled and fully connected to obtain the corresponding target copper embedding link semantic vector;

[0109] Secondly, according to the semantic vector of the target copper embedding link of the target printed circuit board, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the target printed circuit board can be reconstructed through the target image reconstruction network, and the copper embedding reconstructed image of the next copper embedding link of the target copper embedding link can be output (exemplarily, the target image reconstruction network can be an image decoding network, and the various vector mining mentioned above can be implemented based on the corresponding image encoding network), and the copper embedding reconstructed image of the next copper embedding link of the target copper embedding link and the copper embedding monitoring image of the next copper embedding link of the target copper embedding link can be compared and analyzed to output the copper embedding quality analysis data of the target printed circuit board. For example, the copper embedding reconstructed image and the copper embedding monitoring image of the next copper embedding link of the target copper embedding link can be calculated. The similarity between the copper embedding monitoring images is determined, and then the copper embedding quality analysis data is determined based on the similarity. The quality represented by the copper embedding quality analysis data can have a positive correlation with the similarity. It should be noted that the target image reconstruction network can be formed by corresponding training based on the copper embedding image data of the circuit board with higher quality, so that the image reconstructed based on the target image reconstruction network is actually a predicted image of a normal copper embedding process. Therefore, the more similar the copper embedding monitoring image of the next copper embedding link is to the copper embedding reconstruction image, the more normal the copper embedding process is, that is, the greater the probability of no abnormal deformation or cracks; conversely, the greater the difference between the copper embedding monitoring image of the next copper embedding link and the copper embedding reconstruction image, the more abnormal the copper embedding process is, that is, the greater the probability of abnormal deformation or cracks.

[0110] For another example, in another alternative implementation, in order to fully utilize the data processing capability of the neural network, the above step S135 may also include the following contents:

[0111] First, a circuit board semantic vector, a copper embedding link depth vector, a copper embedding related depth vector, and a copper embedding monitoring image semantic vector of a last copper embedding monitoring image of the target printed circuit board may be aggregated to output a target copper embedding link semantic vector of the target printed circuit board;

[0112] Secondly, the copper embedding monitoring image semantic vector of the copper embedding monitoring image of the next copper embedding link of the target copper embedding link can be determined, and, through the target copper embedding quality analysis network, the target copper embedding link semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of the copper embedding monitoring image of the next copper embedding link of the target copper embedding link are subjected to copper embedding quality analysis processing, and the copper embedding quality analysis data of the target printed circuit board is output (for example, the target copper embedding link semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of the copper embedding monitoring image of the next copper embedding link of the target copper embedding link can be spliced ​​to obtain a corresponding splicing vector, and then the splicing vector is fully connected, and then the fully connected processed vector is processed through a corresponding classification function to obtain copper embedding quality analysis data), wherein the target image reconstruction network and the target copper embedding quality analysis network are both trained neural networks.

[0113] It can be understood that in order to enable the above-mentioned target image reconstruction network and the target copper embedding quality analysis network to have reliable image reconstruction or quality analysis capabilities, the printed circuit board copper embedding quality analysis method may further include the following contents:

[0114] First, the circuit board semantic vector of the training printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the training printed circuit board in the target copper embedding link may be determined, as described in the aforementioned step S120;

[0115] Secondly, the copper embedding link semantic vector of each copper embedding monitoring image can be mined out according to the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of each copper embedding monitoring image in the copper embedding monitoring image set, and the copper embedding link depth vector of the training printed circuit board under the target copper embedding link can be mined out according to the copper embedding link semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set, as described in the aforementioned steps S131 and S132;

[0116] Then, based on the circuit board semantic vector corresponding to the training printed circuit board, the copper embedded monitoring image semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set, and the first image association information between the copper embedded monitoring images in the copper embedded monitoring image set, the copper embedded related semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set can be mined, and based on the copper embedded related semantic vector of each copper embedded monitoring image, the copper embedded related depth vector of the training printed circuit board under the target copper embedded link can be mined, as described in the aforementioned steps S133 and S134;

[0117] Afterwards, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the training printed circuit board may be compared and analyzed based on the copper embedding link depth vector and the copper embedding related depth vector of the training printed circuit board, and the copper embedding quality analysis data of the training printed circuit board may be output, as in the aforementioned step S135;

[0118] Finally, the network parameters of the candidate image processing network (it should be noted that the determination and mining of the vectors in the aforementioned various steps and the comparative analysis of the embedded copper monitoring images are all implemented by the candidate image processing network) can be updated based on the acquired embedded copper quality label data (i.e., the actual quality) and the embedded copper quality analysis data of the training printed circuit board to form a corresponding target image processing network, wherein the target image processing network belongs to the target image reconstruction network or the target embedded copper quality analysis network; that is, the difference between the embedded copper quality analysis data and the embedded copper quality label data can be calculated as the corresponding network error, and then the network parameters of the candidate image processing network can be updated in the direction of reducing the network error until the network error converges, such as being less than a preset value or the reduction amplitude is less than a preset amplitude.

[0119] It can be understood that, in the above-mentioned training process, in order to reduce the dependence of the neural network on certain specific neurons and reduce the occurrence of overfitting, so that the neural network can better generalize to unknown data (data in actual applications), the copper embedding link depth vector may include a first copper embedding link depth vector and a second copper embedding link depth vector, and the copper embedding-related depth vector may include a first copper embedding-related depth vector and a second copper embedding-related depth vector.

[0120] Based on this, the steps of mining the copper embedding link semantic vector of each copper embedding monitoring image based on the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of each copper embedding monitoring image in the copper embedding monitoring image set, and mining the copper embedding link depth vector of the training printed circuit board under the target copper embedding link based on the copper embedding link semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set, may include:

[0121] Deactivate some parameters in the mining process and output the corresponding first copper-embedded link depth vector and second copper-embedded link depth vector. For example, the first copper-embedded link depth vector may be the copper-embedded link depth vector obtained without deactivation, and the second copper-embedded link depth vector may be the copper-embedded link depth vector obtained after deactivation. For example, when aggregating the copper-embedded first vector associated with the last copper-embedded monitoring image in the copper-embedded monitoring image set and the copper-embedded second vector associated with the last copper-embedded monitoring image in the copper-embedded monitoring image set, one of the copper-embedded first vector and the copper-embedded second vector may be used as the second copper-embedded link depth vector.

[0122] Similarly, the step of mining the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set based on the circuit board semantic vector corresponding to the training printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set, and mining the copper-embedded related depth vector of the training printed circuit board under the target copper-embedded link based on the copper-embedded related semantic vector of each copper-embedded monitoring image, may include:

[0123] Deactivate some parameters in the mining process and output the corresponding first copper-embedded related depth vector and second copper-embedded related depth vector; for example, splice or superimpose the copper-embedded related semantic vectors of each copper-embedded monitoring image to obtain a non-deactivated first copper-embedded related depth vector; and, when splicing or superimposing the copper-embedded related semantic vectors of each copper-embedded monitoring image, hide some of the copper-embedded related semantic vectors (or assign a weight of 0) to obtain a deactivated second copper-embedded related depth vector.

[0124] Similarly, the step of comparing and analyzing the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the training printed circuit board based on the copper embedding link depth vector and the copper embedding related depth vector of the training printed circuit board, and outputting the copper embedding quality analysis data of the training printed circuit board may include:

[0125] First, based on the circuit board semantic vector, the copper embedding monitoring image semantic vector, the first copper embedding link depth vector and the first copper embedding related depth vector of the training printed circuit board, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the training printed circuit board can be compared and analyzed to output the first copper embedding quality analysis data of the training printed circuit board, as described above;

[0126] Secondly, based on the circuit board semantic vector, the copper embedding monitoring image semantic vector, the second copper embedding link depth vector and the second copper embedding related depth vector of the training printed circuit board, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the training printed circuit board can be compared and analyzed to output the second copper embedding quality analysis data of the training printed circuit board, as described above;

[0127] Finally, the final copper embedded quality analysis data of the training printed circuit board can be obtained based on the first copper embedded quality analysis data and the second copper embedded quality analysis data; for example, the first copper embedded quality analysis data and the second copper embedded quality analysis data can be fused, such as weighted superposition, wherein the weighting coefficient corresponding to the first copper embedded quality analysis data can be greater than the weighting coefficient corresponding to the second copper embedded quality analysis data, so as to avoid excessive influence of the deactivation corresponding parameter.

[0128] Combination Figure 4 The embodiment of the present application also provides a printed circuit board copper embedding quality analysis system applicable to the above electronic device. The printed circuit board copper embedding quality analysis system may include an image data acquisition module, a semantic vector mining module and a copper embedding quality analysis module.

[0129] The image data acquisition module is used to acquire target copper embedding monitoring image data of a target printed circuit board, wherein the target copper embedding monitoring image data includes a copper embedding monitoring image set of the target printed circuit board in a target copper embedding link, and the target copper embedding link refers to a copper embedding link other than the last copper embedding link among multiple copper embedding links included in the copper embedding process of the target printed circuit board, and the copper embedding link is formed by segmenting the copper embedding process of the target printed circuit board according to a preset time granularity in the time dimension or according to a preset space granularity in the space dimension. In an embodiment of the present application, the image data acquisition module can be used to perform Figure 2 As shown in step S110, the relevant contents of the image data acquisition module can refer to the above description of step S110.

[0130] The semantic vector mining module is used to determine the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, wherein the circuit board semantic vector is used to reflect the semantic information of the circuit board included in the target printed circuit board before copper embedding, and the copper embedding monitoring image semantic vector is used to reflect the semantic information of the image of the target printed circuit board during the copper embedding process, and the semantic information includes the semantic information of the circuit board and the semantic information of the embedded copper block. In an embodiment of the present application, the semantic vector mining module can be used to perform Figure 2 As shown in step S120, the relevant contents of the semantic vector mining module can refer to the above description of step S120.

[0131] The copper embedding quality analysis module is used to compare and analyze the copper embedding monitoring image of the next copper embedding link of the target copper embedding link based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, and output the copper embedding quality analysis data of the target printed circuit board, wherein the copper embedding quality analysis data is used to reflect the quality of the target printed circuit board, and there is a negative correlation between the quality and the degree of damage formed on the target printed circuit board during the copper embedding process. In an embodiment of the present application, the copper embedding quality analysis module can be used to perform Figure 2 As shown in step S130, the relevant contents of the embedded copper quality analysis module can refer to the above description of step S130.

[0132] In the embodiment of the present application, corresponding to the above-mentioned printed circuit board embedded copper quality analysis method applied to the electronic device, a computer-readable storage medium is also provided, in which a computer program is stored, and when the computer program is run, each step of the printed circuit board embedded copper quality analysis method is executed. Among them, each step executed when the aforementioned computer program is run will not be repeated here one by one, and reference can be made to the explanation of the printed circuit board embedded copper quality analysis method in the previous text.

[0133] In summary, the printed circuit board copper embedding quality analysis method and system provided by the present application first collects the target copper embedding monitoring image data of the target printed circuit board; secondly, determines the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link; then, based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, compares and analyzes the copper embedding monitoring image of the next copper embedding link of the target copper embedding link, and outputs the copper embedding quality analysis data of the target printed circuit board. Based on the above content, when conducting comparative analysis, not only the circuit board semantic vector of the target printed circuit board will be used as a basis, but also the embedded copper monitoring image semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set of the target copper embedded link will be used as a basis, so that the basis for comparative analysis can be richer, and since the analysis object is the next copper embedded link of the target copper embedded link, it has a strong correlation in the copper embedded link. Therefore, it is reliable to use the embedded copper monitoring image semantic vector of the copper embedded monitoring image of the target copper embedded link as a basis. Therefore, the reliability of the analysis can be improved to a certain extent, and the problem of relatively low reliability of printed circuit board copper embedded quality analysis existing in the prior art can be improved.

[0134] In several embodiments provided in the embodiments of the present application, it should be understood that the disclosed device and method can also be implemented in other ways. The device and method embodiments described above are merely schematic, for example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the device, method and computer program product according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0135] In addition, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0136] If the function is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, electronic device, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code. It should be noted that in this article, the term "include", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements that are not explicitly listed, or also includes elements inherent to such process, method, article or device. Without more constraints, an element defined by the phrase "comprising a..." does not exclude the existence of other identical elements in the process, method, article or apparatus comprising the element.

[0137] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for analyzing the quality of copper embedded in a printed circuit board, characterized in that: include: The target copper embedding monitoring image data of the target printed circuit board is collected, wherein the target copper embedding monitoring image data includes a copper embedding monitoring image set of the target printed circuit board in a target copper embedding link, the target copper embedding link refers to a copper embedding link other than the last copper embedding link among multiple copper embedding links included in the copper embedding process of the target printed circuit board, and the copper embedding link is formed by segmenting the copper embedding process of the target printed circuit board according to a preset time granularity in a time dimension or according to a preset space granularity in a space dimension; Determine a circuit board semantic vector of the target printed circuit board and a copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, wherein the circuit board semantic vector is used to reflect the semantic information of the circuit board included in the target printed circuit board before copper embedding, and the copper embedding monitoring image semantic vector is used to reflect the semantic information of the image of the target printed circuit board during the copper embedding process, and the semantic information includes the semantic information of the circuit board and the semantic information of the embedded copper block; Based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, a copper embedding monitoring image of a subsequent copper embedding link of the target copper embedding link is compared and analyzed, and copper embedding quality analysis data of the target printed circuit board is output, wherein the copper embedding quality analysis data is used to reflect the quality of the target printed circuit board, and there is a negatively correlated corresponding relationship between the quality and the degree of damage formed on the target printed circuit board during the copper embedding process; Among them, the step of comparing and analyzing the copper embedding monitoring image of the next copper embedding link of the target copper embedding link based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, and outputting the copper embedding quality analysis data of the target printed circuit board includes: Mining the copper embedding link semantic vector of each copper embedding monitoring image according to the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of the copper embedding monitoring image in the copper embedding monitoring image set, wherein the set distribution relationship is used to reflect the collection order information of the corresponding copper embedding monitoring image; Mining the copper embedding link depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding link semantic vector; Mining the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set according to the circuit board semantic vector corresponding to the target printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set; Mining the copper embedding-related depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding-related semantic vector; According to the copper embedding link depth vector and the copper embedding related depth vector of the target printed circuit board, a copper embedding monitoring image of a subsequent copper embedding link of the target copper embedding link is compared and analyzed, and copper embedding quality analysis data of the target printed circuit board is output.

2. The method for analyzing the quality of embedded copper in printed circuit boards according to claim 1, characterized in that: The step of mining the copper embedding link semantic vector of each copper embedding monitoring image according to the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of the copper embedding monitoring image in the copper embedding monitoring image set comprises: Performing vector space mapping processing according to the set distribution relationship of each of the copper-embedded monitoring images in the copper-embedded monitoring image set, and outputting a set distribution semantic vector of each copper-embedded monitoring image; Aggregate the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link and the set distribution semantic vector corresponding to each copper embedding monitoring image, and output the copper embedding link semantic vector of each copper embedding monitoring image.

3. The method for analyzing the quality of embedded copper in printed circuit boards according to claim 1, characterized in that: The step of mining the copper embedding link depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding link semantic vector comprises: mining a copper-embedded first vector associated with each copper-embedded monitoring image in the copper-embedded monitoring image set according to the copper-embedded link semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set and the second image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set, wherein the second image association information and the first image association information are association relationships between the copper-embedded monitoring images determined based on different dimensions; mining out a copper-embedded second vector of each embedded copper monitoring image in the copper-embedded monitoring image set based on the copper-embedded first vector of the last copper-embedded monitoring image in the copper-embedded monitoring image set, the copper-embedded link semantic vector of each embedded copper monitoring image in the copper-embedded monitoring image set, and the second image association information between the embedded copper monitoring images in the copper-embedded monitoring image set; Aggregate the copper embedding-related first vector of the last copper embedding monitoring image in the copper embedding monitoring image set and the copper embedding-related second vector of the last copper embedding monitoring image in the copper embedding monitoring image set, and output the copper embedding link depth vector of the target printed circuit board under the target copper embedding link.

4. The method for analyzing the quality of embedded copper in printed circuit boards according to claim 3, characterized in that: The step of mining the copper-embedded first vector associated with each copper-embedded monitoring image in the copper-embedded monitoring image set according to the copper-embedded link semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set and the second image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set comprises: Analyze the first image association information between the copper embedded monitoring images in the copper embedded monitoring image set according to the copper embedded link semantic vector of each copper embedded monitoring image in the copper embedded monitoring image set, so as to output the first image association vector corresponding to each copper embedded monitoring image, wherein the first image association information is used to reflect the association relationship between the copper embedded link semantic vectors of the copper embedded monitoring images, and the first image association information is used as a kind of second image association information; The first image correlation vector of each copper-embedded monitoring image is respectively focused, and the first copper-embedded correlation vector of each copper-embedded monitoring image in the copper-embedded monitoring image set is output.

5. The method for analyzing the quality of embedded copper in printed circuit boards according to claim 3, characterized in that: The step of mining the copper-embedded second vector of each embedded copper monitoring image in the embedded copper monitoring image set based on the copper-embedded first vector of the last embedded copper monitoring image in the embedded copper monitoring image set, the copper-embedded link semantic vector of each embedded copper monitoring image in the embedded copper monitoring image set, and the second image association information between the embedded copper monitoring images in the embedded copper monitoring image set comprises: Analyze the image second association information between the embedded copper monitoring images in the embedded copper monitoring image set according to the copper-embedded first association vector of the last copper-embedded monitoring image in the copper-embedded monitoring image set and the copper-embedded link semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set, so as to output the image second association vector corresponding to each copper-embedded monitoring image, wherein the image second association information is a kind of second image association information; The second image correlation vector of each copper-embedded monitoring image is respectively focused, and the copper-embedded correlation second vector of each copper-embedded monitoring image in the copper-embedded monitoring image set is output.

6. The method for analyzing the quality of embedded copper in printed circuit boards according to claim 1, characterized in that: The step of mining the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set according to the circuit board semantic vector corresponding to the target printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set comprises: For each copper embedded monitoring image in the copper embedded monitoring image set, based on the circuit board semantic vector corresponding to the target printed circuit board, cross-focus processing is performed on the copper embedded monitoring image semantic vector of the copper embedded monitoring image, and the copper embedded cross-focus vector of the copper embedded monitoring image is output; determining first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set based on an association relationship between copper-embedded cross-focus vectors of the copper-embedded monitoring images in the copper-embedded monitoring image set; For each copper-embedded monitoring image in the copper-embedded monitoring image set, based on first image association information between the copper-embedded monitoring image and other copper-embedded monitoring images previously collected in the copper-embedded monitoring image set, weighted sum calculation is performed on copper-embedded cross-focus vectors of other copper-embedded monitoring images previously collected in the copper-embedded monitoring image set to obtain a corresponding weighted copper-embedded cross-focus vector; Based on the weighted copper-embedded cross-focusing vector, cross-focusing processing is performed on the copper-embedded monitoring image semantic vector or the copper-embedded cross-focusing vector to output a copper-embedded related semantic vector.

7. The method for analyzing the quality of embedded copper in printed circuit boards according to any one of claims 1 to 6, characterized in that: The step of comparing and analyzing the copper embedding monitoring image of a subsequent copper embedding link of the target copper embedding link based on the copper embedding link depth vector and the copper embedding related depth vector of the target printed circuit board, and outputting the copper embedding quality analysis data of the target printed circuit board comprises: Aggregating the circuit board semantic vector, the copper embedding link depth vector, the copper embedding related depth vector and the copper embedding monitoring image semantic vector of the last copper embedding monitoring image of the target printed circuit board, and outputting the target copper embedding link semantic vector of the target printed circuit board; According to the target copper embedding link semantic vector of the target printed circuit board, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the target printed circuit board is reconstructed through a target image reconstruction network, and the copper embedding reconstructed image of the next copper embedding link of the target copper embedding link is output, and the copper embedding reconstructed image of the next copper embedding link of the target copper embedding link and the copper embedding monitoring image of the next copper embedding link of the target copper embedding link are compared and analyzed, and the copper embedding quality analysis data of the target printed circuit board is output; or, the copper embedding monitoring image semantic vector of the copper embedding monitoring image of the next copper embedding link of the target copper embedding link is determined, and the copper embedding quality analysis processing is performed on the target copper embedding link semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of the copper embedding monitoring image of the next copper embedding link of the target copper embedding link through a target copper embedding quality analysis network, and the copper embedding quality analysis data of the target printed circuit board is output, wherein the target image reconstruction network and the target copper embedding quality analysis network are both trained neural networks.

8. The method for analyzing the quality of embedded copper in printed circuit boards according to claim 7, characterized in that: The printed circuit board copper embedded quality analysis method also includes: Determine a circuit board semantic vector of a training printed circuit board and a copper embedding monitoring image semantic vector of each copper embedding monitoring image in a copper embedding monitoring image set of the training printed circuit board in a target copper embedding link; Mining out the copper embedding link semantic vector of each copper embedding monitoring image based on the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of each copper embedding monitoring image in the copper embedding monitoring image set, and mining out the copper embedding link depth vector of the training printed circuit board under the target copper embedding link based on the copper embedding link semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set; Mining out the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set based on the circuit board semantic vector corresponding to the training printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set; and mining out the copper-embedded related depth vector of the training printed circuit board under the target copper-embedded link based on the copper-embedded related semantic vector of each copper-embedded monitoring image; According to the copper embedding link depth vector and the copper embedding related depth vector of the training printed circuit board, the copper embedding monitoring image of the next copper embedding link of the target copper embedding link of the training printed circuit board is compared and analyzed, and the copper embedding quality analysis data of the training printed circuit board is output; Based on the acquired copper embedding quality label data and the copper embedding quality analysis data of the training printed circuit board, the network parameters of the candidate image processing network are updated to form a corresponding target image processing network, wherein the target image processing network belongs to the target image reconstruction network or the target copper embedding quality analysis network.

9. A printed circuit board copper embedded quality analysis system, characterized in that: include: An image data acquisition module, used for acquiring target copper embedding monitoring image data of a target printed circuit board, wherein the target copper embedding monitoring image data includes a copper embedding monitoring image set of the target printed circuit board in a target copper embedding link, wherein the target copper embedding link refers to a copper embedding link other than the last copper embedding link among multiple copper embedding links included in the copper embedding process of the target printed circuit board, and the copper embedding link is formed by segmenting the copper embedding process of the target printed circuit board according to a preset time granularity in a time dimension or according to a preset space granularity in a space dimension; A semantic vector mining module, used to determine the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, wherein the circuit board semantic vector is used to reflect the semantic information of the circuit board included in the target printed circuit board before copper embedding, and the copper embedding monitoring image semantic vector is used to reflect the semantic information of the image of the target printed circuit board during the copper embedding process, and the semantic information includes the semantic information of the circuit board and the semantic information of the embedded copper block; A copper embedding quality analysis module, configured to compare and analyze the copper embedding monitoring image of the next copper embedding link of the target copper embedding link based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link, and output copper embedding quality analysis data of the target printed circuit board, wherein the copper embedding quality analysis data is used to reflect the quality of the target printed circuit board, and there is a negatively correlated corresponding relationship between the quality and the degree of damage formed on the target printed circuit board during the copper embedding process; Among them, the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the copper embedding monitoring image set of the target copper embedding link is based on the circuit board semantic vector of the target printed circuit board and the copper embedding monitoring image semantic vector of each copper embedding monitoring image in the target copper embedding link, and the copper embedding monitoring image of the next copper embedding link of the target copper embedding link is compared and analyzed to output the copper embedding quality analysis data of the target printed circuit board, including: Mining the copper embedding link semantic vector of each copper embedding monitoring image according to the copper embedding monitoring image semantic vector of each copper embedding monitoring image and the set distribution relationship of the copper embedding monitoring image in the copper embedding monitoring image set, wherein the set distribution relationship is used to reflect the collection order information of the corresponding copper embedding monitoring image; Mining the copper embedding link depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding link semantic vector; Mining the copper-embedded related semantic vector of each copper-embedded monitoring image in the copper-embedded monitoring image set according to the circuit board semantic vector corresponding to the target printed circuit board, the copper-embedded monitoring image semantic vector of each copper-embedded monitoring image, and the first image association information between the copper-embedded monitoring images in the copper-embedded monitoring image set; Mining the copper embedding-related depth vector of the target printed circuit board under the target copper embedding link according to the copper embedding-related semantic vector; According to the copper embedding link depth vector and the copper embedding related depth vector of the target printed circuit board, a copper embedding monitoring image of a subsequent copper embedding link of the target copper embedding link is compared and analyzed, and copper embedding quality analysis data of the target printed circuit board is output.