CN116152166A - Defect detection method and related device based on feature correlation - Google Patents

Abstract

The application provides a defect detection method and a related device based on feature correlation. The method comprises the following steps: acquiring a circuit board image to be tested and a sample circuit board image; extracting first characteristic information of a circuit board image to be detected and second characteristic information of a sample circuit board image; performing defect detection on the circuit board image to be detected based on the correlation characteristic between the first characteristic information and the second characteristic information to obtain a defect detection result; the correlation features are used for representing the same degree of corresponding same areas between the circuit board image to be tested and the sample circuit board image. By adopting the method and the device, the accuracy of PCB defect detection is improved.

Description

Defect detection method and related device based on feature correlation

technical field

The present application relates to the technical field of image processing, in particular to a defect detection method and related devices based on feature correlation.

Background technique

With the wide application of automotive electronics, communication equipment, transformers, inductance devices, and power modules in daily life, as well as the rapid development of electronic information technology and communication technology, the market has put forward higher requirements for high-transmission, high-voltage electronic products. . As the basic bearing part of electronic products - printed circuit board (Printed circuit board, referred to as PCB), its performance directly affects the performance of corresponding electronic products.

Because PCBs inevitably have a large number of defects during the production process, and the existing defects are mainly located at the circuit components of the PCB. Therefore, it is necessary to detect defective circuit components for subsequent PCB repair work. At present, the defect detection of PCB is mainly through the automatic optical inspection (Automated Optical Inspection, AOI) equipment to use the optical camera to collect the image of the PCB, and use image processing, machine learning and other methods to detect and locate the defects on the PCB.

However, due to the detection accuracy of AOI detection PCB and the shape, quantity and type of defects on the PCB and other influencing factors, defects in some similar areas or unknown types of defects cannot be detected, so that the accuracy of PCB defect detection is not high. high.

Contents of the invention

Based on this, it is necessary to provide a defect detection method based on feature correlation and a related device for the above technical problems, which can improve the accuracy of defect detection on PCBs.

In the first aspect, the present application provides a defect detection method based on feature correlation, including:

Obtain the image of the circuit board to be tested and the image of the sample circuit board;

extracting the first characteristic information of the circuit board image to be tested and the second characteristic information of the sample circuit board image;

Based on the correlation feature between the first feature information and the second feature information, defect detection is performed on the image of the circuit board to be tested to obtain a defect detection result; wherein, the correlation feature is used to characterize the relationship between the image of the circuit board to be tested and the image of the sample circuit board between corresponding to the same extent in the same area.

In the second aspect, the present application also provides a defect detection device based on feature correlation, including:

An acquisition unit, configured to acquire the image of the circuit board to be tested and the image of the sample circuit board;

An extraction unit, configured to extract the first feature information of the circuit board image to be tested and the second feature information of the sample circuit board image;

The detection unit is configured to perform defect detection on the image of the circuit board to be tested based on the correlation feature between the first feature information and the second feature information to obtain a defect detection result; wherein the correlation feature is used to characterize the image of the circuit board to be tested and The same degree of correspondence to the same area between sample board images.

In a third aspect, the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the defect detection method based on feature correlation as described above is implemented.

In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned defect detection method based on feature correlation is implemented.

In a fifth aspect, the present application further provides a computer program product, the computer program product includes a computer program, and when the computer program is executed by a processor, the above-mentioned defect detection method based on feature correlation is implemented.

The above-mentioned defect detection method and related devices based on feature correlation, on the one hand, only use the feature information of the circuit board image to be tested and the sample circuit board image to determine the same degree of corresponding regions between the two, so that it is easy to carry out subsequent PCB Defect detection process to improve the efficiency of defect detection and ensure the real-time detection; Defects can improve the accuracy of PCB defect detection and ensure a high recall rate of circuit boards.

Description of drawings

FIG. 1 is an application environment diagram of a defect detection method based on feature correlation provided by an embodiment of the present application;

Fig. 2 is a schematic flow chart of the first defect detection method based on feature correlation provided by the embodiment of the present application;

Fig. 3 is a schematic flow chart of defect detection on the image of the circuit board to be tested provided by the embodiment of the present application;

Fig. 4 is a schematic flowchart of determining the correlation feature between the first characteristic information and the second characteristic information provided by the embodiment of the present application;

FIG. 5 is a schematic flow diagram of defect detection on a scaled circuit board image to be tested provided by an embodiment of the present application;

FIG. 6 is a schematic flow diagram of defect identification and defect labeling for a scaled image of a circuit board to be tested provided by an embodiment of the present application;

FIG. 7 is a schematic flowchart of a second defect detection method based on feature correlation provided by an embodiment of the present application;

FIG. 8 is a schematic flow diagram of updating a defective circuit board provided by an embodiment of the present application;

FIG. 9 is a schematic flowchart of a third defect detection method based on feature correlation provided by an embodiment of the present application;

Fig. 10 is a block diagram of a defect detection device based on feature correlation provided by an embodiment of the present application;

Fig. 11 is a block diagram of an electronic device provided by an embodiment of the present application;

Fig. 12 is a block diagram of a computer-readable storage medium provided by an embodiment of the present application;

Fig. 13 is a block diagram of a computer program product provided by an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application.

The term "and/or" in the embodiments of the present application refers to any and all possible combinations of one or more of the associated listed items. It should also be noted that: when used in this specification, "comprises/comprises" specifies the presence of stated features, integers, steps, operations, elements and/or components, but does not exclude one or more other features, integers , the presence or addition of steps, operations, elements and/or components and/or groups thereof.

The terms "first", "second", etc. in this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "include" and "have", as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but optionally also includes unlisted steps or units, or optionally further includes For other steps or units inherent in these processes, methods, products or apparatuses.

In addition, although the terms "first", "second" and the like are used many times in this application to describe various operations (or various elements or various applications or various instructions or various data), etc., these operations (or various element or application or instruction or data) should not be limited by these terms. These terms are only used to distinguish one operation (or element or application or instruction or data) from another operation (or element or application or instruction or data). For example, the pixel features of the first quantity can be called the pixel features of the second quantity, and the pixel features of the second quantity can also be called the pixel features of the first quantity, only the scopes of the two are different, and not Out of the scope of the present application, the first number of pixel features and the second number of pixel features are sets of corresponding numbers of pixel features on various circuit boards, but the two are not the corresponding number of pixel features on the same circuit board. Collection only.

The defect detection method based on feature correlation provided in the embodiment of the present application can be applied to the application environment shown in FIG. 1 . Wherein, the electronic device 102 communicates with the server 104 through a communication network. The data storage system can store data that needs to be processed by the server 104 . The data storage system can be integrated on the server 104, or placed on the cloud or other network servers.

In some embodiments, referring to FIG. 1, the server 104 first obtains the circuit board image to be tested and the sample circuit board image; then, extracts the first feature information of the circuit board image to be tested and the second feature information of the sample circuit board image; Finally, based on the correlation feature between the first feature information and the second feature information, the defect detection is performed on the image of the circuit board to be tested, and the defect detection result is obtained; where the correlation feature is used to characterize the image of the circuit board to be tested and the sample circuit The same degree of correspondence to the same region between plate images.

In some embodiments, the electronic device 102 (such as a mobile terminal, a fixed terminal) may be implemented in various forms. Among them, the electronic device 102 may include features such as mobile phones, smart phones, notebook computers, portable handheld devices, personal digital assistants (PDA, Personal Digital Assistant), tablet computers (PAD), etc., which may be based on at least two images. Corresponding correlation features between information, the mobile terminal for detecting defects on the circuit board image, and the electronic device 102 can also be an automatic teller machine (Automated Teller Machine, ATM), an access control machine, a digital TV, a desktop computer, a solid-state computer, etc. A fixed terminal capable of detecting defects on circuit board images based on the correlation features corresponding to feature information of at least two images. In the following, it is assumed that the electronic device 102 is a fixed terminal. However, those skilled in the art will understand that the configuration according to the embodiments disclosed in the present application can also be applied to mobile-type electronic devices 102 if there are operations or elements specifically for mobile purposes.

In some embodiments, the image processing component and the data processing component run by the server 104 can load any one of various additional server applications and/or middle-tier applications that are being executed, such as HTTP (Hypertext Transfer Protocol ), FTP (File Transfer Protocol), CGI (Common Gateway Interface), RDBMS (Relational Database Management System), etc.

In some embodiments, the server 104 may be implemented by an independent server or a server cluster composed of multiple servers. The server 104 may be adapted to run one or more application services or software components that provide the electronic device 102 described in the foregoing disclosure.

In some embodiments, the application service may include, for example, a service for extracting the feature information of the circuit board image to be tested and the feature information of the sample circuit board image, and after extracting the feature information of the image, provide the user with a subsequent circuit board image Services for defect detection, etc. The software component may include, for example, an APP or a client with a defect detection function on a circuit board image.

In some embodiments, the APP or client with the function of detecting defects on circuit board images includes a portal port that provides one-to-one application services to users in the foreground and a plurality of business systems that perform data processing in the background to detect defects. The application of the detection function is extended to APP or client, so that users can use and access the defect detection function anytime and anywhere.

In some embodiments, the defect detection function of the APP or the client can be a computer program running in the user mode to complete one or more specific tasks, which can interact with the user and has a visual user interface. Wherein, the APP or the client may include two parts: a graphical user interface (GUI) and an engine (engine), which are digital client systems that can provide users with various application services in the form of a user interface.

In some embodiments, the user can input corresponding code data or control parameters to the APP or the client through the input device, so as to execute the application service of the computer program and display the application service in the user interface. For example, when the feature information of the circuit board image to be tested and the feature information of the sample circuit board image need to be extracted, the user operates through the input device and displays through the user interface. Optionally, the input device may be touch screen input, key input, voice input or pupil focus input, and the like.

Apple/>

和/或Linux操作系统、各种商用或类/>

操作系统(包括但不限于各种GNU/Linux操作系统、Google/>

OS等)和/或移动操作系统，诸如

Phone、/>

OS、/>

OS、/>

OS操作系统，以及其它在线操作系统或者离线操作系统。In some embodiments, the operating system run by APP or client may include various versions of Microsoft

Apple/>

and/or Linux operating systems, various commercial or class/>

Operating system (including but not limited to various GNU/Linux operating systems, Google/>

OS, etc.) and/or mobile operating systems such as

Phone, />

OS, />

OS, />

OS operating system, and other online operating systems or offline operating systems.

In some embodiments, as shown in FIG. 2 , a first defect detection method based on feature correlation is provided. The application of the method to the server 104 in FIG. 1 is used as an example for illustration. The method includes the following steps:

Step S11, acquiring the image of the circuit board to be tested and the image of the sample circuit board.

In some embodiments, the user collects the captured images of the circuit board to be tested and the sample circuit board in real time through the camera device in the electronic device, and then the electronic device sends the collected captured images to the server for subsequent data processing.

In some embodiments, the photographed image of the circuit board to be tested and the photographed image of the sample circuit board have been pre-photographed by a camera or other device in the electronic device and stored in a third-party organization (such as an image database, a cloud storage platform) etc.), when the server responds to receiving an instruction selected by the user to start acquiring the captured image, the server acquires the image of the circuit board to be tested and the image of the sample circuit board from the corresponding third-party organization.

In some embodiments, the camera device in the electronic device is an automatic optical inspection device (Automated Optical Inspection, AOI) or an automatic visual inspection (automated vision inspection, AVI), wherein, the AOI or AVI integrates image sensing technology, data Processing technology, motion control technology, which is based on optical principles to detect common defects encountered in PCB circuit board welding production. When automatic detection, the AOI machine or AVI machine automatically scans the PCB through the camera, and collects the captured image of the circuit board to be tested and the captured image of the sample circuit board.

In some embodiments, the AOI machine or AVI machine itself can be equipped with an image acquisition device of one of a depth camera, a 3D camera, a monocular camera or a binocular camera, and generate corresponding control information according to user input, to collect the photographed image of the circuit board to be tested and the photographed image of the sample circuit board.

In some embodiments, the captured image of the circuit board to be tested and the captured image of the sample circuit board are PCB (Printed circuit boards, printed circuit board) images. Among them, the PCB board surface is divided into a circuit board area and a non-circuit board area.

In some embodiments, there are PCB lines etched by chemicals on the circuit board area of the PCB, and on the PCB lines are small and dense circuit elements and their sharp corners (that is, the sharp corner areas extend from the four right angles of the square area to The two adjacent straight lines in the square area form oblique lines that intersect each other at the ends that are close to each other, and two adjacent oblique lines intersect each other to form sharp angles; the oblique lines are called sharp angle segments). Due to the tension of the agent, there are inevitably a large number of defects in the PCB during the production process (due to the manufacturing process, in the production process of the circuit board, it is inevitable that there will be holes, mouse erosion, open circuit, short circuit, etc. burrs, copper slag and other defects), and the sharp corners on the PCB line are often false sharp corners (due to the shiny holes of the sharp corners, the reflection of the substrate, local oxidation and dirty spots, etc. more false points). Therefore, the image of the circuit board to be tested collected by the AOI machine or the AVI machine is a PCB image with various defects and sharp corners.

Among them, the sample circuit board image is the schematic design drawing corresponding to the PCB. Compared with the PCB image, there are no defects and sharp corners on the board surface, and there are no false points and sharp corners. It is the most ideal PCB board.

Step S12, extracting the first characteristic information of the circuit board image to be tested and the second characteristic information of the sample circuit board image.

In some embodiments, the server extracts the feature information of the image from the collected image of the circuit board to be tested through a preset neural network, and extracts the feature information of the image from the image of the sample circuit board.

In some embodiments, the first feature information of the circuit board image to be tested and the second feature information of the sample circuit board image include information such as color features, texture features, shape features, and spatial relationship features of the images.

In some embodiments, the server first crops the image of the circuit board to be tested and the image of the sample circuit board, and then uses the Harris corner detection method to extract all the corner points in the image of the circuit board to be tested and the image of the sample circuit board, and finally, the server Feature analysis is performed on the extracted corner points through a preset neural network (eg, convolutional neural network, semantic segmentation network, etc.) to extract feature information.

In some embodiments, the server may input the image of the circuit board to be tested and the image of the sample circuit board into a pre-trained convolutional neural network model (for example, a UNet network model, a CNN network model, a RNN network model, etc.), so as to obtain The feature information of the circuit board image to be tested and the sample circuit board image is directly obtained in the convolutional neural network model.

As an example, the pre-trained convolutional neural network model is a UNet convolutional neural network. Among them, the UNet convolutional neural network can be divided into two parts, one is the feature extraction part, which, like other convolutional neural networks, extracts image features through stacked convolutions, and compresses feature maps through pooling. The other part is the image restoration part, which restores the compressed image through upsampling and convolution. Among them, the feature extraction part can use the residual network structure, such as: Resnet50, VGG, etc. Among them, the sample balance loss function (that is, focal loss) can also be used as the loss function of the neural network to balance positive and negative examples, and finally a stable semantic segmentation model that can output feature information without reducing the loss function is obtained.

Step S13 , based on the correlation feature between the first feature information and the second feature information, perform defect detection on the image of the circuit board to be tested to obtain a defect detection result.

In some embodiments, the correlation feature is used to characterize the degree to which the image of the circuit board under test and the image of the sample circuit board correspond to the same region.

In some embodiments, the server first determines, based on the first feature information of the image of the circuit board to be tested and the second feature information of the image of the sample circuit board, the correlation between the image of the circuit board to be tested and the image of the sample circuit board corresponding to the same areas. Then, the server determines the defect information on the image of the circuit board to be tested according to the correlation characteristics of each same area, so as to complete the defect detection of the image of the circuit board to be tested.

In some embodiments, the correlation feature may be the feature similarity between the image of the circuit board to be tested and the image of the sample circuit board corresponding to each of the same regions. For example, the similarity of texture features, the similarity of color features, the similarity of shape features and so on.

In some embodiments, the server compares the similarity of features corresponding to the same regions between the image of the circuit board to be tested and the image of the sample circuit board with a preset similarity threshold, and if the similarity of the features is less than the similarity threshold, it indicates that the The image of the circuit board to be tested in the corresponding area includes defects, and if the feature similarity is greater than or equal to a similarity threshold, it indicates that there is no defect in the image of the circuit board to be tested in the corresponding area.

In the above-mentioned defect detection method based on feature correlation, on the one hand, only the feature information of the circuit board image to be tested and the sample circuit board image is used to determine the same degree of corresponding regions between the two, so that subsequent defect detection can be easily performed process to improve the efficiency of defect detection and ensure real-time detection; on the other hand, using the same degree of corresponding areas between the image of the circuit board to be tested and the image of the sample circuit board to detect defects in the image of the circuit board to be tested can be Improve the accuracy of defect detection and ensure a high recall rate of circuit boards.

Those skilled in the art can understand that among the above-mentioned methods in specific implementation manners, the disclosed methods can be implemented in more specific ways. For example, the implementation of the above-mentioned defect detection method based on feature correlation is only a schematic description.

Exemplarily, the process of extracting the feature information of the circuit board image under test and the feature information of the sample circuit board image, etc., is only a collection method, and there may be another division method in actual implementation, such as the circuit under test The feature information of the board image and the feature information of the sample circuit board image can be combined or can be integrated into another system, or some features can be ignored, or not implemented.

In a more specific embodiment, in the process that the server implements the extraction of the feature information of the image of the circuit board to be tested and the feature information of the sample circuit board image, it may also include the need to use various interpolation algorithms to firstly extract the image of the circuit board to be tested and the sample The board image is scaled. Subsequently, a defect detection process is performed based on the scaled image of the circuit board to be tested and the image of the sample circuit board.

In some embodiments, after the server obtains the image of the circuit board to be tested and the image of the sample circuit board, it may further include:

The image of the circuit board to be tested and the image of the sample circuit board are scaled to obtain the image of the circuit board to be tested and the image of the sample circuit board after scaling with the same pixel size.

In some embodiments, the server may scale the image of the circuit board under test and the image of the sample circuit board through an interpolation method.

In some embodiments, interpolation algorithms may include adaptive interpolation algorithms and non-adaptive interpolation algorithms. Adaptive methods can change according to the content of the interpolation (eg, sharp edges or smooth textures included in the image), while non-adaptive methods need to perform the same scaling process on all pixels in the feature map. Among them, non-adaptive algorithms include: nearest neighbor, bilinear, bicubic, spline, sinc, lanczos, etc.

In some embodiments, since the complexities of the features included in the image of the circuit board to be tested and the image of the sample circuit board are not the same, the server uses an interpolation algorithm for the features in the image of the circuit board to be tested and the image of the sample circuit board to use values ranging from 0 to 256 (or more) adjacent pixels are interpolated and scaled (including interpolated distortion) to obtain the scaled image of the circuit board under test and the image of the sample circuit board with the same pixel size.

In some embodiments, the server interpolates and scales the features in the test circuit board image and the sample circuit board image, the more adjacent pixels are included, the more accurate the distorted or scaled image is, but the longer it takes.

In some embodiments, the server extracts the first characteristic information of the circuit board image to be tested and the second characteristic information of the sample circuit board image, which may specifically include:

The first feature information of the scaled circuit board image to be tested and the second feature information of the scaled sample circuit board image are extracted.

In some embodiments, the server extracts the feature information of the image from the collected image of the circuit board to be tested through a preset neural network, and extracts the feature information of the image from the image of the sample circuit board.

In some embodiments, both the scaled first feature information of the circuit board image to be tested and the scaled second feature information of the sample circuit board image include color features, texture features, shape features, At least one of the spatial relationship features.

In some embodiments, the spatial relationship feature of the corresponding pixel position of the image is used to represent the distance and direction relationship between the corresponding pixel position of a feature and the pixel position of adjacent features in the feature information of the image.

In some embodiments, the color feature of the corresponding pixel position of the image includes the pixel proportion of the appearance color of a corresponding feature in the feature information of the image and the spatial relationship of color transformation with distance.

In some embodiments, the shape features corresponding to pixel positions in the image include contour features and area features. Among them, the contour feature is mainly aimed at the outer boundary of a corresponding feature in the feature information of the image, and the region feature is mainly aimed at the entire shape area corresponding to the feature information of the image.

In some embodiments, referring to FIG. 3 , FIG. 3 is a schematic flow chart of an embodiment of defect detection in an image of a circuit board to be tested in this application. In step S13, the server performs defect detection on the image of the circuit board to be tested based on the correlation feature between the first feature information and the second feature information, and obtains a defect detection result, which can be specifically implemented in the following manner:

Step S131, determining a correlation feature between the first feature information and the second feature information.

In some embodiments, referring to FIG. 4 , FIG. 4 is a schematic flowchart of an embodiment of determining a correlation feature between first feature information and second feature information in this application. In step S131, the correlation feature between the server's first feature information and the second feature information can be specifically implemented in the following manner:

Step a1, determining each pair of pixels corresponding to the same pixel position between the scaled image of the circuit board to be tested and the scaled image of the sample circuit board.

In some embodiments, the server first determines that the pixel positions of the two images correspond to the same pair of pixel points in the image of the circuit board under test and the image of the sample circuit board with the same pixel size.

Step a2, based on the first feature information and the second feature information, calculate the similarity of each pair of pixel points corresponding to at least one feature information.

In some embodiments, the server may calculate the similarity between each pair of pixel points at the same pixel point position corresponding to at least one feature information by using the Ousian distance or the corotational distance.

Among them, after calculating the eigenvector of the pixel point A of the circuit board image to be tested and the eigenvector of the pixel point B of the sample circuit board image, the Ousian distance or the corotation distance, and then based on the Ousce distance or the corotation distance Determine the similarity between the pixel point A of the image of the circuit board to be tested and the pixel point B of the image of the sample circuit board corresponding to at least one kind of feature information.

Step S132 , performing defect detection on the scaled image of the circuit board to be tested according to the correlation feature, and obtaining a defect detection result.

In some embodiments, referring to FIG. 5 , FIG. 5 is a schematic flowchart of an embodiment of performing defect detection on the scaled image of the circuit board to be tested in the present application. In step S132, the server performs defect detection on the scaled image of the circuit board to be tested according to the correlation feature, and obtains a defect detection result, which can be specifically implemented in the following manner:

Step a3, determining that the scaled image of the circuit board to be tested corresponds to the first number of pixel features and the scaled image of the sample circuit board corresponds to the second number of pixel features.

In some embodiments, after the server scales the image of the circuit board to be tested and the image of the sample circuit board, the server then determines that the feature information of the image of the circuit board to be tested corresponds to the pixel feature whose number of channels on the image is the first number of channels, and The feature information of the image of the sample circuit board corresponds to the features of pixels whose channel number is the second channel number on the image.

As an example, the feature information of the image of the circuit board to be tested corresponds to A1 features on the image of the circuit board to be tested after zooming, and the A1 features are distributed on the image of the circuit board to be tested after zooming to occupy A2 pixels, then The A2 pixel points are pixel features whose channel number is the first channel number on the image. The feature information of the sample circuit board image corresponds to B1 features on the sample circuit board image after scaling, and the B1 features occupy B2 pixels on the scaled sample circuit board image, then the B2 pixels are is the pixel feature whose channel number is the second channel number on the image.

Step a4, based on the first number of pixel features, the second number of pixel features, and the similarity between each pair of pixel points corresponding to at least one feature information, perform defect identification and defect labeling on the scaled circuit board image to be tested, and obtain Defect detection results.

In some embodiments, referring to FIG. 6 , FIG. 6 is a schematic flowchart of an embodiment of performing defect identification and defect labeling on the scaled circuit board image to be tested in the present application. In step a4, based on the first number of pixel features, the second number of pixel features, and the similarity between each pair of pixel points corresponding to at least one feature information, the server performs defect identification and defect identification on the scaled image of the circuit board to be tested. Marking to get defect detection results, which can be achieved in the following ways:

Step b1, based on the first number of pixel features, the second number of pixel features and the similarity between each pair of pixel points corresponding to at least one kind of feature information, traversing and identifying each pixel in the scaled image of the circuit board to be tested, Get the defect recognition result.

In some embodiments, the server inputs the pixel features of the first number of channels of the circuit board image to be tested, the pixel features of the second number of channels of the sample circuit board image, and the similarity of at least one feature information corresponding to the image into an established In the trained convolutional neural network model, each pixel in the scaled image of the circuit board to be tested is traversed to identify whether it belongs to the pixel corresponding to the defect feature.

Step b2: Based on the defect identification result, mark the corresponding pixel area to obtain the defect detection result.

In some embodiments, if the convolutional neural network model traverses and recognizes that a part of the pixels corresponding to the scaled circuit board image to be tested belongs to the pixels corresponding to the defect feature, the convolutional neural network model will The area is marked, and finally the convolutional neural network model outputs the marked and scaled image of the circuit board to be tested to the server as the defect detection result.

In some embodiments, the scaled image of the circuit board to be tested after the labeling includes the convolutional neural network model marking the pixel area corresponding to the defect feature by means of frame selection or adding a specific color, and the labeling The adjacent area of the area is also marked with the position information of the pixel point area corresponding to the defect feature.

In some embodiments, as shown in FIG. 7 , a second defect detection method based on feature correlation is provided. Taking the application of the method to the server 104 in FIG. 1 as an example for illustration, the method includes the following steps:

Step S21 , using the sample circuit board image, traverse whether the similarity between the scaled circuit board image to be tested and the sample circuit board image corresponding to the same area satisfies a preset threshold.

In some embodiments, the sample circuit board image is any one of a preset collected image of a non-defective circuit board and a design image of the circuit board to be tested. Wherein, the design image of the circuit board to be tested is an image of a designed circuit board without defects.

In some embodiments, the server will design a defect-free sample circuit board image to traverse every identical region of the test circuit board image matching the same pixel size. Wherein, the matching process includes determining whether the similarity of each of the same regions satisfies a preset threshold.

Step S22 , if the preset threshold is not met, mark the image area corresponding to the image area not satisfying the preset threshold in the zoomed image of the circuit board to be tested as a defect area in the zoomed image of the circuit board to be tested.

In some embodiments, if the server determines that the similarity of at least a part of the image of the circuit board to be tested corresponds to the same area as the sample circuit board image does not meet the preset threshold requirements, the server will The at least a part of the area is marked in the board image, and the marked area is used as a defect area in the image of the circuit board to be tested.

In some embodiments, the sample circuit board image may be a preset collected image of a defective circuit board. Wherein, the defective circuit board is a circuit board with multiple known defects fabricated on the circuit board by the design engineer in advance, that is, the sample circuit board can be regarded as a defect sample template.

In some embodiments, the server can use the captured image of the defective circuit board to traverse each pixel area in the zoomed image of the circuit board to be tested, and mark the image in the zoomed image of the circuit board to be tested that matches the captured image region, as the defect region in the scaled image of the circuit board under test.

Specifically, the server uses the sample circuit board image designed as a defect sample template to traverse each same area of the scaled circuit board image to be tested that matches the same pixel size. Wherein, the matching process includes determining whether the similarity of each of the same regions satisfies a preset threshold.

Further, if the server determines that the similarity of at least a part of the image of the circuit board under test after zooming corresponds to the same area as the image of the sample circuit board satisfies the preset threshold requirement, then the server At least a part of the area is marked in the image, and the marked area is used as a defect area in the image of the circuit board to be tested.

In some embodiments, referring to FIG. 8 , FIG. 8 is a schematic flowchart of an embodiment of updating a defective circuit board in the present application. After the server marks the image area that matches the captured image in the scaled image of the circuit board to be tested, the following implementation process can also be carried out specifically:

Step c1, extracting image regions matching the collected images.

In some embodiments, the server extracts the marked image region that matches the collected image from the zoomed image of the circuit board to be tested.

Step c2, performing one or more image processing of flipping, rotating, zooming in, zooming out, and chroma adjustment on the extracted image area to obtain a plurality of extended image areas.

In some embodiments, the server can use the image processing program carried by its own equipment, or the optical detection device to generate the extracted multiple image regions into corresponding multiple images to be trained, and send them to a third-party organization (such as image processing platform, cloud server, etc.), to perform one or more image processing in flipping, rotating, zooming in, zooming out, and chroma adjustment to the multiple images to be trained to obtain multiple extended image regions.

Step c3, using multiple extended image areas to update the preset defective circuit board.

In some embodiments, the server adds the obtained plurality of extended image regions to the sample circuit board image designed as a defect sample template, so as to update the sample circuit board image. There can be one or more than one sample circuit board image, that is, the obtained multiple extended image regions can be partially added to one sample circuit board image, and another part can be added to other multiple sample circuit board images. The specific manner and steps of updating the preset defective circuit board are specifically defined here.

In order to more clearly illustrate the defect detection method based on the feature correlation provided by the embodiment of the present disclosure, the defect detection method based on the feature correlation will be described in detail below with a specific embodiment. In an exemplary embodiment, refer to FIG. 9. FIG. 9 is a flow chart of the second defect detection method based on feature correlation provided by the embodiment of the present application. The defect detection method based on feature correlation is used in the server 104. , including the following:

Step S31: The server extracts the feature information of the defect map and the feature information of the standard map.

Wherein, the server takes pictures of the PCB board through the AOI to obtain the AOI image of the defect map. Or take pictures of the PCB board through AVI to get the AVI image of the defect map. The defect map image to be detected can be black and white or color.

Wherein, the server extracts the feature information of the defect map and the standard map through the neural network, for example, the server can use the convolutional neural network to perform feature extraction on the defect map and the standard map, and obtain the feature map corresponding to the defect map and the standard map, As the feature information corresponding to the defect map and the standard map.

Wherein, the standard drawing refers to the defect-free parent drawing corresponding to the defect drawing, which may be a design drawing or a CAM drawing.

Among them, feature extraction refers to using a computer to extract image information and determine whether each image point belongs to an image feature. The result of feature extraction is to divide the points on the image into different subsets, which often belong to isolated points, continuous curves or continuous areas. Commonly used image features include color features, texture features, shape features, and spatial relationship features.

In this embodiment of the present application, feature extraction is performed on the AOI image to be detected, the black and white AVI image, and the color AVI image, wherein different networks may be used to extract different features.

Exemplarily, the extracted features include image color features and shape features. Color features can be represented by a color correlation map, which is another way of expressing the color distribution of an image. It not only depicts the proportion of pixels of a certain color, but also expresses the spatial relationship of color changes with distance, reflecting the The spatial relationship between colors. There are two types of representation methods for shape features, one is contour features and the other is area features. The contour feature of the image is mainly aimed at the outer boundary of the object, while the regional feature of the image is related to the entire shape area. This application does not limit the specific feature representation method.

Step S32: The server scales the feature maps corresponding to the feature information of the defect map and the standard map to the same size.

Wherein, the server uses the feature maps corresponding to the defect map and the standard map as feature information of the defect map and the standard map.

In some embodiments, the server may scale the feature maps of the two through an interpolation method. Among them, the interpolation algorithm may include two types: adaptive type and non-adaptive type. Adaptive methods can change according to the content of the interpolation (for example, sharp edges or smooth textures included in the feature map), and non-adaptive methods need to perform the same scaling process on all pixels in the feature map.

Among them, non-adaptive algorithms include: nearest neighbor, bilinear, bicubic, spline, sinc, lanczos, etc.

Among them, because the complexity of the feature map corresponding to the defect map and the standard map is not the same, the server uses an interpolation algorithm for the features in the feature map from 0 to 256 (or more) adjacent pixels, that is, the adjacent pixels included in the feature interpolation The more, the more accurate the warped or scaled image, but the longer it takes.

Step S33: The server calculates the correlation feature of each pixel in the feature map corresponding to the defect map and the standard map according to the scaled feature map.

Among them, the correlation feature on each pixel includes the cosine similarity on the same pixel corresponding to the defect map and the standard map.

Among them, the cosine similarity measures the similarity between two vector inner product spaces by measuring the cosine value between them, and is applicable to vector comparisons of any dimension, so it belongs to the machine learning algorithm that is widely used in high-dimensional spaces. Among them, digital images contain many feature codes, and these feature groups belong to high-dimensional space. The server converts the feature groups of each image into vectors in high-dimensional space, and the cosine value of the angle between two vectors can be used to determine two Whether the vectors point roughly in the same direction.

In the embodiment of the present application, the correlation feature on each pixel is determined by calculating the cosine value of the angle included in the inner product space of the vector representing the feature of each pixel.

Step S34: The server inputs the correlation feature including each pixel into the trained neural network for defect detection, and outputs the result of defect detection.

Wherein, the defect detection result includes detection result information such as the type, size, and position of the defect in the defect map.

Among them, the server will correspond to the correlation features on the same pixel of the defect map and the standard map after zooming (after zooming, the size of the image is one-quarter or one-eighth of the original image), and the number of feature channels of the corresponding image (256 , 512) are jointly input into the neural network for defect detection, to output a feature map of a new defect map with the same size and with detection labels, and the feature map of the new defect map is the result of defect detection.

As an example, the server extracts A (channel) features from the defect map, extracts B (channel) features from the standard map, and uses the pixel similarity feature as one (channel) feature. Finally, the server takes the A+B+ One (channel) feature is input into the neural network for comparison and judgment calculation process to obtain the defect result.

It should be understood that although the various steps in the flow charts of FIGS. 2-9 are shown sequentially as indicated by the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in FIGS. 2-9 may include multiple steps or stages. These steps or stages are not necessarily performed at the same time, but may be performed at different times. These steps or stages The execution sequence is not necessarily performed sequentially, but may be performed alternately or alternately with other steps or at least a part of steps or stages in other steps.

It can be understood that the same/similar parts between the various embodiments of the above-mentioned methods in this specification can be referred to each other, and each embodiment focuses on the differences from other embodiments, and for relevant parts, refer to other method embodiments description of the .

Fig. 10 is a block diagram of a defect detection device based on feature correlation provided by an embodiment of the present application. Referring to Fig. 10, the defect detection device 10 based on feature correlation includes:

An acquisition unit 11, configured to acquire an image of a circuit board to be tested and a sample circuit board image;

An extraction unit 12, configured to extract the first feature information of the circuit board image to be tested and the second feature information of the sample circuit board image;

The detection unit 13 is configured to perform defect detection on the image of the circuit board to be tested based on the correlation feature between the first feature information and the second feature information to obtain a defect detection result; wherein the correlation feature is used to characterize the image of the circuit board to be tested The same degree that corresponds to the same area between the sample board images.

In some embodiments, after acquiring the image of the circuit board to be tested and the image of the sample circuit board, the defect detection device 10 based on feature correlation is further specifically used for:

The image of the circuit board to be tested and the image of the sample circuit board are scaled to obtain the image of the circuit board to be tested and the image of the sample circuit board after scaling with the same pixel size.

In some embodiments, in terms of extracting the first feature information of the circuit board image to be tested and the second feature information of the sample circuit board image, the extraction unit 12 is specifically used for:

The first feature information of the scaled circuit board image to be tested and the second feature information of the scaled sample circuit board image are extracted.

In some embodiments, the detection unit 13 is specifically configured to:

determining a correlation feature between the first feature information and the second feature information;

According to the correlation feature, defect detection is carried out on the scaled image of the circuit board to be tested, and the defect detection result is obtained.

In some embodiments, in terms of determining the correlation feature between the first feature information and the second feature information, the detection unit 13 is specifically configured to:

Determining each pair of pixels corresponding to the same pixel position between the scaled circuit board image to be tested and the scaled sample circuit board image;

Based on the first feature information and the second feature information, the similarity of each pair of pixel points corresponding to at least one feature information is calculated.

Wherein, both the first feature information and the second feature information include at least one of color features, texture features, shape features, and spatial relationship features at corresponding pixel positions of the image.

In some embodiments, the detection unit 13 is specifically used to:

Determining that the scaled circuit board image to be tested corresponds to the first number of pixel features and the scaled sample circuit board image corresponds to the second number of pixel features;

Based on the first number of pixel features, the second number of pixel features, and the similarity between each pair of pixel points corresponding to at least one feature information, perform defect identification and defect labeling on the scaled image of the circuit board to be tested to obtain a defect detection result .

In some embodiments, based on the first number of pixel features, the second number of pixel features, and the similarity between each pair of pixel points corresponding to at least one feature information, the zoomed image of the circuit board to be tested is used for defect identification and In terms of defect labeling and defect detection results, the detection unit 13 is specifically used for:

Based on the first number of pixel features, the second number of pixel features and the similarity between each pair of pixel points corresponding to at least one feature information, traverse and identify each pixel in the scaled circuit board image to be tested to obtain defect identification result;

Based on the defect recognition result, the corresponding pixel area is marked to obtain the defect detection result.

In some embodiments, the defect detection device 10 based on feature correlation is also specifically used for:

Using the sample circuit board image, traversing whether the similarity between the scaled circuit board image to be tested and the sample circuit board image corresponding to the same area satisfies a preset threshold;

If the preset threshold value is not satisfied, in the image of the circuit board to be tested after zooming, mark the image area corresponding to the image area not satisfying the preset threshold value as a defective area in the image of the circuit board to be tested after zooming.

Wherein, the sample circuit board image is any one of a preset collected image of a non-defective circuit board and a design image of the circuit board to be tested.

In some embodiments, the defect detection device 10 based on feature correlation is also specifically used for:

Use the captured image of the defective circuit board to traverse each pixel area in the image of the circuit board under test after zooming, and mark the image area in the image of the circuit board under test after zooming that matches the collected image, as the image of the circuit board under test after zooming defect area in .

Wherein, the sample circuit board image is a preset collected image of a defective circuit board.

In some embodiments, the defect detection device 10 based on feature correlation is also specifically used for:

After marking the image area matching the collected image in the image of the circuit board to be tested after scaling, extracting the image area matching the collected image;

Perform one or more image processings of flipping, rotating, zooming in, zooming out, and chroma adjustment on the extracted image area to obtain multiple extended image areas;

Preset defective boards are updated with multiple extended image areas.

Regarding the apparatus in the foregoing embodiments, the specific manner in which each module executes operations has been described in detail in the embodiments related to the method, and will not be described in detail here.

FIG. 11 is a block diagram of an electronic device 20 provided by an embodiment of the present application. For example, electronic device 20 may be a server. Referring to FIG. 11 , the electronic device 20 includes a processor 21, further the processor 21 may be a set of processors, which may include one or more processors, and the electronic device 20 includes a memory resource represented by a memory 22, wherein the memory Computer programs, such as application programs, are stored on 22 . The computer program stored in memory 22 may include one or more modules each corresponding to a set of executable instructions. In addition, the processing component 21 is configured to implement the defect detection method based on feature correlation as described above when executing the computer program.

In some embodiments, electronic device 20 is a server in which a computing system can run one or more operating systems, including any of the operating systems discussed above as well as any commercially available server operating systems. The server can also run any of a variety of additional server applications and/or middle-tier applications, including HTTP (Hypertext Transfer Protocol) servers, FTP (File Transfer Protocol) servers, CGI (Common Gateway Interface) servers, servers, database server etc. Exemplary database servers include, but are not limited to, those commercially available from (International Business Machines) and the like.

In some embodiments, processing component 21 generally controls the overall operation of electronic device 20, such as operations associated with display, data processing, data communication, and recording operations. The processing component 21 may include one or more processors to execute computer programs, so as to complete all or part of the steps of the above method. Additionally, processing component 21 may include one or more modules to facilitate interaction between processing component 21 and other components. For example, the processing component 21 may include a multimedia module to facilitate the use of the multimedia component to control the interaction between the user electronic device and the processing component 21 .

In some embodiments, the processor in the processing component 21 may also be called a CPU (Central Processing Unit, central processing unit). A processor may be an electronic chip with signal processing capabilities. The processor can also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array ( Field-Programmable GateArray, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. A general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like. In addition, the processors may be jointly implemented by integrated circuit chips.

In some embodiments, memory 22 is configured to store various types of data to support operations at electronic device 20 . Examples of such data include instructions, collected data, messages, pictures, videos, etc. for any application or method operating on the electronic device 20 . The memory 22 can be implemented by any type of volatile or non-volatile storage device or their combination, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic Disk, Optical Disk or Graphene Memory.

In some embodiments, the memory 22 can be a memory stick, a TF card, etc., and can store all information in the electronic device 20, including input raw data, computer programs, intermediate running results and final running results are all stored in an embodiment In, in the memory 22. It stores and retrieves information from locations specified by the processor. With the memory 22 in one embodiment, the electronic device 20 has a memory function to ensure normal operation. In one embodiment of the electronic device 20, the memory 22 can be divided into main memory (internal memory) and auxiliary memory (external memory) according to the purpose, and there are also classification methods for external memory and internal memory. External storage is usually magnetic media or optical discs, which can store information for a long time. Memory refers to the storage unit on the motherboard, which is used to store data and programs currently being executed, but it is only used to store programs and data temporarily, and the data will be lost when the power is turned off or cut off.

In some embodiments, the electronic device 20 may further include: a power supply component 23 configured to perform power management of the electronic device 20, a wired or wireless network interface 24 configured to connect the electronic device 20 to a network, and an input/output (I/O O) Interface 25. The electronic device 20 can operate based on an operating system stored in the memory 22, such as Windows Server, MacOS X, Unix, Linux, FreeBSD or the like.

In some embodiments, the power supply component 23 provides power to various components of the electronic device 20 . Power supply components 23 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 20 .

In some embodiments, wired or wireless network interface 24 is configured to facilitate wired or wireless communications between electronic device 20 and other devices. The electronic device 20 can access a wireless network based on a communication standard, such as WiFi, an operator network (such as 2G, 3G, 4G or 5G), or a combination thereof.

In some embodiments, the wired or wireless network interface 24 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, wired or wireless network interface 24 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

In some embodiments, an input/output (I/O) interface 25 provides an interface between the processing component 21 and a peripheral interface module, which may be a keyboard, a click wheel, buttons, and the like. These buttons may include, but are not limited to: a home button, volume buttons, start button, and lock button.

FIG. 12 is a block diagram of a computer-readable storage medium 30 provided by an embodiment of the present application. A computer program 31 is stored on the computer-readable storage medium 30 , wherein when the computer program 31 is executed by a processor, the defect detection method based on feature correlation as described above is realized.

If the integrated units of the functional units in various embodiments of the present application are implemented in the form of software functional units and sold or used as independent products, they can be stored in the computer-readable storage medium 30 . Based on this understanding, the technical solution of the present application is essentially or part of the contribution to the prior art, or all or part of the technical solution can be embodied in the form of software products, and the computer-readable storage medium 30 is stored in a computer In the program 31, several instructions are included to make a computer device (which can be a personal computer, a system server, or a network device, etc.), an electronic device (such as MP3, MP4, etc., or a mobile phone, a tablet computer, a wearable device, etc.) The smart terminal (also can be a desktop computer, etc.) or a processor (processor) to execute all or part of the steps of the methods in various embodiments of the present application.

Fig. 13 is a block diagram of a computer program product 40 provided by an embodiment of the present application. The computer program product 40 includes program instructions 41 that can be executed by the processor of the electronic device 20 to implement the defect detection method based on feature correlation as described above.

Those skilled in the art should understand that the embodiment of the present application may provide a defect detection method based on feature correlation, a defect detection device 10 based on feature correlation, an electronic device 20, a computer-readable storage medium 30 or a computer program product 40. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product 40 implemented on one or more computer program instructions 41 (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer usable program code embodied therein.

This application refers to the flow chart and/or the defect detection method based on feature correlation, the defect detection device 10 based on feature correlation, the electronic device 20, the computer-readable storage medium 30 or the computer program product 40 according to the embodiment of the application. block diagram to describe. It should be understood that each process and/or block in the flowchart and/or block diagram, and a combination of processes and/or blocks in the flowchart and/or block diagram can be realized by the computer program product 40 . These computer program products 40 may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the program instructions executed by the processor of the computer or other programmable data processing equipment 41 Produce means for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program products 40 may also be stored in a computer readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner such that program instructions 41 stored in the computer program product 40 produce an article of manufacture comprising instruction means , the instruction means realizes the functions specified in one or more procedures of the flow chart and/or one or more blocks of the block diagram.

These program instructions 41 can also be loaded on a computer or other programmable data processing device, so that a series of operation steps are performed on the computer or other programmable device to produce computer-implemented processing, so that the process executed on the computer or other programmable device The program instructions 41 provide steps for implementing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

It should be noted that the above-mentioned various methods, devices, electronic devices, computer-readable storage media, computer program products, etc. may also include other implementations according to the descriptions of the method embodiments. For specific implementation methods, please refer to the relevant method embodiments description and will not be repeated here.

Other embodiments of the present disclosure will be readily apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. The present disclosure is intended to cover any modification, use or adaptation of the present disclosure. These modifications, uses or adaptations follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. . The specification and examples are to be considered exemplary only, with the true scope and spirit of the disclosure indicated by the appended claims.

It should be understood that the present disclosure is not limited to the precise constructions which have been described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims (11)

1. A defect detection method based on feature correlation, comprising:

acquiring a circuit board image to be tested and a sample circuit board image;

extracting first characteristic information of the circuit board image to be detected and second characteristic information of the sample circuit board image;

performing defect detection on the circuit board image to be detected based on the correlation characteristic between the first characteristic information and the second characteristic information to obtain a defect detection result; the correlation feature is used for representing the same degree of the corresponding same area between the circuit board image to be tested and the sample circuit board image.

2. The method of claim 1, wherein after the obtaining the circuit board image to be tested and the sample circuit board image, the method further comprises:

scaling the circuit board image to be tested and the sample circuit board image to obtain a scaled circuit board image to be tested and a sample circuit board image with the same pixel size;

the extracting the first characteristic information of the circuit board image to be detected and the second characteristic information of the sample circuit board image includes:

extracting first characteristic information of the scaled circuit board image to be tested and second characteristic information of the scaled sample circuit board image;

performing defect detection on the circuit board image to be detected based on the correlation characteristic between the first characteristic information and the second characteristic information to obtain a defect detection result, wherein the defect detection result comprises:

determining a correlation feature between the first feature information and the second feature information;

and performing defect detection on the scaled circuit board image to be detected according to the correlation characteristics to obtain a defect detection result.

3. The method of claim 2, wherein the first feature information and the second feature information each include at least one of a color feature, a texture feature, a shape feature, and a spatial relationship feature at a location of a corresponding pixel point of the image;

The determining a correlation feature between the first feature information and the second feature information includes:

determining each pair of pixel points corresponding to the same pixel point position between the scaled circuit board image to be detected and the scaled sample circuit board image;

and calculating the similarity of at least one characteristic information corresponding to each pair of pixel points based on the first characteristic information and the second characteristic information.

4. The method of claim 3, wherein performing defect detection on the scaled circuit board image to be tested according to the correlation characteristic to obtain a defect detection result comprises:

determining a first number of pixel features corresponding to the scaled circuit board image to be tested and a second number of pixel features corresponding to the scaled sample circuit board image;

and performing defect identification and defect labeling on the scaled circuit board image to be detected based on the first number of pixel features, the second number of pixel features and the similarity of at least one feature information corresponding to each pair of pixel points to obtain a defect detection result.

5. The method of claim 4, wherein performing defect recognition and defect labeling on the scaled circuit board image to be tested based on the first number of pixel features, the second number of pixel features, and the similarity of at least one feature information corresponding to each pair of pixel points to obtain a defect detection result comprises:

Traversing and identifying each pixel point in the scaled circuit board image to be detected based on the first number of pixel features, the second number of pixel features and the similarity of at least one feature information corresponding to each pair of pixel points to obtain a defect identification result;

and marking the corresponding pixel point areas based on the defect identification result to obtain a defect detection result.

6. The method of claim 2, wherein the sample circuit board image is any one of a preset collection image of a defect-free circuit board and a design image of the circuit board to be tested;

the method further comprises the steps of:

traversing whether the similarity of the corresponding same region between the zoomed circuit board image to be tested and the sample circuit board image meets a preset threshold value or not by utilizing the sample circuit board image;

if the preset threshold value is not met, marking an image area which corresponds to the preset threshold value not met in the zoomed circuit board image to be detected as a defect area in the zoomed circuit board image to be detected.

7. The method of claim 2, wherein the sample circuit board image is an acquisition image of a pre-set defective circuit board;

The method further comprises the steps of:

and traversing each pixel point area in the zoomed circuit board image to be detected by utilizing the acquired image of the defect circuit board, and marking an image area matched with the acquired image in the zoomed circuit board image to be detected as the defect area in the zoomed circuit board image to be detected.

8. The method of claim 7, wherein after said labeling the image area in the scaled circuit board image to be tested that matches the captured image, the method further comprises:

extracting an image area matched with the acquired image;

performing one or more of image processing of turning, rotating, amplifying, shrinking and chromaticity adjustment on the extracted image area to obtain a plurality of extended image areas;

and updating the preset defect circuit board by utilizing the plurality of extended image areas.

9. A defect detection apparatus based on feature correlation, comprising:

the acquisition unit is used for acquiring the circuit board image to be detected and the sample circuit board image;

the extraction unit is used for extracting first characteristic information of the circuit board image to be detected and second characteristic information of the sample circuit board image;

The detection unit is used for carrying out defect detection on the circuit board image to be detected based on the correlation characteristic between the first characteristic information and the second characteristic information to obtain a defect detection result; the correlation feature is used for representing the same degree of the corresponding same area between the circuit board image to be tested and the sample circuit board image.

10. An electronic device comprising a memory storing a computer program and a processor implementing the feature correlation-based defect detection method as claimed in any one of claims 1 to 8 when the computer program is executed.

11. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the feature correlation-based defect detection method as claimed in any one of claims 1 to 8.

Images