CN115641330B - Flexible circuit board defect detection method and system based on image processing - Google Patents

Flexible circuit board defect detection method and system based on image processing Download PDF

Info

Publication number
CN115641330B
CN115641330B CN202211440566.7A CN202211440566A CN115641330B CN 115641330 B CN115641330 B CN 115641330B CN 202211440566 A CN202211440566 A CN 202211440566A CN 115641330 B CN115641330 B CN 115641330B
Authority
CN
China
Prior art keywords
description
array
image
analysis space
defect
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202211440566.7A
Other languages
Chinese (zh)
Other versions
CN115641330A (en
Inventor
汪永新
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Dongguan Zonja Printing Co ltd
Original Assignee
Dongguan Zonja Printing Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Dongguan Zonja Printing Co ltd filed Critical Dongguan Zonja Printing Co ltd
Priority to CN202211440566.7A priority Critical patent/CN115641330B/en
Publication of CN115641330A publication Critical patent/CN115641330A/en
Application granted granted Critical
Publication of CN115641330B publication Critical patent/CN115641330B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

According to the soft circuit board defect detection method and system based on image processing, the image sequence to be identified is separated to obtain a plurality of separated images, the plurality of separated images are respectively subjected to first analysis space description array extraction and second analysis space description array extraction and description array fusion, a target fusion description array is obtained, then the image definition description arrays corresponding to the plurality of separated images are acquired, based on the image definition description arrays, the information of two analysis spaces can be contained in the extracted image definition description arrays at the same time, the extracted image definition description arrays can restore basic properties of the images extremely, soft circuit board defect credibility coefficients corresponding to the plurality of separated images are acquired, accuracy of image defect detection is improved, a plurality of defect image chains are determined in the image sequence, the corresponding defect definition description arrays are determined, unified defect image chain combination is finally obtained, and the accuracy of reasoning and the accuracy of unified defect image chain combination are improved.

Description

Flexible circuit board defect detection method and system based on image processing
Technical Field
The application relates to the field of image processing and artificial intelligence, in particular to a soft circuit board defect detection method and system based on image processing.
Background
The flexible circuit board has a great number of advantages of saving space, reducing weight, high flexibility and the like, and has very wide application in production and life. Common defects of flexible circuit boards include short-circuit or missing of welding spots, short-circuit of component burning circuits, welding bridges, open circuits, welding spots and wiring cracks of circuit boards, in conventional defect detection, manual mode is adopted, obviously, the efficiency and accuracy of manual detection cannot be ensured in the face of large-batch defect circuit board detection, and the trend is that auxiliary defect detection is carried out in an artificial intelligent mode. In the face of a large-batch circuit board defect detection scene, such as factory-returning batch detection, reasonable collection of defects of the flexible circuit board is needed, unified processing and product yield analysis are convenient, and therefore, the rapid collection of defective products on the premise of ensuring efficient identification is a technical problem to be solved.
It should be noted that, the foregoing description of the defect collection is merely an introduction to the background of the technology of the present application, and is not representative of the premise and the reference for evaluating the creativity of the present application.
Disclosure of Invention
The invention aims to provide a soft circuit board defect detection method and system based on image processing so as to solve the technical problems.
In order to achieve the above objective, the implementation procedure of the embodiment of the present application is as follows:
an embodiment of the present application provides a method for detecting a defect of a flexible circuit board based on image processing, which is applied to a defect detection system, and the method includes:
acquiring an image sequence to be identified, and separating the image sequence to be identified to obtain a plurality of separated images; the image sequence to be identified comprises a plurality of images of the flexible circuit board to be detected;
extracting a first analysis space description array from the plurality of separated images respectively to obtain a first analysis space description array corresponding to the plurality of separated images, wherein the first analysis space description array comprises a first analysis space temporary description array and a first analysis space final description array;
extracting second analysis space description arrays of the plurality of separated images respectively to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array;
Performing description array fusion through a first analysis space temporary description array and a second analysis space temporary description array corresponding to the plurality of separation images to obtain a target fusion description array corresponding to the plurality of separation images;
extracting an image interpretation description array through a first analysis space final description array, a second analysis space final description array and a target fusion description array corresponding to the plurality of separated images to obtain an image interpretation description array corresponding to the plurality of separated images, and detecting image defects through the image interpretation description array to obtain a soft circuit board defect credible coefficient corresponding to the plurality of separated images;
determining a plurality of defect image chains from the image sequence to be identified through the flexible circuit board defect credibility coefficient, and determining a defect definition description array corresponding to the plurality of defect image chains through the image definition description array;
and carrying out defect identification on the defect image chains through the defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination.
Optionally, the identifying the defect of the defect image chain by the defect definition description array corresponding to the plurality of defect image chains to obtain a unified defect image chain combination includes:
Performing specified array extraction through defect definition description arrays corresponding to the defect image chains to obtain a selected description array;
analyzing and restoring the soft circuit board defect credible coefficients corresponding to the selected description array and the plurality of separated images to obtain a target defect interpretation description array corresponding to the plurality of defect image chains;
and carrying out defect identification on the plurality of defect image chains through a target defect definition description array corresponding to the plurality of defect image chains to obtain the unified defect image chain combination.
Optionally, the extracting the specified array from the defect paraphrasing description arrays corresponding to the defect image chains to obtain a selected description array includes:
extracting original image description arrays corresponding to the multiple separated images respectively, and determining defect image chain original description arrays corresponding to the multiple defect image chains from the original image description arrays corresponding to the multiple separated images respectively;
integrating the original description arrays of the defect image chains corresponding to the plurality of defect image chains with the corresponding defect definition description arrays respectively to obtain target integration description arrays corresponding to the plurality of defect image chains;
And putting the target integrated description arrays corresponding to the defect image chains into a description array extraction network of the description array processing model for processing to obtain a target selected description array.
Optionally, the performing defect recognition on the multiple defect image chains through the target defect paraphrasing description arrays corresponding to the multiple defect image chains to obtain the unified defect image chain combination includes:
determining the matching degree among the defect image chains through the target defect definition description arrays corresponding to the defect image chains;
and performing image barrel division according to the matching degree among the defect image chains to obtain the unified defect image chain combination.
Optionally, the extracting the first analysis space description arrays of the multiple separated images respectively to obtain first analysis space description arrays corresponding to the multiple separated images, where the first analysis space description arrays include a first analysis space temporary description array and a first analysis space final description array, and the extracting includes:
respectively adopting a first analysis space convolution operation for the plurality of separated images to obtain a plurality of temporary convolution description arrays and ending convolution description arrays corresponding to the plurality of separated images;
Performing dimension transformation operation on the temporary convolution description arrays to obtain a plurality of first analysis space temporary description arrays corresponding to the plurality of separated images;
performing dimension transformation operation on the ending convolution description array to obtain a first analysis space final description array corresponding to the plurality of separated images;
extracting the second analysis space description arrays of the plurality of separated images to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array, and the method comprises the following steps of:
extracting original image description arrays corresponding to the multiple separated images respectively;
and performing a second analysis space convolution operation on the original image description arrays corresponding to the multiple separation images to obtain multiple second analysis space temporary description arrays and multiple second analysis space final description arrays corresponding to the multiple separation images.
Optionally, the number of the first analysis space temporary description arrays is a plurality, and the number of the second analysis space temporary description arrays is a plurality;
the method for obtaining the target fusion description array corresponding to the plurality of separation images by fusing the description array through the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separation images comprises the following steps:
Integrating a first analysis space temporary description array tensor1-1 in the plurality of first analysis space temporary description arrays with a corresponding second analysis space temporary description array tensor2-1 in the plurality of second analysis space temporary description arrays to obtain a first integrated description array, and performing convolution processing on the first integrated description array to obtain a first fusion description array;
integrating the first fusion description array, a first analysis space temporary description array tensor1-2 in the plurality of first analysis space temporary description arrays and a corresponding second analysis space temporary description array tensor2-2 in the plurality of second analysis space temporary description arrays to obtain a second integration description array, and performing convolution processing on the second integration description array to obtain a second fusion description array;
and obtaining a target fusion description array when the plurality of first analysis space temporary description arrays and the plurality of second analysis space temporary description arrays are completed.
Optionally, the extracting the image interpretation description array by the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images to obtain the image interpretation description array corresponding to the plurality of separated images, and performing image defect detection by the image interpretation description array to obtain the soft circuit board defect credible coefficient corresponding to the plurality of separated images, including:
Integrating the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separation images to obtain a target integration description array corresponding to the plurality of separation images;
the convolution processing is adopted through the target integration description arrays corresponding to the plurality of separation images, so that the convolution description arrays corresponding to the plurality of separation images are obtained;
acquiring description array maximum values and description array equal values corresponding to each order in the convolution description arrays through the convolution description arrays corresponding to the plurality of separation images;
summing the maximum value of the description array and the equal value of the description array to obtain paraphrase extraction description array values corresponding to all orders in the convolution description array, and obtaining paraphrase extraction description arrays corresponding to the plurality of separated images through the paraphrase extraction description array values corresponding to all orders in the convolution description array;
performing linear transformation on the paraphrasing extraction description arrays corresponding to the plurality of separated images to obtain image paraphrasing description arrays corresponding to the plurality of separated images;
and performing detection and judgment on defects and non-defects of the flexible circuit board through the image interpretation description arrays corresponding to the plurality of separated images to obtain the credible coefficients of the defects of the flexible circuit board corresponding to the plurality of separated images.
Optionally, the method further comprises:
the image sequence to be identified is put into an image defect detection model, and the image sequence to be identified is separated through the image defect detection model, so that a plurality of separated images are obtained;
extracting first analysis space description arrays of the plurality of separated images through the image defect detection model to obtain first analysis space description arrays corresponding to the plurality of separated images, wherein the first analysis space description arrays comprise a first analysis space temporary description array and a first analysis space final description array;
extracting second analysis space description arrays of the plurality of separated images respectively to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array;
performing description array fusion on a first analysis space temporary description array and a second analysis space temporary description array corresponding to the plurality of separated images through the image defect detection model to obtain a target fusion description array corresponding to the plurality of separated images;
And extracting an image interpretation description array from the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images through the image defect detection model to obtain the image interpretation description array corresponding to the plurality of separated images, and detecting the image defect through the image interpretation description array to obtain the soft circuit board defect credible coefficient corresponding to the plurality of separated images.
Optionally, the image defect detection model includes a first analysis space description array extraction network, a second analysis space description array extraction network, a description array fusion network, an image interpretation description array extraction network, and a defect discrimination network;
the method further comprises the steps of:
the image sequence to be identified is put into an image defect detection model, and the image sequence to be identified is separated through the image defect detection model, so that a plurality of separated images are obtained;
the plurality of separated images are put into the first analysis space description array extraction network to carry out second analysis space description array extraction, and a first analysis space temporary description array and a first analysis space final description array are obtained;
The plurality of separated images are put into the second analysis space description array extraction network to extract a second analysis space description array, so that a second analysis space temporary description array and a second analysis space final description array are obtained;
the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separation images are put into the description array fusion network to be subjected to description array fusion, and the target fusion description array corresponding to the plurality of separation images is obtained;
and inputting the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images into the image interpretation description array extraction network to extract the image interpretation description array to obtain the image interpretation description array corresponding to the plurality of separated images, and inputting the image interpretation description array into the defect discrimination network to detect the image defects to obtain the soft circuit board defect credible coefficients corresponding to the plurality of separated images.
A second aspect of the embodiments of the present application provides a defect detection system, including a processor and a memory, where the memory stores a computer program, and when the processor runs the program, the method described above is implemented.
According to the soft circuit board defect detection method and system based on image processing, a plurality of separation images are obtained through separation of an image sequence to be identified, then a first analysis space description array is respectively carried out on the plurality of separation images to obtain a first analysis space temporary description array and a first analysis space final description array, a second analysis space description array is respectively carried out on the plurality of separation images to obtain a second analysis space temporary description array and a second analysis space final description array, then description array fusion is carried out on the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separation images to obtain a target fusion description array corresponding to the plurality of separation images, and the target fusion description array obtained through description array fusion enables the target fusion description array to contain nesting information of different analysis spaces. Then, extracting the image interpretation description arrays through the first analysis space final description arrays, the second analysis space final description arrays and the target fusion description arrays corresponding to the plurality of separated images to obtain the image interpretation description arrays corresponding to the plurality of separated images, so that the extracted image interpretation description arrays simultaneously contain the information of the two analysis spaces, and meanwhile, the extracted image interpretation description arrays can restore the basic properties of the images to the greatest extent. Then, performing image defect detection through the image interpretation description array to obtain a flexible circuit board defect credible coefficient corresponding to a plurality of separated images, so that the accuracy of image defect detection can be improved, then determining a plurality of defect image chains from an image sequence to be identified through the flexible circuit board defect credible coefficient, and determining a defect interpretation description array corresponding to the plurality of defect image chains through the image interpretation description array; and carrying out defect identification on the defect image chains through the defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination, so that the accuracy of defect identification of the defect image chains is improved, and further, the accuracy of the obtained unified defect image chain combination is ensured.
In the following description, other features will be partially set forth. Upon review of the ensuing disclosure and the accompanying figures, those skilled in the art will in part discover these features or will be able to ascertain them through production or use thereof. The features of the present application may be implemented and obtained by practicing or using the various aspects of the methods, tools, and combinations that are set forth in the detailed examples described below.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings that are required to be used in the description of the embodiments of the present application will be briefly described below.
Fig. 1 is a flowchart of a method for detecting defects of a flexible circuit board based on image processing according to an embodiment of the present application.
Fig. 2 is a schematic view of a scene of a method for detecting defects of a flexible circuit board based on image processing according to an embodiment of the present application.
Fig. 3 is a schematic diagram of a functional module architecture of a defect detecting device according to an embodiment of the present application.
Fig. 4 is a schematic diagram of a defect detection system according to an embodiment of the present application.
Detailed Description
Embodiments of the present application are described below with reference to the accompanying drawings in the embodiments of the present application. The terminology used in the description of the embodiments of the application is for the purpose of describing particular embodiments of the application only and is not intended to be limiting of the application.
An execution subject of the method for detecting defects of a flexible printed circuit board based on image processing in the embodiment of the application is a defect detection system, including but not limited to a defect detection system with high performance data processing capability such as a server, a personal computer, a notebook computer and the like. For example, the defect detection system is a server, specifically including but not limited to a single network server, a server group formed by a plurality of network servers, or a cloud formed by a large number of computers or network servers in cloud computing, wherein the cloud computing is a distributed computing type, and is a super virtual computer formed by a group of loosely coupled computer sets. The computer device can be used for realizing the application by running alone, and can also be accessed into a network and realized by interaction with other computer devices in the network. Wherein the network in which the computer device is located includes, but is not limited to, the internet, a wide area network, a metropolitan area network, a local area network, a VPN network, and the like. The defect detection system can communicate with the terminal equipment, and the receiving terminal equipment collects photographed images, such as images of the flexible circuit board to be detected.
The embodiment of the application provides a flexible circuit board defect detection method based on image processing, which is applied to a defect detection system, as shown in fig. 1, and comprises the following steps:
Step 110, acquiring an image sequence to be identified, and separating the image sequence to be identified to obtain a plurality of separated images.
In this embodiment, the image sequence to be identified includes images of a plurality of flexible circuit boards to be detected, where the images are images obtained by photographing a plurality of flexible circuit boards to be detected, and it should be specifically noted that, because of the requirement of the size of the circuit boards and the definition of the images, the number of the images to be photographed is a plurality of, for example, 5, in other words, the 5 images are images obtained by photographing the same flexible circuit board, each image corresponds to a portion of the flexible circuit board, and the images are continuous images that may include a repeated area. In practical application, the image corresponding to the same flexible circuit board can be packaged as one image package, and the image sequence to be identified comprises image packages corresponding to a plurality of flexible circuit boards, in other words, the image sequence to be identified comprises images of a plurality of flexible circuit boards to be detected. The embodiment of the application aims to identify images with defects in a plurality of flexible circuit boards to be detected, and divide and collect the images according to the types of the defects. For example, please refer to fig. 2, which is a schematic view of a scene of an image sequence to be identified provided in the embodiment of the present application, which includes a plurality of image packages, for example A, B, C, D … …, wherein image package a includes images a1, a2, a3, a4, a5, image package B includes images B1, B2, B3, B4, B5, image package C includes images C1, C2, C3, C4, C5, image package D includes images D1, D2, D3, D4, D5 … …, and the soft circuit board corresponding to image package a includes defects, specifically, defects indicated by images a2, a3, a4, i.e., defects a ', a' are type I defects, such as circuit board tin ring defects; correspondingly, the flexible circuit board corresponding to the image package B has defects, specifically, among a plurality of images contained in the flexible circuit board, the defects indicated by the images B1, B2 and B3, namely, the defect types of B ', B' are II type defects, such as the defects of the board surface of the circuit board; the soft circuit board corresponding to the image package C is provided with defects, specifically defects indicated by the images C3, C4 and C5 in the plurality of images contained in the soft circuit board, namely, the defect type of C ', C' is a type I defect; the flexible circuit board corresponding to the image package D has defects, specifically, among the multiple images included in the flexible circuit board, the defects indicated by the images D2, D3, D4 and D5, namely, the defect type of D ', D' is the type II defect … …, and the defect types of a 'and C' are consistent, so that the two are collected into the same set, the defect types of B 'and D' are consistent, and the two are collected into the same set. The following description describes how to perform the above steps, it can be understood that in step 110, the image sequence to be identified is separated to obtain a plurality of separated images, and the separated images are a1, a2, a3, a4, a5, b1, b2, b3, b4, b5, c1, c2, c3, c4, c5, d1, d2, d3, d4, d5 and … …, in other words, the minimum unit obtained by separation is the minimum component unit of the image packet corresponding to the flexible circuit board to be detected.
Step 120, extracting a first analysis space description array from the plurality of separated images respectively to obtain a first analysis space description array corresponding to the plurality of separated images, wherein the first analysis space description array comprises a first analysis space temporary description array and a first analysis space final description array.
The first analysis space description array is an image description array extracted in a first analysis space, in this embodiment of the present application, the first analysis space representation is analyzed in an image space domain or an image frequency domain, in this embodiment of the present application, the first analysis space representation corresponds to the first analysis space, and further includes a second analysis space, where the second analysis space is analyzed in the image space domain or the image frequency domain, and it can be understood that when the first analysis space is analyzed in the image space domain, the second analysis space is analyzed in the image frequency domain, whereas when the first analysis space is analyzed in the image frequency domain, the second analysis space is analyzed in the image space domain, and the first second is only used to distinguish the first analysis space from the second analysis space. The first analysis space temporary description array is a paraphrase array obtained when the first analysis space final description array is extracted, the first analysis space final description array is a first analysis space description array corresponding to the separated image obtained by the final extraction, and the paraphrase array is a characteristic numerical expression for representing the semantics of the image. The first analysis space description array may be regarded as a feature of the image obtained in the first analysis space.
For example, the convolution processing may be repeatedly performed on the separated image, each time processing is performed to obtain a first analysis space temporary description array, the current first analysis space temporary description array is taken as loading information of the next processing, the processing is cycled down until the end, and the finally processed array is taken as the first analysis space final description array. It is easy to understand that all the separate images are processed once, so as to obtain a first analysis space temporary description array and a first analysis space final description array which respectively correspond to the separate images.
And 130, respectively extracting second analysis space description arrays of the plurality of separated images to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array.
And 140, fusing the description arrays through the first analysis space temporary description arrays and the second analysis space temporary description arrays corresponding to the plurality of separation images to obtain a target fusion description array corresponding to the plurality of separation images.
The process of the description array fusion is to negotiate the image content by the first analysis space temporary description array and the corresponding second analysis space temporary description array, so that the stability of defect identification is improved, and a deeper paraphrasing array can be obtained. The target fusion description array is a paraphrase array obtained after the first analysis space description array and the second analysis space description array are fused. For example, the first analysis space temporary description array and the second analysis space temporary description array corresponding to the separated images are combined to obtain the target fusion description array corresponding to the separated images, and the first analysis space temporary description array and the second analysis space temporary description array corresponding to each separated image are combined to obtain the target fusion description array corresponding to each separated image.
And 150, extracting an image interpretation description array through a first analysis space final description array, a second analysis space final description array and a target fusion description array corresponding to the plurality of separated images to obtain an image interpretation description array corresponding to the plurality of separated images, and detecting image defects through the image interpretation description array to obtain a flexible circuit board defect credible coefficient corresponding to the plurality of separated images.
The image interpretation description array is an interpretation array obtained by integrating the first analysis space description array, the second analysis space description array and the fusion description array, and each separated image comprises a corresponding image interpretation description array. The image defect detection is a process of judging an image, specifically, judging whether the image has a flexible circuit board defect, and the detection result comprises the defect of the flexible circuit board and the defect of the flexible circuit board. The reliability coefficient of the soft circuit board defect is used for indicating the confidence level that the corresponding separated image is the soft circuit board defect, and the greater the reliability coefficient of the soft circuit board defect is, the higher the confidence level that the corresponding separated image is the soft circuit board defect is. For example, the image interpretation description arrays are obtained by aggregating the first analysis space final description arrays, the second analysis space final description arrays and the target fusion description arrays corresponding to each separated image, so as to obtain aggregated description arrays, and further obtain the image interpretation description arrays corresponding to each separated image. And then performing defect identification through the image interpretation description array, and determining whether the separated image is a flexible circuit board defect or a non-flexible circuit board defect to obtain a flexible circuit board defect credible coefficient corresponding to each separated image.
Step 160, determining a plurality of defect image chains from the image sequence to be identified through the flexible circuit board defect credibility coefficient, and determining a defect definition description array corresponding to the plurality of defect image chains through the image definition description array.
The defect image chain is an image combination obtained by combining a plurality of continuous separated images with flexible circuit board defects, and concretely, reference may be made to a ', B', C ', D' in fig. 2, where the separated images with flexible circuit board defects are separated images with flexible circuit board defect reliability coefficients greater than preset flexible circuit board defect reliability coefficients, and the preset flexible circuit board defect reliability coefficients are preset values, and the sizes of the preset flexible circuit board defect reliability coefficients can be selected according to practical situations. The defect definition description array can indicate definition of a defect image chain, and is obtained by fusing the image definition description arrays corresponding to the defects of each flexible circuit board, for example, comparing the reliability coefficient of the defects of the flexible circuit board corresponding to each separated image with the reliability coefficient of the defects of the preset flexible circuit board, and if the reliability coefficient of the defects of the flexible circuit board is larger than the reliability coefficient of the defects of the preset flexible circuit board, the separated image corresponding to the reliability coefficient of the defects of the flexible circuit board has the defects of the flexible circuit board. And combining the separated images corresponding to the defects of the flexible circuit boards belonging to one image package in the image sequence to be identified into a defect image chain through spatial ordering of the images to obtain a plurality of defect image chains, and combining the image definition description arrays corresponding to the defects of the flexible circuit boards in the defect image chains to obtain a defect definition description array corresponding to the defect image chains, wherein each defect image chain is processed to obtain a corresponding defect definition description array of each defect image chain.
And step 170, performing defect identification on the defect image chains through the defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination.
The defect image chain defect identification is used for determining whether the defect image chains are identical defect image chains, and the unified defect image chain combination comprises identical defect image chains (such as I type defects and II type defects in fig. 2), wherein the identical defect image chains are defect image chains with matching degree larger than preset matching degree.
The plurality of defect image chains can be grouped through the defect definition description arrays corresponding to the plurality of defect image chains, for example, a preset grouping operator such as a mean value algorithm is adopted, and a grouping result is obtained through distance clustering of the description arrays, so that at least one unified defect image chain combination is obtained.
According to the soft circuit board defect detection method based on image processing, a plurality of separation images are obtained through separation of an image sequence to be identified, first analysis space description arrays are extracted from the plurality of separation images respectively to obtain a first analysis space temporary description array and a first analysis space final description array, second analysis space description arrays are extracted from the plurality of separation images respectively to obtain a second analysis space temporary description array and a second analysis space final description array, description array fusion is conducted through the first analysis space temporary description arrays and the second analysis space temporary description arrays corresponding to the plurality of separation images to obtain target fusion description arrays corresponding to the plurality of separation images, and the target fusion description arrays obtained through description array fusion enable the target fusion description arrays to contain nesting information of different analysis spaces. Then, extracting the image interpretation description arrays through the first analysis space final description arrays, the second analysis space final description arrays and the target fusion description arrays corresponding to the plurality of separated images to obtain the image interpretation description arrays corresponding to the plurality of separated images, so that the extracted image interpretation description arrays simultaneously contain the information of the two analysis spaces, and meanwhile, the extracted image interpretation description arrays can restore the basic properties of the images to the greatest extent. Then, performing image defect detection through the image interpretation description array to obtain a flexible circuit board defect credible coefficient corresponding to a plurality of separated images, so that the accuracy of image defect detection can be improved, then determining a plurality of defect image chains from an image sequence to be identified through the flexible circuit board defect credible coefficient, and determining a defect interpretation description array corresponding to the plurality of defect image chains through the image interpretation description array; and carrying out defect identification on the defect image chains through the defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination, so that the accuracy of defect identification of the defect image chains is improved, and further, the accuracy of the obtained unified defect image chain combination is ensured.
In one implementation of the embodiment of the present application, for step 170, performing defect identification on a defect image chain through a defect definition description array corresponding to a plurality of defect image chains to obtain a unified defect image chain combination, including:
and 171, extracting the designated array through the defect definition description arrays corresponding to the defect image chains to obtain a selected description array.
Specifically, the embodiment of the application provides a description array processing model, which can be built by adopting a transducer architecture, wherein the designated array extraction is the process of encoding the Encoder in the description array processing model, which is essentially the process of extracting specific characteristics (description arrays), and the selection of the description arrays completes the encoding result of the defect definition description arrays of the designated array extraction. For training of the description array processing model, a general training process can be adopted, namely training data (sample-sample label) is prepared, a preset model is input, an inference result is obtained, comparison and cost of the sample label are compared, model parameters are updated, and iteration is carried out until convergence, and in addition, the description array processing model can be used by directly selecting an existing adaptive model. the transducer is a mature model, and specific structures can be referred to in the prior art, and will not be described herein.
Step 172, analyzing and restoring the soft circuit board defect credible coefficients corresponding to the selected description array and the plurality of separated images to obtain a target defect definition description array corresponding to the plurality of defect image chains.
And then, in step 171, the parsing and restoring process is to describe the result of the Decoder output by the Decoder in the array processing model, for example, the soft circuit board defect credible coefficient of the separated image corresponding to the current defect image chain is identified in the soft circuit board defect credible coefficients corresponding to the plurality of separated images, then, the selected description array corresponding to the defect image chain and the soft circuit board defect credible coefficient of the separated image corresponding to the current defect image chain are put into the Decoder describing the array processing model, so as to obtain the target defect definition description array corresponding to the defect image chain, and all the defect image chains perform the above processes, so as to obtain the target defect definition description array corresponding to all the defect image chains.
Step 173, performing defect recognition on the multiple defect image chains through the target defect definition description arrays corresponding to the multiple defect image chains to obtain a unified defect image chain combination.
For example, the target defect definition description arrays corresponding to the defect image chains are grouped by a mean algorithm to obtain a plurality of defect image chains, and each type of defect image chain is determined to be the same defect image chain, so as to obtain a unified defect image chain combination of unified types, such as type I defects and type II defects in fig. 2.
In one implementation of the embodiment of the present application, for step 171, performing specified array extraction through defect definition description arrays corresponding to a plurality of defect image chains to obtain a selected description array, including: extracting original image description arrays corresponding to the multiple separated images respectively, and determining defect image chain original description arrays corresponding to the multiple defect image chains from the original image description arrays corresponding to the multiple separated images respectively; integrating the original description arrays of the defect image chains corresponding to the plurality of defect image chains with the corresponding defect definition description arrays respectively to obtain target integration description arrays corresponding to the plurality of defect image chains; and putting the target integrated description arrays corresponding to the defect image chains into a description array extraction network of the description array processing model for processing to obtain a target selected description array.
The original image description array is the most basic description information of the image, the original description array of the defect image chain is the original image description array corresponding to the defect image chain, the original image description arrays of a plurality of separated images corresponding to the defect image chain are integrated, the target integrated description array is the description array after the original characteristic information is combined, and the target selected description array is the selected description array after the original characteristic information is combined. For example, the original image description arrays corresponding to the separate images are extracted, then the original image description arrays of the separate images corresponding to the defect image chains are integrated to obtain the original description arrays of the defect image chains corresponding to each defect image chain, for example, the original image description arrays of the separate images corresponding to the defect image chains are connected (e.g. spliced), then the original description arrays of the defect image chains corresponding to each defect image chain are respectively connected with the defect definition description arrays corresponding to each defect image chain to obtain the target integration description arrays corresponding to the defect image chains, and then the target integration description arrays corresponding to the defect image chains are sequentially put into the description array extraction network of the description array processing model to be processed to obtain the target selected description arrays. Through the process, the original description arrays of the defect image chain are respectively integrated with the corresponding defect definition description arrays and then are processed, so that the accuracy of the obtained target selected description arrays is improved, and the accuracy of the obtained target defect definition description arrays is further improved.
In one implementation of the examples of the present application, for step 173: performing defect identification on the multiple defect image chains through a target defect definition description array corresponding to the multiple defect image chains to obtain a unified defect image chain combination, wherein the method comprises the following steps of: determining the matching degree among a plurality of defect image chains through a target defect definition description array corresponding to the defect image chains; and performing image barrel division according to the matching degree among the defect image chains to obtain a unified defect image chain combination.
The matching degree between the plurality of defect image chains represents the proximity degree between the plurality of defect image chains, the matching degree can be obtained by calculating vector distances (such as Euclidean distance, manhattan distance and cosine distance) of the plurality of defect image chains, and the distance and the matching degree are in negative correlation. If the target defect definition description arrays corresponding to the defect image chains are used, determining the target defect definition description arrays A and the target defect definition description arrays B in the target defect definition description arrays corresponding to the defect image chains, determining the matching degree of the target defect definition description arrays A and the target defect definition description arrays B, determining the matching degree among all the target defect definition description arrays, classifying all the matching degrees in a barrel, integrating the defect image chains corresponding to the target defect definition description arrays with the matching degree larger than the preset matching degree into a unified defect image chain combination, performing image barrel division in a mode of determining the matching degree, omitting the process of pre-specifying the mass center number in barrel division, and facilitating the improvement of the accuracy and the acquisition speed of the acquired unified defect image chain combination.
In one implementation of the examples of the present application, for step 120: extracting a first analysis space description array from the plurality of separated images to obtain a first analysis space description array corresponding to the plurality of separated images, wherein the first analysis space description array comprises a first analysis space temporary description array and a first analysis space final description array, and the method comprises the following steps: respectively adopting a first analysis space convolution operation for a plurality of separated images to obtain a plurality of temporary convolution description arrays and ending convolution description arrays corresponding to the plurality of separated images; carrying out dimension transformation operation on the temporary convolution description arrays to obtain a plurality of first analysis space temporary description arrays corresponding to the plurality of separated images; and carrying out dimension transformation operation on the ending convolution description array to obtain a first analysis space final description array corresponding to the plurality of separated images.
The first analysis space convolution operation can acquire a description array of an image in a first analysis space, the ending convolution description array is a finally obtained convolution description array, the temporary convolution description array is a convolution description array obtained before the ending convolution description array is obtained, and the dimension transformation operation is to transform the first analysis space description array into an operation with the same dimension as the second analysis space description array. For example, a first analysis space convolution operation is specifically performed on each of the separate images to obtain a plurality of temporary convolution description arrays and a final convolution description array corresponding to each of the separate images, then dimension transformation is performed on each of the temporary convolution description arrays to obtain a plurality of first analysis space temporary description arrays corresponding to the plurality of separate images, and dimension transformation is performed on the final convolution description arrays to obtain a first analysis space final description array corresponding to the plurality of separate images.
In one implementation of the examples of the present application, for step 130: extracting the plurality of separated images by adopting a second analysis space description array respectively to obtain a second analysis space description array corresponding to the plurality of separated images, wherein the second analysis space description array comprises a second analysis space temporary description array and a second analysis space final description array, and the method comprises the following steps of: extracting original image description arrays corresponding to the multiple separated images respectively; performing a second analysis space convolution operation on the original image description arrays corresponding to the plurality of separated images to obtain a plurality of second analysis space temporary description arrays and a second analysis space final description array corresponding to the plurality of separated images, wherein the second analysis space convolution operation can obtain the description arrays of the images in the second analysis space, extract the description arrays of the images to the original image description arrays corresponding to the plurality of separated images, then perform the second analysis space convolution operation on the original image description arrays for a plurality of times (equal to the number of times of the first analysis space convolution operation), finally obtain the second analysis space final description arrays, obtain the second analysis space temporary description arrays before obtaining the second analysis space final description arrays, and finally obtain a plurality of second analysis space temporary description arrays and second analysis space final description arrays corresponding to the plurality of separated images. Extracting original image description arrays corresponding to the multiple separated images respectively; and then, a second analysis space convolution operation is adopted by the original image description array, so that a plurality of second analysis space temporary description arrays and second analysis space final description arrays corresponding to a plurality of separated images are obtained, and the accuracy of the obtained second analysis space description arrays is improved.
In one implementation of the embodiments of the present application, the first analysis space temporary description array includes a plurality and the second analysis space temporary description array includes a plurality. For step 140: performing description array fusion through a first analysis space temporary description array and a second analysis space temporary description array corresponding to a plurality of separation images to obtain a target fusion description array corresponding to the plurality of separation images, wherein the method comprises the following steps: step 141, integrating the first analysis space temporary description array tensor1-1 in the plurality of first analysis space temporary description arrays with the corresponding second analysis space temporary description array tensor2-1 in the plurality of second analysis space temporary description arrays to obtain a first integrated description array, and performing convolution processing on the first integrated description array to obtain a first fusion description array. The integrated description array is a description array obtained by adding the description arrays, and the fused description array is a description array obtained by fusing the description arrays. For example, a first analysis space temporary description array tensor1-1 and a corresponding second analysis space temporary description array tensor2-1 are obtained, the first analysis space temporary description array tensor1-1 and the corresponding second analysis space temporary description array tensor2-1 are obtained according to a first convolution calculation, then the first analysis space temporary description array tensor1-1 and the corresponding second analysis space temporary description array tensor2-1 are subjected to splicing processing at a set level (channel and dimension) to obtain a first integrated description array, and then the first integrated description array is subjected to convolution processing to obtain a first fusion description array. And step 142, integrating the first fusion description array, the first analysis space temporary description array tensor1-2 in the plurality of first analysis space temporary description arrays and the corresponding second analysis space temporary description array tensor2-2 in the plurality of second analysis space temporary description arrays to obtain a second integration description array, and performing convolution processing on the second integration description array to obtain a second fusion description array. In the step, in the process of fusing the first analysis space temporary description array and the second analysis space temporary description array, the first fusion description array obtained in the previous time is integrated together to obtain a second integration description array, and then convolution processing is carried out on the second integration description array to obtain a second fusion description array. In step 143, a target fusion description array is obtained when the plurality of first analysis space temporary description arrays and the plurality of second analysis space temporary description arrays are completed. For example, the first analysis space temporary description arrays and the corresponding second analysis space temporary description arrays are sequentially fused to obtain the last fused description arrays, the last fused description arrays are integrated with the current first analysis space temporary description arrays and the current second analysis space temporary description arrays, convolution processing is then carried out on the integrated description arrays to obtain the current fused description arrays, and when the description arrays are fused finally, the fused description arrays, the final first analysis space temporary description arrays and the final second analysis space temporary description arrays are integrated to obtain the final integrated description arrays, and convolution processing is then carried out on the final integrated description arrays to obtain the target integrated description arrays. According to the method and the device, the first analysis space temporary description array and the corresponding second analysis space temporary description array are fused to enable the description arrays of different analysis spaces to be nested and complement each other, in addition, the top network is clear of information contained in the lower network, and the acquired target integrated description array is enabled to have accuracy.
In one implementation of the examples of the present application, step 150: extracting the image interpretation description array through the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images to obtain the image interpretation description array corresponding to the plurality of separated images, and detecting the image defect through the image interpretation description array to obtain the soft circuit board defect credible coefficient corresponding to the plurality of separated images, wherein the method comprises the following steps:
and 151, integrating the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separation images to obtain a target integration description array corresponding to the plurality of separation images.
And 152, performing convolution processing on the target integrated description arrays corresponding to the plurality of separated images to obtain convolution description arrays corresponding to the plurality of separated images.
The target integration description array is a description array obtained by integrating the first analysis space final description array, the second analysis space final description array and the target fusion description array, and the convolution description array is obtained by adopting convolution processing to the target integration description array. For example, the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to each separation image are connected on a channel in sequence to obtain the target integration description array corresponding to each separation image, and then convolution processing is performed on the target integration description array corresponding to each separation image to obtain convolution description arrays corresponding to a plurality of separation images.
And step 153, determining the description array maximum value and the description array equal value corresponding to each order in the convolution description array through the convolution description arrays corresponding to the plurality of separated images.
And 154, summing the maximum value of the description array and the equal value of the description array to obtain paraphrase extraction description array values corresponding to all the steps in the convolution description array, and obtaining paraphrase extraction description arrays corresponding to a plurality of separated images through the paraphrase extraction description array values corresponding to all the steps in the convolution description array.
The maximum value of the description array is the largest description array in all the description arrays corresponding to the corresponding order (the dimension of the description array can be understood), the equal value of the description arrays is the average result of all the description arrays corresponding to the corresponding order, and the paraphrasing extraction description array value is the description array representing the paraphrasing description array of the image obtained by extraction. For example, the paraphrase extraction description arrays corresponding to each separate image are sequentially determined, the convolution description arrays corresponding to the separate images to be acquired at present are determined, and then the description array maximum values and the description array equal values corresponding to the respective steps in the convolution description arrays are acquired, in other words, the description array equal values and the description array maximum values of all the description arrays corresponding to the respective steps are determined. And then, summing the maximum value of the description array and the equal value of the description array to obtain the paraphrase extraction description array value corresponding to each order in the convolution description array, and taking the paraphrase extraction description array value corresponding to each order as the paraphrase extraction description array of the current separation image.
And 155, linearly transforming the paraphrasing extraction description arrays corresponding to the plurality of separated images to obtain the image paraphrasing description arrays corresponding to the plurality of separated images.
Step 156, performing detection and judgment of defects and non-defects of the flexible circuit board through the image definition description arrays corresponding to the plurality of separated images, and obtaining the credible coefficients of the defects of the flexible circuit board corresponding to the plurality of separated images.
For example, the paraphrasing extraction description arrays corresponding to the separate images are sequentially subjected to linear transformation through a Sigmoid function to obtain image paraphrasing description arrays corresponding to the separate images, and then a classifier (normalized) is selected for soft circuit board defect and non-defect detection judgment through the image paraphrasing description arrays to obtain soft circuit board defect credible coefficients corresponding to the separate images. The description array maximum value and the description array equal value are determined, the description array is extracted through the description array maximum value and the description array equal value, the description array maximum value can reflect the most prominent description information, the description array equal value can reflect the description information of the whole image, the extracted and obtained image definition description array is helped to contain better accuracy, defect identification is carried out through the image definition description array, and accuracy of the obtained soft circuit board defect credible coefficient is facilitated to be improved.
In an implementation manner of the embodiment of the present application, the method for detecting a defect of a flexible circuit board based on image processing further includes:
and 10, putting the image sequence to be identified into an image defect detection model, and separating the image sequence to be identified through the image defect detection model to obtain a plurality of separated images.
Step 20, extracting a first analysis space description array from a plurality of separated images through an image defect detection model to obtain a first analysis space description array corresponding to the plurality of separated images, wherein the first analysis space description array comprises a first analysis space temporary description array and a first analysis space final description array; and respectively extracting the second analysis space description arrays of the plurality of separated images to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array.
And step 30, performing description array fusion on the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separated images through the image defect detection model to obtain a target fusion description array corresponding to the plurality of separated images.
And step 40, extracting the image interpretation description arrays of the first analysis space final description arrays, the second analysis space final description arrays and the target fusion description arrays corresponding to the plurality of separated images through the image defect detection model to obtain the image interpretation description arrays corresponding to the plurality of separated images, and detecting the image defects through the image interpretation description arrays to obtain the soft circuit board defect credible coefficients corresponding to the plurality of separated images.
In this embodiment, the image defect detection model is configured to identify defects and non-defects in the image sequence, and the image defect detection model may be built by unrestricted selection of a machine learning or deep learning network, such as a convolutional neural network. In practical application, firstly, an image sequence to be identified is acquired, the image sequence to be identified is put into an image defect detection model, the image defect detection model can comprise two sub-networks, a second analysis space final description array and a first analysis space final description array corresponding to the image sequence to be identified are extracted based on the two sub-modules, the description arrays are fused, the extracted second analysis space temporary description array and the first analysis space temporary description array are fused to obtain a target fusion description array, then, an paraphrasing array is extracted through the obtained second analysis space final description array, the first analysis space final description array and the target fusion description array, and image defect detection is carried out through the extracted paraphrasing array. By adopting the process, the image defect detection is carried out through the image defect detection model, so that the reliability coefficient of the soft circuit board defect corresponding to the multiple separated images is obtained, and the process of image defect detection is accelerated.
In one implementation of the embodiment of the present application, the image defect detection model includes a first analysis space description array extraction network, a second analysis space description array extraction network, a description array fusion network, an image interpretation description array extraction network, and a defect discrimination network; the method for detecting the defects of the flexible circuit board based on image processing can further comprise the following steps:
step 101, inputting the image sequence to be identified into an image defect detection model, and separating the image sequence to be identified through the image defect detection model to obtain a plurality of separated images.
And step 102, putting a plurality of separated images into a first analysis space description array extraction network to extract a second analysis space description array, and obtaining a first analysis space temporary description array and a first analysis space final description array.
And step 103, putting the plurality of separated images into a second analysis space description array extraction network to extract the second analysis space description array, and obtaining a second analysis space temporary description array and a second analysis space final description array.
And 104, inputting the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separated images into a description array fusion network to fuse the description arrays, so as to obtain a target fusion description array corresponding to the plurality of separated images.
And 105, inputting the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images into an image interpretation description array extraction network to extract the image interpretation description array, obtaining the image interpretation description array corresponding to the plurality of separated images, inputting the image interpretation description array into a defect discrimination network to detect the image defects, and obtaining the soft circuit board defect credible coefficients corresponding to the plurality of separated images.
The first analysis space description array extraction network is configured to extract a first analysis space description array of the image, the second analysis space description array extraction network is configured to extract a second analysis space description array of the image, the description array fusion network is configured to fuse the second analysis space temporary description array and the first analysis space temporary description array, the image interpretation description array extraction network is configured to extract a interpretation array of the image, and the defect discrimination network is configured to judge defects and non-defects.
For example, a plurality of separate images are input into a first analysis space description array extraction network to perform second analysis space description array extraction, that is, a convolution module based on the first analysis space description array extraction network generates a first analysis space description array, a final convolution module based on the final convolution module generates a first analysis space temporary description array, a previous convolution module generates a first analysis space temporary description array, a plurality of separate images are input into the second analysis space description array extraction network to perform second analysis space description array extraction, that is, a convolution module based on the second analysis space description array extraction network generates a second analysis space description array, the final convolution module generates a second analysis space final description array, the previous convolution module obtains the second analysis space temporary description array, and the convolution times of the first analysis space description array extraction network and the second analysis space description array extraction network are equal. And performing description array fusion on the first analysis space temporary description array and the second analysis space temporary description array through a description array fusion network to obtain a target fusion description array, extracting an image definition description array based on an image definition description array extraction network, and performing image defect detection through a defect discrimination network to obtain soft circuit board defect credible coefficients corresponding to a plurality of separated images.
In one implementation of the embodiment of the application, the image defect detection model may include a shared network and two sub-networks corresponding to different analysis spaces respectively, the image defect detection model inputs the image sequence to be identified into the two sub-networks respectively, and the first analysis space final description array and the second analysis space final description array including the same dimension and channel are obtained by adopting a plurality of rolling and pooling operations.
In one implementation of the examples of the present application, the image defect detection model is optimally trained by:
step 100, an image sequence sample and a matched indicator are obtained.
The image sequence sample is an image sequence required by model training, and the indication mark is a real record corresponding to the image sequence sample, such as whether the mark is defective or non-defective.
And 200, putting the image sequence sample into a to-be-trained image defect detection model, and separating the image sequence sample by using the to-be-trained image defect detection model to obtain a plurality of separated image samples.
Step 300, extracting a first analysis space description array from each separated image sample by using a to-be-trained image defect detection model to obtain an original first analysis space description array corresponding to a plurality of separated image samples, wherein the original first analysis space description array comprises an original first analysis space temporary description array and an original first analysis space final description array; and respectively extracting second analysis space description arrays of each separated image sample to obtain an original second analysis space description array corresponding to the plurality of separated image samples, wherein the original second analysis space description array comprises an original second analysis space temporary description array and an original second analysis space final description array.
And 400, performing description array fusion on the original first analysis space temporary description array and the original second analysis space temporary description array corresponding to each separated image sample by using the to-be-trained image defect detection model to obtain original fusion description arrays corresponding to a plurality of separated image samples.
And 500, extracting an image interpretation description array by utilizing a to-be-trained image defect detection model to the original first analysis space final description array, the original second analysis space final description array and the original fusion description array corresponding to each separated image sample to obtain original image interpretation description arrays corresponding to a plurality of separated image samples, and performing image defect detection through the original image interpretation description arrays to obtain original flexible circuit board defect credibility coefficients corresponding to the plurality of separated image samples.
The method comprises the steps that a separated image sample is a separated image obtained in a training process, an original first analysis space description array is a first analysis space description array obtained through parameter extraction to be optimized, an original second analysis space description array is a second analysis space description array obtained through parameter extraction to be optimized, an original flexible circuit board defect trusted coefficient is a flexible circuit board defect trusted coefficient obtained through parameter reasoning to be optimized, for example, a training image defect detection model is built based on a neural network, then a first image defect detection reasoning is carried out on the image sequence sample through the training image defect detection model, the original flexible circuit board defect trusted coefficient corresponding to each separated image sample is obtained, and the image defect detection reasoning carried out by the training image defect detection model and the reasoning of the trained image defect detection model remain the same.
Step 600, obtaining training cost through the original soft circuit board defect credibility coefficient corresponding to each separated image sample and the indication mark matched with the image sequence sample, obtaining cost value, and debugging the to-be-trained image defect detection model through the cost value, thus obtaining a debugged image defect detection model.
And 700, determining the debug image defect detection model as an image defect detection model to be trained, and repeating the training step until the model converges to finally obtain the image defect detection model.
Training the image defect detection model to be trained through the image sequence sample and the matched indication mark to obtain the image defect detection model, wherein the image defect detection model is independently deployed, so that an optimization result is accurate, high in cost and low in cost after optimization is completed, and the inference accuracy of the image defect detection model is improved.
Specifically, an image sequence processing network to be trained is deployed, the image sequence processing network to be trained is optimized through a training sample, an image sequence processing network is obtained, an image sequence to be recognized is separated based on the image sequence processing network, a plurality of separation images are obtained, a first analysis space description array corresponding to the plurality of separation images is obtained through extraction of the first analysis space description array, the first analysis space description array comprises a first analysis space temporary description array and a first analysis space final description array, extraction of the second analysis space description array is carried out on the plurality of separation images respectively, a second analysis space description array corresponding to the plurality of separation images is obtained, the second analysis space description array comprises a second analysis space temporary description array and a second analysis space final description array, the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separation images are fused, a target fusion array corresponding to the plurality of separation images is obtained, a soft defect array is obtained through extraction of a plurality of release line corresponding to the plurality of separation space temporary description arrays, a soft defect array is detected through a plurality of release line corresponding to the plurality of separation image, a soft defect array is detected through a soft defect array is detected, and carrying out defect identification on the defect image chains through defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination.
In addition, in one implementation of the embodiment of the present application, the method for detecting the defects of the flexible circuit board based on the image processing is performed by a defect detection system, and then the method includes the following steps:
step I, acquiring an image sequence to be identified, putting the image sequence to be identified into an image defect detection model, separating the image sequence to be identified through the image defect detection model to obtain a plurality of separated images, wherein the image defect detection model comprises a first analysis space description array extraction network, a second analysis space description array extraction network, a description array fusion network, an image interpretation description array extraction network and a defect discrimination network.
And II, putting the plurality of separated images into a first analysis space description array extraction network to perform first analysis space convolution operation to obtain a temporary convolution description array and a final convolution description array corresponding to the plurality of separated images, and performing dimension transformation operation on the temporary convolution description array and the final convolution description array to obtain a first analysis space temporary description array and a first analysis space description array corresponding to the plurality of separated images.
And III, extracting original image description arrays corresponding to the multiple separated images, putting the original image description arrays corresponding to the multiple separated images into a second analysis space description array extraction network to perform second analysis space convolution operation, and obtaining a second analysis space temporary description array and a second analysis space final description array corresponding to the multiple separated images. And integrating the first analysis space temporary description array with the second analysis space temporary description array to obtain a first integrated description array, and performing convolution processing on the first integrated description array to obtain a target fusion description array.
And IV, inputting a first analysis space final description array, a second analysis space final description array and a target fusion description array corresponding to the plurality of separation images into an image interpretation description array extraction network for integration to obtain a target integration description array corresponding to the plurality of separation images, performing convolution processing on the target integration description array corresponding to the plurality of separation images to obtain a convolution description array corresponding to the plurality of separation images, determining equal values of description array maximum values and description arrays corresponding to each order in the convolution description array through the convolution description arrays corresponding to the plurality of separation images, and summing the equal values of the description array maximum values and the description arrays to obtain interpretation extraction description array values corresponding to each order in the convolution description array, and obtaining the interpretation extraction description array corresponding to the plurality of separation images through the interpretation extraction description array values corresponding to each order in the convolution description array.
And V, inputting the image definition description array into a defect judging network to detect and judge defects and non-defects of the flexible circuit board, and obtaining the credible coefficients of the defects of the flexible circuit board corresponding to the plurality of separated images. And determining a plurality of defect image chains from the image sequence to be identified through the soft circuit board defect credibility coefficients corresponding to the plurality of separated images, and determining a defect definition description array corresponding to the plurality of defect image chains through the image definition description array.
And step VI, inputting the defect definition description arrays corresponding to the plurality of defect image chains into a description array extraction network of the description array processing model to perform specified array extraction to obtain selected description arrays corresponding to the plurality of defect image chains, and inputting the selected description arrays corresponding to the plurality of defect image chains and corresponding soft circuit board defect credibility coefficients into an analysis and reduction network of the description array processing model to perform analysis and reduction to obtain target defect definition description arrays corresponding to the plurality of defect image chains.
And step VII, determining the matching degree among the plurality of defect image chains through the target defect definition description arrays corresponding to the plurality of defect image chains, and carrying out image barrel division through the matching degree among the plurality of defect image chains to obtain a unified defect image chain combination.
In summary, according to the image processing-based flexible circuit board defect detection method provided by the embodiment of the application, a plurality of separation images are obtained by separating an image sequence to be identified, a first analysis space description array is respectively extracted from the plurality of separation images to obtain a first analysis space temporary description array and a first analysis space final description array, a second analysis space description array is respectively extracted from the plurality of separation images to obtain a second analysis space temporary description array and a second analysis space final description array, and then the description arrays are fused through the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separation images to obtain a target fusion description array corresponding to the plurality of separation images, and the target fusion description array obtained through the fusion of the description arrays contains nesting information of different analysis spaces. Then, extracting the image interpretation description arrays through the first analysis space final description arrays, the second analysis space final description arrays and the target fusion description arrays corresponding to the plurality of separated images to obtain the image interpretation description arrays corresponding to the plurality of separated images, so that the extracted image interpretation description arrays simultaneously contain the information of the two analysis spaces, and meanwhile, the extracted image interpretation description arrays can restore the basic properties of the images to the greatest extent. Then, performing image defect detection through the image interpretation description array to obtain a flexible circuit board defect credible coefficient corresponding to a plurality of separated images, so that the accuracy of image defect detection can be improved, then determining a plurality of defect image chains from an image sequence to be identified through the flexible circuit board defect credible coefficient, and determining a defect interpretation description array corresponding to the plurality of defect image chains through the image interpretation description array; and carrying out defect identification on the defect image chains through the defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination, so that the accuracy of defect identification of the defect image chains is improved, and further, the accuracy of the obtained unified defect image chain combination is ensured.
Based on the same principle as the method shown in fig. 1, there is also provided a defect detection apparatus 10 according to an embodiment of the present application, as shown in fig. 3, the apparatus 10 includes:
the separation module 11 is configured to obtain an image sequence to be identified, and separate the image sequence to be identified to obtain a plurality of separated images; the image sequence to be identified comprises a plurality of images of the flexible circuit board to be detected.
The first extraction module 12 is configured to extract a first analysis space description array from the plurality of separate images, so as to obtain a first analysis space description array corresponding to the plurality of separate images, where the first analysis space description array includes a first analysis space temporary description array and a first analysis space final description array.
And the second extraction module 13 is configured to extract the second analysis space description arrays of the multiple separate images respectively, so as to obtain second analysis space description arrays corresponding to the multiple separate images, where the second analysis space description arrays include a second analysis space temporary description array and a second analysis space final description array.
And the fusion module 14 is configured to fuse the description arrays through the first analysis space temporary description arrays and the second analysis space temporary description arrays corresponding to the multiple separate images, so as to obtain a target fusion description array corresponding to the multiple separate images.
The trusted coefficient determining module 15 is configured to extract an image interpretation description array through the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images, obtain an image interpretation description array corresponding to the plurality of separated images, and perform image defect detection through the image interpretation description array, so as to obtain a trusted coefficient of the soft circuit board defect corresponding to the plurality of separated images.
The defect determining module 16 is configured to determine a plurality of defect image chains from the image sequence to be identified according to the flexible circuit board defect credibility coefficient, and determine a defect definition description array corresponding to the plurality of defect image chains according to the image definition description array.
And the combination module 17 is used for carrying out defect identification on the defect image chains through the defect definition description arrays corresponding to the defect image chains to obtain a unified defect image chain combination.
The above embodiment describes the defect detecting device 10 from the viewpoint of a virtual module, and the following describes a defect detecting system from the viewpoint of a physical module, specifically as follows:
an embodiment of the present application provides a defect detection system, as shown in fig. 4, the defect detection system 100 includes: a processor 101 and a memory 103. Wherein the processor 101 is coupled to the memory 103, such as via bus 102. Optionally, the defect detection system 100 may also include a transceiver 104. It should be noted that, in practice, the transceiver 104 is not limited to one, and the structure of the defect detecting system 100 is not limited to the embodiment of the present application.
The processor 101 may be a CPU, general purpose processor, GPU, DSP, ASIC, FPGA or other programmable logic device, transistor logic device, hardware component, or any combination thereof. Which may implement or perform the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. The processor 101 may also be a combination that implements computing functionality, e.g., comprising one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.
Bus 102 may include a path to transfer information between the aforementioned components. Bus 102 may be a PCI bus or an EISA bus, etc. The bus 102 may be classified as an address bus, a data bus, a control bus, or the like. For ease of illustration, only one thick line is shown in fig. 4, but not only one bus or one type of bus.
Memory 103 may be, but is not limited to, a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, an EEPROM, a CD-ROM or other optical disk storage, optical disk storage (including compact disks, laser disks, optical disks, digital versatile disks, blu-ray disks, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.
The memory 103 is used for storing application program codes for executing the present application and is controlled to be executed by the processor 101. The processor 101 is configured to execute application code stored in the memory 103 to implement what is shown in any of the method embodiments described above.
An embodiment of the present application provides a defect detection system, where the defect detection system in the embodiment of the present application includes: one or more processors; a memory; one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, which when executed by the one or more processors, implement the image processing-based flexible circuit board defect detection method described above. According to the technical scheme, a plurality of separated images are obtained by separating an image sequence to be identified, then a first analysis space description array extraction is respectively carried out on the plurality of separated images to obtain a first analysis space temporary description array and a first analysis space final description array, a second analysis space description array extraction is respectively carried out on the plurality of separated images to obtain a second analysis space temporary description array and a second analysis space final description array, description array fusion is carried out on the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separated images to obtain a target fusion description array corresponding to the plurality of separated images, and the target fusion description array obtained through the description array fusion order contains nesting information of different analysis spaces. Then, extracting the image interpretation description arrays through the first analysis space final description arrays, the second analysis space final description arrays and the target fusion description arrays corresponding to the plurality of separated images to obtain the image interpretation description arrays corresponding to the plurality of separated images, so that the extracted image interpretation description arrays simultaneously contain the information of the two analysis spaces, and meanwhile, the extracted image interpretation description arrays can restore the basic properties of the images to the greatest extent. Then, performing image defect detection through the image interpretation description array to obtain a flexible circuit board defect credible coefficient corresponding to a plurality of separated images, so that the accuracy of image defect detection can be improved, then determining a plurality of defect image chains from an image sequence to be identified through the flexible circuit board defect credible coefficient, and determining a defect interpretation description array corresponding to the plurality of defect image chains through the image interpretation description array; and carrying out defect identification on the defect image chains through the defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination, so that the accuracy of defect identification of the defect image chains is improved, and further, the accuracy of the obtained unified defect image chain combination is ensured.
Embodiments of the present application provide a computer readable storage medium having a computer program stored thereon, which when executed on a processor, enables the processor to perform the corresponding content of the foregoing method embodiments.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited in order and may be performed in other orders, unless explicitly stated herein. Moreover, at least some of the steps in the flowcharts of the figures may include a plurality of sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, the order of their execution not necessarily being sequential, but may be performed in turn or alternately with other steps or at least a portion of the other steps or stages.
The foregoing is only a partial embodiment of the present application, and it should be noted that, for a person skilled in the art, several improvements and modifications can be made without departing from the principle of the present application, and these improvements and modifications should also be considered as the protection scope of the present application.

Claims (9)

1. The method is characterized by being applied to a defect detection system, and comprises the following steps:
acquiring an image sequence to be identified, and separating the image sequence to be identified to obtain a plurality of separated images; the image sequence to be identified comprises a plurality of images of the flexible circuit board to be detected;
extracting a first analysis space description array from the plurality of separated images respectively to obtain a first analysis space description array corresponding to the plurality of separated images, wherein the first analysis space description array comprises a first analysis space temporary description array and a first analysis space final description array;
extracting second analysis space description arrays of the plurality of separated images respectively to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array;
performing description array fusion through a first analysis space temporary description array and a second analysis space temporary description array corresponding to the plurality of separation images to obtain a target fusion description array corresponding to the plurality of separation images;
Extracting an image interpretation description array through a first analysis space final description array, a second analysis space final description array and a target fusion description array corresponding to the plurality of separated images to obtain an image interpretation description array corresponding to the plurality of separated images, and detecting image defects through the image interpretation description array to obtain a soft circuit board defect credible coefficient corresponding to the plurality of separated images;
determining a plurality of defect image chains from the image sequence to be identified through the flexible circuit board defect credibility coefficient, and determining a defect definition description array corresponding to the plurality of defect image chains through the image definition description array;
performing defect identification on the defect image chains through the defect definition description arrays corresponding to the plurality of defect image chains to obtain a unified defect image chain combination;
the method includes the steps of extracting a first analysis space description array of the plurality of separated images respectively to obtain a first analysis space description array corresponding to the plurality of separated images, wherein the first analysis space description array comprises a first analysis space temporary description array and a first analysis space final description array, and the method comprises the following steps:
Respectively adopting a first analysis space convolution operation for the plurality of separated images to obtain a plurality of temporary convolution description arrays and ending convolution description arrays corresponding to the plurality of separated images;
performing dimension transformation operation on the temporary convolution description arrays to obtain a plurality of first analysis space temporary description arrays corresponding to the plurality of separated images;
performing dimension transformation operation on the ending convolution description array to obtain a first analysis space final description array corresponding to the plurality of separated images;
extracting the second analysis space description arrays of the plurality of separated images to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array, and the method comprises the following steps of:
extracting original image description arrays corresponding to the multiple separated images respectively;
and performing a second analysis space convolution operation on the original image description arrays corresponding to the multiple separation images to obtain multiple second analysis space temporary description arrays and multiple second analysis space final description arrays corresponding to the multiple separation images.
2. The method of claim 1, wherein the performing defect image chain defect recognition by the defect definition description array corresponding to the plurality of defect image chains to obtain a unified defect image chain combination comprises:
performing specified array extraction through defect definition description arrays corresponding to the defect image chains to obtain a selected description array;
analyzing and restoring the soft circuit board defect credible coefficients corresponding to the selected description array and the plurality of separated images to obtain a target defect interpretation description array corresponding to the plurality of defect image chains;
and carrying out defect identification on the plurality of defect image chains through a target defect definition description array corresponding to the plurality of defect image chains to obtain the unified defect image chain combination.
3. The method according to claim 2, wherein the performing the specified array extraction by the defect definition description arrays corresponding to the plurality of defect image chains to obtain the selected description array includes:
extracting original image description arrays corresponding to the multiple separated images respectively, and determining defect image chain original description arrays corresponding to the multiple defect image chains from the original image description arrays corresponding to the multiple separated images respectively;
Integrating the original description arrays of the defect image chains corresponding to the plurality of defect image chains with the corresponding defect definition description arrays respectively to obtain target integration description arrays corresponding to the plurality of defect image chains;
and putting the target integrated description arrays corresponding to the defect image chains into a description array extraction network of the description array processing model for processing to obtain a target selected description array.
4. The method according to claim 2, wherein the performing defect recognition on the plurality of defect image chains through the target defect paraphrasing description arrays corresponding to the plurality of defect image chains to obtain the unified defect image chain combination includes:
determining the matching degree among the defect image chains through the target defect definition description arrays corresponding to the defect image chains;
and performing image barrel division according to the matching degree among the defect image chains to obtain the unified defect image chain combination.
5. The method of claim 1, wherein the number of the first analysis space temporary description arrays is a plurality and the number of the second analysis space temporary description arrays is a plurality;
The method for obtaining the target fusion description array corresponding to the plurality of separation images by fusing the description array through the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separation images comprises the following steps:
integrating a first analysis space temporary description array tensor1-1 in the plurality of first analysis space temporary description arrays with a corresponding second analysis space temporary description array tensor2-1 in the plurality of second analysis space temporary description arrays to obtain a first integrated description array, and performing convolution processing on the first integrated description array to obtain a first fusion description array;
integrating the first fusion description array, a first analysis space temporary description array tensor1-2 in the plurality of first analysis space temporary description arrays and a corresponding second analysis space temporary description array tensor2-2 in the plurality of second analysis space temporary description arrays to obtain a second integration description array, and performing convolution processing on the second integration description array to obtain a second fusion description array;
and obtaining a target fusion description array when the plurality of first analysis space temporary description arrays and the plurality of second analysis space temporary description arrays are completed.
6. The method of claim 1, wherein the extracting the image paraphrasing description array by the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images to obtain the image paraphrasing description array corresponding to the plurality of separated images, and performing image defect detection by the image paraphrasing description array to obtain the flexible circuit board defect credible coefficients corresponding to the plurality of separated images, comprises:
integrating the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separation images to obtain a target integration description array corresponding to the plurality of separation images;
the convolution processing is adopted through the target integration description arrays corresponding to the plurality of separation images, so that the convolution description arrays corresponding to the plurality of separation images are obtained;
acquiring description array maximum values and description array equal values corresponding to each order in the convolution description arrays through the convolution description arrays corresponding to the plurality of separation images;
summing the maximum value of the description array and the equal value of the description array to obtain paraphrase extraction description array values corresponding to all orders in the convolution description array, and obtaining paraphrase extraction description arrays corresponding to the plurality of separated images through the paraphrase extraction description array values corresponding to all orders in the convolution description array;
Performing linear transformation on the paraphrasing extraction description arrays corresponding to the plurality of separated images to obtain image paraphrasing description arrays corresponding to the plurality of separated images;
and performing detection and judgment on defects and non-defects of the flexible circuit board through the image interpretation description arrays corresponding to the plurality of separated images to obtain the credible coefficients of the defects of the flexible circuit board corresponding to the plurality of separated images.
7. The method according to claim 1, wherein the method further comprises:
the image sequence to be identified is put into an image defect detection model, and the image sequence to be identified is separated through the image defect detection model, so that a plurality of separated images are obtained;
extracting first analysis space description arrays of the plurality of separated images through the image defect detection model to obtain first analysis space description arrays corresponding to the plurality of separated images, wherein the first analysis space description arrays comprise a first analysis space temporary description array and a first analysis space final description array;
extracting second analysis space description arrays of the plurality of separated images respectively to obtain second analysis space description arrays corresponding to the plurality of separated images, wherein the second analysis space description arrays comprise a second analysis space temporary description array and a second analysis space final description array;
Performing description array fusion on a first analysis space temporary description array and a second analysis space temporary description array corresponding to the plurality of separated images through the image defect detection model to obtain a target fusion description array corresponding to the plurality of separated images;
and extracting an image interpretation description array from the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images through the image defect detection model to obtain the image interpretation description array corresponding to the plurality of separated images, and detecting the image defect through the image interpretation description array to obtain the soft circuit board defect credible coefficient corresponding to the plurality of separated images.
8. The method of claim 7, wherein the image defect detection model comprises a first analysis space description array extraction network, a second analysis space description array extraction network, a description array fusion network, an image paraphrasing description array extraction network, and a defect discrimination network;
the method further comprises the steps of:
the image sequence to be identified is put into an image defect detection model, and the image sequence to be identified is separated through the image defect detection model, so that a plurality of separated images are obtained;
The plurality of separated images are put into the first analysis space description array extraction network to carry out second analysis space description array extraction, and a first analysis space temporary description array and a first analysis space final description array are obtained;
the plurality of separated images are put into the second analysis space description array extraction network to extract a second analysis space description array, so that a second analysis space temporary description array and a second analysis space final description array are obtained;
the first analysis space temporary description array and the second analysis space temporary description array corresponding to the plurality of separation images are put into the description array fusion network to be subjected to description array fusion, and the target fusion description array corresponding to the plurality of separation images is obtained;
and inputting the first analysis space final description array, the second analysis space final description array and the target fusion description array corresponding to the plurality of separated images into the image interpretation description array extraction network to extract the image interpretation description array to obtain the image interpretation description array corresponding to the plurality of separated images, and inputting the image interpretation description array into the defect discrimination network to detect the image defects to obtain the soft circuit board defect credible coefficients corresponding to the plurality of separated images.
9. A defect detection system comprising a processor and a memory, the memory storing a computer program which, when run by the processor, implements the method of claims 1-8.
CN202211440566.7A 2022-11-17 2022-11-17 Flexible circuit board defect detection method and system based on image processing Active CN115641330B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211440566.7A CN115641330B (en) 2022-11-17 2022-11-17 Flexible circuit board defect detection method and system based on image processing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211440566.7A CN115641330B (en) 2022-11-17 2022-11-17 Flexible circuit board defect detection method and system based on image processing

Publications (2)

Publication Number Publication Date
CN115641330A CN115641330A (en) 2023-01-24
CN115641330B true CN115641330B (en) 2023-06-30

Family

ID=84948196

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211440566.7A Active CN115641330B (en) 2022-11-17 2022-11-17 Flexible circuit board defect detection method and system based on image processing

Country Status (1)

Country Link
CN (1) CN115641330B (en)

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109766925B (en) * 2018-12-20 2021-05-11 深圳云天励飞技术有限公司 Feature fusion method and device, electronic equipment and storage medium
CN113554611B (en) * 2021-07-19 2024-02-02 广东电网有限责任公司 Insulator self-explosion defect detection method and device, terminal and storage medium
CN113516661B (en) * 2021-09-15 2022-02-08 深圳新视智科技术有限公司 Defect detection method and device based on feature fusion
CN113989600A (en) * 2021-09-28 2022-01-28 上海师范大学 Image foreign matter detection method
CN114612411A (en) * 2022-03-04 2022-06-10 腾讯科技（深圳）有限公司 Image tampering detection method, device, equipment and storage medium

Also Published As

Publication number Publication date
CN115641330A (en) 2023-01-24

Similar Documents

Publication Publication Date Title
CN112199375A (en) Cross-modal data processing method and device, storage medium and electronic device
CN111583229A (en) Road surface fault detection method based on convolutional neural network
CN111526099B (en) Internet of things application flow detection method based on deep learning
CN114821282B (en) Image detection device and method based on domain antagonistic neural network
CN110533654A (en) The method for detecting abnormality and device of components
US20210096530A1 (en) System and method for identifying manufacturing defects
CN114170184A (en) Product image anomaly detection method and device based on embedded feature vector
CN111210412B (en) Packaging detection method and device, electronic equipment and storage medium
CN116823793A (en) Device defect detection method, device, electronic device and readable storage medium
CN115546223A (en) Method and system for detecting loss of fastening bolt of equipment under train
CN113962324A (en) Picture detection method and device, storage medium and electronic equipment
CN115641330B (en) Flexible circuit board defect detection method and system based on image processing
CN114596244A (en) Infrared image identification method and system based on visual processing and multi-feature fusion
CN116580232A (en) Automatic image labeling method and system and electronic equipment
CN116028893A (en) Intelligent data monitoring method, system and electronic equipment
Ziabari et al. A two-tier convolutional neural network for combined detection and segmentation in biological imagery
CN112529881B (en) Power control cabinet cable anomaly identification method and device
CN114596487A (en) Switch on-off state identification method based on self-attention mechanism
CN112966762A (en) Wild animal detection method and device, storage medium and electronic equipment
Li et al. Focus on local: transmission line defect detection via feature refinement
Zhao et al. Lightweight Neural Network for Component Recognition and Fault Detection of Large-Scale Transmission Line
Bhanbhro et al. Symbol Detection in a Multi-class Dataset Based on Single Line Diagrams using Deep Learning Models
CN117113010B (en) Power transmission channel safety monitoring method and system based on convolutional network lightweight
WO2024113541A1 (en) Battery cell electrode sheet inspection method and apparatus, and electronic device
CN116452868A (en) Open set identification method, device, equipment and storage medium based on isolated forest

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant