CN116363567B - Wafer defect identification method and system based on AOI visual inspection - Google Patents
Wafer defect identification method and system based on AOI visual inspection Download PDFInfo
- Publication number
- CN116363567B CN116363567B CN202310645819.2A CN202310645819A CN116363567B CN 116363567 B CN116363567 B CN 116363567B CN 202310645819 A CN202310645819 A CN 202310645819A CN 116363567 B CN116363567 B CN 116363567B
- Authority
- CN
- China
- Prior art keywords
- wafer
- preset
- wafer defect
- defect
- image
- 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
Links
- 230000007547 defect Effects 0.000 title claims abstract description 267
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000011179 visual inspection Methods 0.000 title claims description 25
- 238000005520 cutting process Methods 0.000 claims description 33
- 238000012549 training Methods 0.000 claims description 28
- 238000012216 screening Methods 0.000 claims description 22
- 238000012706 support-vector machine Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 7
- 230000002950 deficient Effects 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 30
- 230000000694 effects Effects 0.000 abstract description 8
- 230000000007 visual effect Effects 0.000 abstract description 5
- 238000012545 processing Methods 0.000 description 8
- 238000004519 manufacturing process Methods 0.000 description 7
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 239000003086 colorant Substances 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 150000002500 ions Chemical class 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000001179 sorption measurement Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a wafer defect identification method and a system based on AOI visual detection, and relates to the field of intelligent wafer defect identification, wherein the method comprises the following steps: obtaining preset wafer characteristic information by obtaining a preset tracking video stream of a preset printed circuit board and analyzing the preset tracking video stream; constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database; and analyzing the preset wafer characteristic information through the wafer defect intelligent identification model to obtain a preset wafer defect identification result. The wafer defect detection device solves the technical problems of low detection efficiency and poor detection accuracy in the wafer defect detection in the prior art, and is easy to misdetect, miss-detect and misdetect. The intelligent identification target of the wafer defects is realized, the detection efficiency of the wafer defects is improved, and meanwhile, the technical effect of the accuracy of wafer defect identification is effectively improved.
Description
Technical Field
The invention relates to the field of intelligent wafer defect identification, in particular to a wafer defect identification method and system based on AOI visual detection.
Background
With the rapid development of electronic manufacturing industry, the position of a printed circuit board (Printed Circuit Board, PCB) in industrial production is higher and higher, however, various pollution is inevitably introduced in the integrated circuit manufacturing process of manual participation, wafer defects are generated, and the performance parameters such as the yield of devices and the like are greatly influenced. Exemplary defects such as those caused by adsorption of atoms, ions, molecules, particles, or films on the wafer surface or in the oxide film of the wafer itself occur on the wafer surface. In addition, the existing method for detecting the defects of the wafer by the traditional manual detection method cannot meet the requirements on accuracy and rapidity of alignment detection, so that an accurate and rapid defect detection means is provided for detecting the defects of the wafer by intelligent automatic optical detection (Automatic Optic Inspection, AOI), and the defect detection method has the necessity and importance.
However, when detecting the wafer defects, the conventional manual detection is inevitably affected by the subjective factors and experience of the existing wafer, so that the accuracy and the efficiency of detecting and identifying the wafer defects are poor.
Disclosure of Invention
The invention aims to provide a wafer defect identification method and system based on AOI visual detection, which are used for solving the technical problems of low detection efficiency and poor detection accuracy in wafer defect detection in the prior art, and the problems of easy false detection, missed detection and defect false alarm.
In view of the above problems, the present invention provides a wafer defect identification method and system based on AOI visual inspection.
In a first aspect, the present invention provides a wafer defect identification method based on AOI visual inspection, where the method is implemented by a wafer defect identification system based on AOI visual inspection, and the method includes: obtaining preset wafer characteristic information by obtaining a preset tracking video stream of a preset printed circuit board and analyzing the preset tracking video stream; constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database; and analyzing the preset wafer characteristic information through the wafer defect intelligent identification model to obtain a preset wafer defect identification result.
In a second aspect, the present invention further provides a wafer defect recognition system based on AOI visual inspection, for performing a wafer defect recognition method based on AOI visual inspection according to the first aspect, where the system includes: the feature information obtaining module 11 is configured to obtain a preset tracking video stream of a preset printed circuit board, and analyze the preset tracking video stream to obtain preset wafer feature information; and an intelligent model construction module 12 for constructing a wafer defect intelligent recognition model, wherein the wafer defect intelligent recognition model is embedded with a wafer defect database; and the defect recognition module 13 is used for analyzing the preset wafer characteristic information through the wafer defect intelligent recognition model to obtain a preset wafer defect recognition result.
In a third aspect, the present invention also provides an electronic device, including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of the first aspects above.
In a fourth aspect, a computer readable storage medium having stored thereon a computer program which, when executed, implements the steps of the method of any of the first aspects above.
One or more technical schemes provided by the invention have at least the following technical effects or advantages:
1. obtaining preset wafer characteristic information by obtaining a preset tracking video stream of a preset printed circuit board and analyzing the preset tracking video stream; constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database; and analyzing the preset wafer characteristic information through the wafer defect intelligent identification model to obtain a preset wafer defect identification result. The intelligent identification target of the wafer defects is realized, the detection efficiency of the wafer defects is improved, and meanwhile, the technical effect of the accuracy of wafer defect identification is effectively improved.
2. The preset tracking video stream is obtained by tracking and monitoring the preset printed circuit board, so that the aim of automatically monitoring the state of the preset printed circuit board on the production line is fulfilled, and a comprehensive and effective video data basis is provided for automatically judging and identifying the wafer defects by a subsequent system.
3. The defect information of the wafer defect sample is analyzed, and a wafer defect database is constructed, so that a training and judging and identifying data basis is provided for automatic judging and identifying of the intelligent wafer defect identifying model. Then, by analyzing the data in the wafer defect database and training the wafer defect intelligent identification model, the aim of providing a model foundation for intelligently analyzing and identifying the wafer defects of the preset printed circuit board is fulfilled, and the effect of improving the intelligent degree of judging and identifying the wafer defects is achieved.
The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent. It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described below, it being obvious that the drawings in the description below are only exemplary and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a wafer defect identification method based on AOI visual inspection according to the present invention;
FIG. 2 is a schematic flow chart of a preset cutting result of a preset tracking video stream obtained in the wafer defect identification method based on AOI visual inspection;
FIG. 3 is a schematic flow chart of embedding a wafer defect database into a wafer defect intelligent recognition model in the wafer defect recognition method based on AOI visual inspection according to the present invention;
FIG. 4 is a schematic flow chart of using a defect judgment support vector machine as a first recognition layer in the wafer defect recognition method based on AOI visual inspection according to the present invention;
FIG. 5 is a schematic flow chart of a wafer defect recognition method based on AOI visual inspection using a defect recognition model as a second recognition layer according to the present invention;
fig. 6 is a schematic structural diagram of a wafer defect recognition system based on AOI visual inspection according to the present invention.
Detailed Description
The invention provides a wafer defect identification method and a wafer defect identification system based on AOI visual detection, which solve the technical problems of low detection efficiency and poor detection accuracy in wafer defect detection in the prior art, and cause easy false detection, missed detection and defect false alarm. The intelligent identification target of the wafer defects is realized, the detection efficiency of the wafer defects is improved, and meanwhile, the technical effect of the accuracy of wafer defect identification is effectively improved.
In the following, the technical solutions of the present invention will be clearly and completely described with reference to the accompanying drawings, and it should be understood that the described embodiments are only some embodiments of the present invention, but not all embodiments of the present invention, and that the present invention is not limited by the exemplary embodiments described herein. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention. It should be further noted that, for convenience of description, only some, but not all of the drawings related to the present invention are shown.
Examples
Referring to fig. 1, the invention provides a wafer defect identification method based on AOI visual inspection, wherein the method is applied to a wafer defect identification system based on AOI visual inspection, and the method specifically comprises the following steps:
step S100: acquiring a preset tracking video stream of a preset printed circuit board, and analyzing the preset tracking video stream to obtain preset wafer characteristic information; and
specifically, the wafer defect identification method based on AOI visual inspection is applied to a wafer defect identification system based on AOI visual inspection, and the wafer defect identification system is in communication connection with high-definition video acquisition equipment, such as a camera device and the like. The high-definition video acquisition equipment acquires real-time video of a preset printed circuit board on a production line, correspondingly obtains a video stream of the preset printed circuit board in a real-time motion state on the production line, namely, the preset tracking video stream is actually transmitted to the wafer defect recognition system in communication connection with the preset tracking video stream, and the wafer defect recognition system analyzes the preset tracking video stream in real time, correspondingly obtains real-time state characteristic information of the preset printed circuit board, namely, obtains the preset wafer characteristic information. Real-time characteristic information of the wafer is obtained by analyzing the real-time video stream, and an accurate and reliable wafer characteristic information basis is provided for carrying out wafer defect identification on a follow-up wafer defect intelligent identification model.
Further, as shown in fig. 2, step S100 of the present invention includes:
step S110: compressing the preset tracking video stream to obtain a compressed tracking video stream, wherein the compressed tracking video stream comprises M tracking images, and M is an integer greater than 1;
step S120: sequentially generating M image color histograms of the M tracking images;
step S130: sequentially acquiring M pieces of image main color information of the M pieces of image color histograms;
step S140: sequentially extracting first main color information and second main color information in the M image main color information, wherein the first main color information and the second main color information have adjacent relations;
step S150: calculating the information difference between the first main color information and the second main color information, and judging whether the information difference meets a preset information difference threshold value or not;
step S160: if yes, generating a cutting instruction, and sequentially reversely matching a first tracking image of the first main color information and a second tracking image of the second main color information according to the cutting instruction; and is also provided with
Step S170: and performing positioning cutting according to the first tracking image and the second tracking image to obtain a preset cutting result of the preset tracking video stream, wherein the preset cutting result comprises N sections of video streams, and N is an integer greater than 1.
Specifically, before video information analysis is performed on the preset tracking video stream acquired by the high-definition video acquisition device, compression processing is performed on the preset tracking video stream, and the compressed tracking video stream is correspondingly obtained. Exemplary compression of the preset trace video stream, such as by a moving picture expert group, reduces the system data throughput and thereby increases the system response speed and operating speed. Wherein, the compressed tracking video stream comprises M tracking images, namely each frame of image in the video is used as one tracking image, and M is an integer greater than 1, namely the compressed tracking video stream comprises more than one frame of image frame.
And then, carrying out image primary color tone analysis on each tracking image in the M tracking images, and correspondingly drawing to generate an image color histogram of each tracking image, thereby obtaining the M image color histograms. And then sequentially acquiring the main color information of each image color histogram in the M image color histograms, such as color hue, brightness and saturation information, to obtain M image main color information of the M image color histograms. Then, randomly extracting one image main color information from the M image main color information, setting the one image main color information as first main color information, further extracting adjacent image main color information positioned in front of or behind the randomly extracted one image main color information from the M image main color information, and correspondingly setting the adjacent image main color information as second main color information. That is, the first main color information and the second main color information have a neighboring positional relationship. And finally, calculating the information difference between the first main color information and the second main color information, and judging whether the information difference meets a preset information difference threshold value. When the information difference meets a preset information difference threshold, the system automatically generates a cutting instruction, and the cutting instruction is used for sequentially and reversely matching a first tracking image of the first main color information and a second tracking image of the second main color information according to the cutting instruction, and positioning cutting is performed according to the first tracking image and the second tracking image. For example, if the difference between the main color information of a certain tracking image frame a and the main color information of a certain tracking image frame B adjacent to the certain tracking image frame a is large, that is, the information difference is within the preset information difference threshold, it is indicated that a significant change occurs between the two image frames, and the two image frames are used as a cutting position, so that the monitored content in the video stream is simply divided by cutting. The preset information difference threshold value refers to the range of image color information which is set by combining with actual production conditions and is used for generating sudden change of shooting acquisition information after the related technicians analyze the colors of the surrounding environments of the wafer and the production line thereof. And finally, cutting to obtain a preset cutting result of the preset tracking video stream through all the positioned cutting positions, wherein the preset cutting result comprises N sections of video streams, and N is an integer greater than 1.
The target of providing a more accurate video information basis for identifying wafer defects for subsequent model analysis is achieved by compressing and cutting the preset tracking video stream, that is, the processing data volume of the system is reduced by compressing the video, the system efficiency is improved, the video is further segmented by clustering, a basis is provided for subsequent screening and determining the target video stream, and a more effective information basis is provided for subsequent defect analysis due to processing and screening of the video data, so that the identification efficiency of the system on the wafer defects is improved.
Further, the present invention further includes the following step S180:
step S181: comparing and analyzing the N segments of video streams and determining a target video stream, wherein the target video stream comprises P target images, and P is an integer greater than 1; and is also provided with
Step S182: obtaining a target image sequence according to the P target images;
step S183: extracting a first frame image in the target image sequence, and setting the first frame image as a first target image;
step S184: acquiring a non-first frame image in the target image sequence, and performing offset calibration on the non-first frame image by taking the first target image as a calibration reference to obtain a calibration result;
step S185: and analyzing the calibration result to obtain the preset wafer characteristic information.
Specifically, the N segments of video streams obtained after the segmentation processing are sequentially analyzed, and the target video streams are determined through comparison. The target video stream is a video clip of which the shot content is the preset printed circuit board, and the target video stream comprises P target images, wherein P is an integer larger than 1.
Firstly, the P target images are images with time marks, namely shooting time sequence characteristics, so that a target image sequence can be obtained. And then setting a first frame image in the target image sequence as a first target image, and taking the first target image as a calibration standard. And then taking the non-first frame images in the target image sequence except the first frame image as images to be calibrated in an offset mode, namely, taking the first frame image as a calibration standard to calibrate each non-first frame image, and correspondingly obtaining a calibration result. And finally, analyzing the calibration result to obtain the preset wafer characteristic information.
And a calibration result is obtained by carrying out offset calibration processing on each image in the target video stream, so that a complete and effective image foundation is provided for subsequent analysis of wafer characteristics.
Further, step S185 of the present invention includes:
step S1851: preprocessing the calibration result to obtain a calibration preprocessing result;
step S1852: acquiring a Harr feature set of the calibration pretreatment result, wherein the Harr feature set comprises Q index parameters of Q indexes, and Q is an integer greater than 1;
step S1853: acquiring a preset index set, and screening the Q indexes based on the preset index set to obtain an index screening result; and is also provided with
Step S1854: and combining the Q index parameters to obtain parameters of each index in the index screening result, setting the parameters as screening index parameters, and taking the screening index parameters as the preset wafer characteristic information.
Specifically, the calibration result is preprocessed before the calibration result is analyzed to obtain the preset wafer characteristic information. The calibration result is an image, and the image is subjected to processing such as enhancement, sharpening and the like to obtain a calibration pretreatment result, so that the complete and accurate extraction of the subsequent features is facilitated. Then, a Harr feature set of the calibration preprocessing result is obtained, wherein the Harr feature set comprises Q index parameters of Q indexes, and Q is an integer greater than 1. And then acquiring a preset index set, wherein the preset index set is used for screening the Q indexes based on the preset index set, and only index parameters with key influences on defect identification are reserved to obtain an index screening result, and the preset index set comprises a plurality of image characteristic indexes which are determined wafer image characteristic indexes which can have important influences on wafer defect identification after data mining and correlation analysis are carried out by related technicians. And finally, combining the Q index parameters to obtain parameters of each index in the index screening result, setting the parameters as screening index parameters, and taking the screening index parameters as the preset wafer characteristic information.
Step S200: constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database;
further, as shown in fig. 3, step S200 of the present invention includes:
step S210: constructing a wafer defect sample library, wherein the wafer defect sample library comprises R wafer defect samples, the R wafer defect samples correspond to R wafer defect types, and the R wafer defect types correspond to R wafer defect characteristic information;
step S220: constructing the wafer defect database according to the R wafer defect types, the R wafer defect characteristic information and the corresponding relation thereof; and is also provided with
Step S230: and embedding the wafer defect database into the wafer defect intelligent identification model.
Specifically, the intelligent wafer defect recognition model is used for intelligently analyzing the preset wafer characteristic information and then automatically judging and recognizing the wafer defects on the preset printed circuit board, and comprises a first recognition layer and a second recognition layer which are two recognition layers. In addition, the intelligent wafer defect identification model is embedded with a wafer defect database, and the wafer defect database is formed by constructing a wafer defect sample library containing R wafer defect samples, wherein each wafer defect sample in the wafer defect sample library corresponds to at least one wafer defect, that is, the R wafer defect samples correspond to R wafer defect types, and characteristic parameters of each wafer defect type are acquired in advance, so that the R wafer defect characteristic information corresponding to the R wafer defect types is obtained. And finally, constructing the wafer defect database according to the R wafer defect types, the R wafer defect characteristic information and the corresponding relation thereof, and embedding the wafer defect database into the wafer defect intelligent identification model.
Further, as shown in fig. 4, step S230 of the present invention includes:
step S231: extracting a first wafer defect type based on the wafer defect database, and matching first wafer defect characteristic information of the first wafer defect type;
step S232: carrying out defect identification on the first wafer defect characteristic information; and is also provided with
Step S233: obtaining a target wafer sample and extracting target wafer characteristic information of the target wafer sample, wherein the target wafer sample is a wafer sample with a defect-free mark;
step S234: taking the first wafer defect characteristic information, the defective mark, the target wafer characteristic information and the defect-free mark as a first training data set; and is also provided with
Step S235: and training according to the first training data set to obtain a defect judgment support vector machine, and taking the defect judgment support vector machine as a first identification layer.
Further, as shown in fig. 5, the present invention further includes the following steps:
step S236: taking the first wafer defect type and the first wafer defect characteristic information as a second training data set; and is also provided with
Step S237: training according to the second training data set to obtain a defect identification model, and taking the defect identification model as a second identification layer;
step S238: and the first identification layer and the second identification layer jointly form the wafer defect intelligent identification model.
Specifically, a first wafer defect type is obtained according to the wafer defect database, the first wafer defect type is any wafer defect condition in the wafer defect database, and then the wafer defect database is traversed to match first wafer defect characteristic information of the first wafer defect type. And then carrying out defect identification on the first wafer defect characteristic information.
And then acquiring a sample without wafer defects, namely a target wafer sample, based on big data or historical wafer detection conditions, and extracting target wafer characteristic information of the target wafer sample. Wherein the target wafer sample is provided with a defect-free mark. And then taking the first wafer defect characteristic information, the defective mark, the target wafer characteristic information and the defect-free mark as a first training data set, and training according to the first training data set to obtain a defect judgment support vector machine. And finally, taking the defect judgment support vector machine as a first identification layer. Further, the first wafer defect type and the first wafer defect characteristic information are used as a second training data set, and a defect identification model is obtained through training according to the second training data set. And finally, taking the defect identification model as a second identification layer. And finally, the first identification layer and the second identification layer jointly form the intelligent wafer defect identification model.
By constructing the first recognition layer of the intelligent wafer defect recognition model, the technical aim of intelligently judging whether the wafer on the preset printed circuit board has defects is achieved, and then constructing the second recognition layer of the intelligent wafer defect recognition model, the technical aim of further determining the types of wafer defects of the wafer on the preset printed circuit board is achieved, the accurate recognition of the wafer defects is achieved, references and bases are provided for the targeted repair processing of related personnel, and further the technical effects of pertinence and effectiveness of defect processing are improved.
Step S300: and analyzing the preset wafer characteristic information through the wafer defect intelligent identification model to obtain a preset wafer defect identification result.
Specifically, the first recognition layer of the intelligent wafer defect recognition model is used for carrying out preliminary analysis on the preset wafer characteristic information to obtain a recognition result of whether defects exist. When the wafer of the preset printed circuit board has defects, the second identification layer of the wafer defect intelligent identification model automatically identifies the defect types of the wafer, and the corresponding wafer defect types of the preset printed circuit board obtain the preset wafer defect identification result.
In summary, the wafer defect identification method based on AOI visual detection provided by the invention has the following technical effects:
obtaining preset wafer characteristic information by obtaining a preset tracking video stream of a preset printed circuit board and analyzing the preset tracking video stream; constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database; and analyzing the preset wafer characteristic information through the wafer defect intelligent identification model to obtain a preset wafer defect identification result. The intelligent identification target of the wafer defects is realized, the detection efficiency of the wafer defects is improved, and meanwhile, the technical effect of the accuracy of wafer defect identification is effectively improved.
Examples
Based on the same inventive concept as the wafer defect recognition method based on AOI visual inspection in the foregoing embodiment, the present invention further provides a wafer defect recognition system based on AOI visual inspection, referring to fig. 6, the system includes:
the characteristic information obtaining module is used for obtaining a preset tracking video stream of a preset printed circuit board and analyzing the preset tracking video stream to obtain preset wafer characteristic information; and
the intelligent model construction module is used for constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database;
and the defect identification module is used for analyzing the preset wafer characteristic information through the intelligent wafer defect identification model to obtain a preset wafer defect identification result.
Further, the feature information obtaining module in the system is further configured to:
compressing the preset tracking video stream to obtain a compressed tracking video stream, wherein the compressed tracking video stream comprises M tracking images, and M is an integer greater than 1;
sequentially generating M image color histograms of the M tracking images;
sequentially acquiring M pieces of image main color information of the M pieces of image color histograms;
sequentially extracting first main color information and second main color information in the M image main color information, wherein the first main color information and the second main color information have adjacent relations;
calculating the information difference between the first main color information and the second main color information, and judging whether the information difference meets a preset information difference threshold value or not;
if yes, generating a cutting instruction, and sequentially reversely matching a first tracking image of the first main color information and a second tracking image of the second main color information according to the cutting instruction; and is also provided with
And performing positioning cutting according to the first tracking image and the second tracking image to obtain a preset cutting result of the preset tracking video stream, wherein the preset cutting result comprises N sections of video streams, and N is an integer greater than 1.
Further, the feature information obtaining module in the system is further configured to:
comparing and analyzing the N segments of video streams and determining a target video stream, wherein the target video stream comprises P target images, and P is an integer greater than 1; and is also provided with
Obtaining a target image sequence according to the P target images;
extracting a first frame image in the target image sequence, and setting the first frame image as a first target image;
acquiring a non-first frame image in the target image sequence, and performing offset calibration on the non-first frame image by taking the first target image as a calibration reference to obtain a calibration result;
and analyzing the calibration result to obtain the preset wafer characteristic information.
Further, the feature information obtaining module in the system is further configured to:
preprocessing the calibration result to obtain a calibration preprocessing result;
acquiring a Harr feature set of the calibration pretreatment result, wherein the Harr feature set comprises Q index parameters of Q indexes, and Q is an integer greater than 1;
acquiring a preset index set, and screening the Q indexes based on the preset index set to obtain an index screening result; and is also provided with
And combining the Q index parameters to obtain parameters of each index in the index screening result, setting the parameters as screening index parameters, and taking the screening index parameters as the preset wafer characteristic information.
Further, the intelligent model building module in the system is further configured to:
constructing a wafer defect sample library, wherein the wafer defect sample library comprises R wafer defect samples, the R wafer defect samples correspond to R wafer defect types, and the R wafer defect types correspond to R wafer defect characteristic information;
constructing the wafer defect database according to the R wafer defect types, the R wafer defect characteristic information and the corresponding relation thereof; and is also provided with
And embedding the wafer defect database into the wafer defect intelligent identification model.
Further, the intelligent model building module in the system is further configured to:
extracting a first wafer defect type based on the wafer defect database, and matching first wafer defect characteristic information of the first wafer defect type;
carrying out defect identification on the first wafer defect characteristic information; and is also provided with
Obtaining a target wafer sample and extracting target wafer characteristic information of the target wafer sample, wherein the target wafer sample is a wafer sample with a defect-free mark;
taking the first wafer defect characteristic information, the defective mark, the target wafer characteristic information and the defect-free mark as a first training data set; and is also provided with
And training according to the first training data set to obtain a defect judgment support vector machine, and taking the defect judgment support vector machine as a first identification layer.
Further, the intelligent model building module in the system is further configured to:
taking the first wafer defect type and the first wafer defect characteristic information as a second training data set; and is also provided with
Training according to the second training data set to obtain a defect identification model, and taking the defect identification model as a second identification layer;
and the first identification layer and the second identification layer jointly form the wafer defect intelligent identification model.
In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the foregoing method and specific example for identifying a wafer defect based on AOI visual inspection in the first embodiment of fig. 1 are equally applicable to a wafer defect identifying system based on AOI visual inspection in the first embodiment, and by the foregoing detailed description of the foregoing method for identifying a wafer defect based on AOI visual inspection, those skilled in the art can clearly know that the foregoing system for identifying a wafer defect based on AOI visual inspection in the first embodiment, so that the description is omitted herein for brevity. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The present invention also provides an electronic device including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of embodiments.
The invention also provides a computer readable storage medium having stored thereon a computer program which when executed implements the steps of the method of any of the first embodiments.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the present invention and the equivalent techniques thereof, the present invention is also intended to include such modifications and variations.
Claims (8)
1. The wafer defect identification method based on AOI visual inspection is characterized by comprising the following steps of:
acquiring a preset tracking video stream of a preset printed circuit board, and analyzing the preset tracking video stream to obtain preset wafer characteristic information; and
constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database;
analyzing the preset wafer characteristic information through the intelligent wafer defect identification model to obtain a preset wafer defect identification result;
before the analyzing the preset tracking video stream to obtain the preset wafer characteristic information, the method comprises the following steps:
compressing the preset tracking video stream to obtain a compressed tracking video stream, wherein the compressed tracking video stream comprises M tracking images, and M is an integer greater than 1;
sequentially generating M image color histograms of the M tracking images;
sequentially acquiring M pieces of image main color information of the M pieces of image color histograms;
sequentially extracting first main color information and second main color information in the M image main color information, wherein the first main color information and the second main color information have adjacent relations;
calculating the information difference between the first main color information and the second main color information, and judging whether the information difference meets a preset information difference threshold value or not;
if yes, generating a cutting instruction, and sequentially reversely matching a first tracking image of the first main color information and a second tracking image of the second main color information according to the cutting instruction; and is also provided with
Positioning and cutting are carried out according to the first tracking image and the second tracking image, and a preset cutting result of the preset tracking video stream is obtained, wherein the preset cutting result comprises N sections of video streams, and N is an integer greater than 1;
after the preset cutting result of the preset tracking video stream is obtained, the method comprises the following steps:
comparing and analyzing the N segments of video streams and determining a target video stream, wherein the target video stream comprises P target images, and P is an integer greater than 1; and is also provided with
Obtaining a target image sequence according to the P target images;
extracting a first frame image in the target image sequence, and setting the first frame image as a first target image;
acquiring a non-first frame image in the target image sequence, and performing offset calibration on the non-first frame image by taking the first target image as a calibration reference to obtain a calibration result;
and analyzing the calibration result to obtain the preset wafer characteristic information.
2. The method of claim 1, wherein analyzing the calibration result to obtain the predetermined wafer characteristic information comprises:
preprocessing the calibration result to obtain a calibration preprocessing result;
acquiring a Harr feature set of the calibration pretreatment result, wherein the Harr feature set comprises Q index parameters of Q indexes, and Q is an integer greater than 1;
acquiring a preset index set, and screening the Q indexes based on the preset index set to obtain an index screening result; and is also provided with
And combining the Q index parameters to obtain parameters of each index in the index screening result, setting the parameters as screening index parameters, and taking the screening index parameters as the preset wafer characteristic information.
3. The method for identifying wafer defects according to claim 1, wherein the constructing the intelligent wafer defect identification model comprises:
constructing a wafer defect sample library, wherein the wafer defect sample library comprises R wafer defect samples, the R wafer defect samples correspond to R wafer defect types, and the R wafer defect types correspond to R wafer defect characteristic information;
constructing the wafer defect database according to the R wafer defect types, the R wafer defect characteristic information and the corresponding relation thereof; and is also provided with
And embedding the wafer defect database into the wafer defect intelligent identification model.
4. The method of claim 3, comprising, prior to said embedding said wafer defect database into said intelligent wafer defect identification model:
extracting a first wafer defect type based on the wafer defect database, and matching first wafer defect characteristic information of the first wafer defect type;
carrying out defect identification on the first wafer defect characteristic information; and is also provided with
Obtaining a target wafer sample and extracting target wafer characteristic information of the target wafer sample, wherein the target wafer sample is a wafer sample with a defect-free mark;
taking the first wafer defect characteristic information, the defective mark, the target wafer characteristic information and the defect-free mark as a first training data set; and is also provided with
And training according to the first training data set to obtain a defect judgment support vector machine, and taking the defect judgment support vector machine as a first identification layer.
5. The method according to claim 4, wherein after training according to the first training data set to obtain a defect judgment support vector machine and using the defect judgment support vector machine as a first recognition layer, comprising:
taking the first wafer defect type and the first wafer defect characteristic information as a second training data set; and is also provided with
Training according to the second training data set to obtain a defect identification model, and taking the defect identification model as a second identification layer;
and the first identification layer and the second identification layer jointly form the wafer defect intelligent identification model.
6. A wafer defect identification system based on AOI visual inspection, the wafer defect identification system comprising:
the characteristic information obtaining module is used for obtaining a preset tracking video stream of a preset printed circuit board and analyzing the preset tracking video stream to obtain preset wafer characteristic information; and
the intelligent model construction module is used for constructing an intelligent wafer defect identification model, wherein the intelligent wafer defect identification model is embedded with a wafer defect database;
the defect identification module is used for analyzing the preset wafer characteristic information through the intelligent wafer defect identification model to obtain a preset wafer defect identification result;
the feature information obtaining module comprises:
compressing the preset tracking video stream to obtain a compressed tracking video stream, wherein the compressed tracking video stream comprises M tracking images, and M is an integer greater than 1;
sequentially generating M image color histograms of the M tracking images;
sequentially acquiring M pieces of image main color information of the M pieces of image color histograms;
sequentially extracting first main color information and second main color information in the M image main color information, wherein the first main color information and the second main color information have adjacent relations;
calculating the information difference between the first main color information and the second main color information, and judging whether the information difference meets a preset information difference threshold value or not;
if yes, generating a cutting instruction, and sequentially reversely matching a first tracking image of the first main color information and a second tracking image of the second main color information according to the cutting instruction; and is also provided with
Positioning and cutting are carried out according to the first tracking image and the second tracking image, and a preset cutting result of the preset tracking video stream is obtained, wherein the preset cutting result comprises N sections of video streams, and N is an integer greater than 1;
comparing and analyzing the N segments of video streams and determining a target video stream, wherein the target video stream comprises P target images, and P is an integer greater than 1; and is also provided with
Obtaining a target image sequence according to the P target images;
extracting a first frame image in the target image sequence, and setting the first frame image as a first target image;
acquiring a non-first frame image in the target image sequence, and performing offset calibration on the non-first frame image by taking the first target image as a calibration reference to obtain a calibration result;
and analyzing the calibration result to obtain the preset wafer characteristic information.
7. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.
8. A computer readable storage medium, characterized in that it has stored thereon a computer program which, when executed, implements the steps of the method according to any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310645819.2A CN116363567B (en) | 2023-06-02 | 2023-06-02 | Wafer defect identification method and system based on AOI visual inspection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310645819.2A CN116363567B (en) | 2023-06-02 | 2023-06-02 | Wafer defect identification method and system based on AOI visual inspection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116363567A CN116363567A (en) | 2023-06-30 |
| CN116363567B true CN116363567B (en) | 2023-08-22 |
Family
ID=86929079
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310645819.2A Active CN116363567B (en) | 2023-06-02 | 2023-06-02 | Wafer defect identification method and system based on AOI visual inspection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116363567B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117250208B (en) * | 2023-11-20 | 2024-02-06 | 青岛天仁微纳科技有限责任公司 | Machine vision-based nano-imprint wafer defect accurate detection system and method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109977808A (en) * | 2019-03-11 | 2019-07-05 | 北京工业大学 | A kind of wafer surface defects mode detection and analysis method |
| CN113344886A (en) * | 2021-06-11 | 2021-09-03 | 长江存储科技有限责任公司 | Wafer surface defect detection method and equipment |
| WO2022147966A1 (en) * | 2021-01-08 | 2022-07-14 | 长鑫存储技术有限公司 | Wafer testing method and apparatus, and device and storage medium |
| CN115439660A (en) * | 2022-09-19 | 2022-12-06 | 抖音视界有限公司 | Detection method, detection device, electronic equipment and medium |
-
2023
- 2023-06-02 CN CN202310645819.2A patent/CN116363567B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109977808A (en) * | 2019-03-11 | 2019-07-05 | 北京工业大学 | A kind of wafer surface defects mode detection and analysis method |
| WO2022147966A1 (en) * | 2021-01-08 | 2022-07-14 | 长鑫存储技术有限公司 | Wafer testing method and apparatus, and device and storage medium |
| CN113344886A (en) * | 2021-06-11 | 2021-09-03 | 长江存储科技有限责任公司 | Wafer surface defect detection method and equipment |
| CN115439660A (en) * | 2022-09-19 | 2022-12-06 | 抖音视界有限公司 | Detection method, detection device, electronic equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116363567A (en) | 2023-06-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110349126B (en) | Convolutional neural network-based marked steel plate surface defect detection method | |
| WO2017181724A1 (en) | Inspection method and system for missing electronic component | |
| CN116363567B (en) | Wafer defect identification method and system based on AOI visual inspection | |
| CN113935666B (en) | Building decoration wall tile abnormity evaluation method based on image processing | |
| CN111275700A (en) | Terminal defect detection method and system based on deep learning | |
| US11580758B2 (en) | Method for processing image, electronic device, and storage medium | |
| CN115995056A (en) | Automatic bridge disease identification method based on deep learning | |
| CN114266743A (en) | FPC defect detection method, system and storage medium based on HSV and CNN | |
| CN115841488A (en) | Hole checking method of PCB (printed Circuit Board) based on computer vision | |
| CN114943875A (en) | Visual analysis method for cable element identification | |
| CN112967224A (en) | Electronic circuit board detection system, method and medium based on artificial intelligence | |
| CN116958052A (en) | Printed circuit board defect detection method based on YOLO and attention mechanism | |
| CN116843677A (en) | Appearance quality detection system and method for sheet metal part | |
| CN111724375A (en) | Screen detection method and system | |
| TW202402117A (en) | Training method for defect detection model and defect detection method for printed circuit board | |
| CN115797314A (en) | Part surface defect detection method, system, equipment and storage medium | |
| CN115082650A (en) | Implementation method of automatic pipeline defect labeling tool based on convolutional neural network | |
| CN114677670A (en) | Automatic identification and positioning method for identity card tampering | |
| CN112686105A (en) | Fog concentration grade identification method based on video image multi-feature fusion | |
| CN111369508A (en) | Defect detection method and system for metal three-dimensional lattice structure | |
| CN113344858B (en) | Feature detection method, device and computer storage medium | |
| KR102129970B1 (en) | Method And Apparatus for Matching inspection Data Electronic Component | |
| CN117853826B (en) | Object surface precision identification method based on machine vision and related equipment | |
| CN117978966A (en) | Method for monitoring autonomous running state of kitchen video through multi-point access | |
| CN118014933A (en) | Defect detection and identification method and device based on image detection |
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 |