CN106153639A - circuit board detecting method based on artificial intelligence and detection device thereof - Google Patents
circuit board detecting method based on artificial intelligence and detection device thereof Download PDFInfo
- Publication number
- CN106153639A CN106153639A CN201510188433.9A CN201510188433A CN106153639A CN 106153639 A CN106153639 A CN 106153639A CN 201510188433 A CN201510188433 A CN 201510188433A CN 106153639 A CN106153639 A CN 106153639A
- Authority
- CN
- China
- Prior art keywords
- product
- data
- image
- circuit board
- processing mechanism
- 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.)
- Granted
Links
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a kind of circuit board detecting method based on artificial intelligence, first first pass through photographic unit and carry out detection crawl image, then by image from display screen display, processing mechanism is to view data Fitting Analysis, judge whether the characteristic information meeting in data base, do respective markers, and marking signal transmission is shown to display screen；When mark product is carried out the confirmation post-processor structure of excellent product, data message is fitted whether comparison meets again, and carries out data class definition storage, finally carry out classifying and sorted examine.Beneficial effects of the present invention: quickly can be marked and show in place against regulation to circuit board so that operator can find and carry out corresponding defect processing in time, meanwhile, can detect polytype circuit board, be greatly improved detection efficiency.
Description
Technical field
The present invention relates to a kind of circuit board detecting method based on artificial intelligence and detection device thereof.
Background technology
Flexible PCB (FPC) is to have height reliability, excellent flexible printed circuit with polyimides or mylar for the one that base material is made.Have the advantages that distribution density is high, lightweight, thickness is thin, bending property is good.In process of production, in order to prevent out short circuit too much to cause yield too low or reduce hole, roll, FPC plate that cutting etc. goes out technological problems and causes is scrapped, the problem of feed supplement, so needing before putting into application, product carries out detection analyze, existing method is typically all by manually detecting, but due to different operator, empirical value is different, detection process wastes time and energy, and causes detection efficiency low.
Summary of the invention
It is an object of the invention to solve above-mentioned technical problem, it is provided that a kind of circuit board detecting method based on artificial intelligence and detection device thereof.
The purpose of the present invention is achieved through the following technical solutions:
Circuit board detecting method based on artificial intelligence, comprises the steps,
S1, detection steps, by being arranged at whether the photographic unit above table top detects with the presence of product on table top；When there is product, it is judged that its most static being placed in detects regional center, the most then carry out shooting and take phase, and send data to processing mechanism, otherwise, do not shoot；
Described photographic unit includes two groups of cameras, often organize that described camera is respectively disposed adjacent for shooting the overall IP Camera group of product and for shooting the area array cameras group of product topography further, described area array cameras group shooting figure carries out the labelling of position in overall figure；Often organize the camera lens of camera at least provided with one.
S2, image procossing and step display, with photographic unit carry out the processing mechanism that is electrically connected with image is fitted or swing become a full member to process, and the view data transmission after processing is to display screen, the product image that display photographs；
S3, analytical procedure；Setting up data base, processing mechanism analyses whether the characteristic information meeting in data base to view data detection, it is judged that does not meets, is marked, and marking signal transmission is shown to display screen, if meeting, then carries out filtering not labelling, and carries out the preservation of image；Described data base is established as self-defined setting or intelligent extraction, and described intelligent extraction comprises the steps；
S31, the number of times lowest limit arranging collecting sample and the range of error allowed, proceed by the collection of sample；
S32, collected by camera sample product, be acquired carrying out the extraction of characteristic to first sample, and identify shape of product corresponding to this feature data and color；
S33, proceed after the collection of a sample, and the feature of a rear sample is mated with the feature of front sample, calculating meansigma methods；
S34, repetition S33 step, collection and meansigma methods to data carry out matching primitives, and judge that the change of this meansigma methods, whether in required value range, if exceeding required value range, then rejects corresponding data；If meeting, the most continual it is added until completing the lowest limit number arranged, and carries out the preservation of data, form data base during detection；
S4, confirms step, and mark product is made whether the confirmation of erasability defect；
S5, analytical procedure again, image mechanism gathers the product information after confirming again, and data message is again carried out detecting whether to meet by processing mechanism with the characteristic information in data base, if judging not meet, then labelling does not eliminates, and change color, if meeting, then labelling eliminates, data after reaffirming carry out class definition storage, and carry out the preservation of respective image；
S6, confirm step again, the product after analyzing again being reaffirmed the erasability of its defect, if confirming to eliminate, then carrying out classifying step, if can eliminate, then carry out the analytical procedure again of S5 after eliminating；
S7, classifying step, the product after detection is carried out the sorting placement of last excellent product, is respectively placed in different boxes；
S8, step is examined in classification, it is arranged at the photographic head above box and box is carried out location positioning, and the product in box is shot, and send data to processing mechanism, it is analyzed comparison with the class definition data of preservation in S5, examine and whether place mistake, if not meeting, then mistake, start the alarm mechanism being electrically connected with processing mechanism.
Preferably, described S1 specifically includes following steps,
S11, divides operating table surface, and coverage is defined as monitored area, monitored area is made the boundary definition of scope, and confirms center in bounds；
S12, when the external boundary monitoring monitored area has covering, then it represents that have product to enter；
S13, before and after the image in contrastive detection region, frame is the most consistent again, when before and after image, frame is consistent, then it represents that product is static；The center in bounds that simultaneously monitors is capped, and border is unobstructed, then it represents that product is positioned at shooting area center, and image mechanism is converted to high-resolution by low resolution, starts shooting, otherwise does not starts.
Preferably, in described S2 image display step, image is the display of product required detection position, comprises the steps:
The view data that photographic unit is photographed by S21, processing mechanism carries out center of gravity or the principal axis of inertia processes to image forward；
S22, use LUT search and then find color space territory；
S23, spatial domain is marked, and uses morphologic method to filter；
S24, the region after label is further carried out the contrast of area or shape, remove incongruent point in label, obtain the image of required detection position.
Preferably, one during approximating method includes feature based matching or Gray Correlation matching in described S2, and combinations thereof matching.
Preferably, described feature based approximating method comprises the steps:
Step one, in data base, the shape data parameter of formal parameter and characteristic point, color data parameter are as normal data；
Step 2, formal parameter to detection product extract and with normal data comparison, to product characteristic point shape, color is maximized with normal data overlaps.
Preferably, in described S3 and S5 graphical analysis include Threshold segmentation, characteristic area compare, color displacement and shape compare in one or more combinations.
Preferably, the parameter that in described S3 and S5, the feature in data base relates to includes the shape data parameter of product, color data parameter.
Preferably, described S8 specifically includes following steps:
S81, extraction to box boundary characteristic, and predefined box data message compares with in data base by extraction feature；If meeting, then it represents that there is box；
S82, dividing carrying out subregion and color in box border, and carry out the determination of boundary, different colours represents the classification that whether qualified corresponding product is；
S83, detect that boundary is blocked when photographic head, then it represents that have product to enter, and carry out placing the shooting of product, and contrasted with the defining classification data of storage in S5 by image data, detect whether identical.
A kind of circuit board detection device based on artificial intelligence, described detection device includes,
Operating table surface, is used for placing product to be detected, and described operating table surface divides and is provided with operating area；
Photographic unit, described image mechanism is placed on above described operating area by propping up, described photographic unit includes the adjacent two groups of cameras being fixedly installed, being respectively for shooting the overall IP Camera group of product and for shooting the area array cameras group of product topography further, the shooting image of described area array cameras group is labeled on general image；
Processing mechanism, described processing mechanism is electrically connected with described photographic unit；
Display screen, described display screen is electrically connected with described processing mechanism；
Box, described box is placed in the side of operating table surface, and described box is at least provided with qualified district and defective district；Certainly, the region of box divides and can divide as required.
Photographic head, described photographic head is placed in the top of box, and described photographic head is electrically connected with described processing mechanism.
Preferably, described device also includes sound-controlled apparatus or the body-sensing assembly being electrically connected with processing mechanism.
Beneficial effects of the present invention: quickly can be marked and show in place against regulation to circuit board so that operator can find and carry out corresponding defect processing in time, meanwhile, can detect polytype circuit board, be greatly improved detection efficiency.
Detailed description of the invention
Present invention is disclosed a kind of circuit board detection device based on artificial intelligence, described detection device includes:
Operating table surface, is used for placing product to be detected, and described operating table surface divides and is provided with operating area；
Photographic unit, described image mechanism is placed on above described operating area by propping up；Described photographic unit is the two groups of cameras being disposed adjacent, one group of shooting for overall figure, and another group is for the shooting of local location, and the local location photographed can carry out the labelling of position in overall figure, preferably to make mark for detection position.Certainly, often organize the concrete number of shots in camera and can carry out the setting of reality as required.
Processing mechanism, described processing mechanism is electrically connected with described photographic unit；
Display screen, described display screen is electrically connected with described processing mechanism；
Box, described box is placed in the side of operating table surface, and described box is respectively arranged with qualified district and defective district；Different colors can be arranged between different regions be distinguish between.
Photographic head, described photographic head is placed in the top of box, and described photographic head is electrically connected with described processing mechanism.
In order to preferably carry out trace-back operation before and after product, described device also includes sound-controlled apparatus or the body-sensing assembly being electrically connected with processing mechanism.By the input of acoustic control and body-sensing more intelligent with operate faster.
The circuit board detecting method based on artificial intelligence realized by this device under set forth below, is comprised the steps,
S1, detection steps, by being arranged at whether the photographic unit above table top detects with the presence of product on table top；When there is product, it is judged that its most static being placed in detects regional center, the most then carry out shooting and take phase, and send data to processing mechanism, otherwise, do not shoot.
Described photographic unit includes two groups of cameras, often organize that described camera is respectively disposed adjacent for shooting the overall IP Camera group of product and for shooting the area array cameras group of product topography further, described area array cameras group shooting figure carries out the labelling of position in overall figure；Often organize the number of camera at least provided with one or more, can be configured according to actual needs.
When mobile product, the local photographed can followed by mobile on overall figure and move, thus the personnel that are easier to operate to can learn the position at the part place of detection now.
Concrete, described S1 comprises the steps,
S11, divides operating table surface, and coverage is defined as monitored area, monitored area is made the boundary definition of scope, and confirms center in bounds；
S12, when the external boundary monitoring monitored area has covering, then it represents that have product to enter；
S13, before and after the image in contrastive detection region, frame is the most consistent again, when before and after image, frame is consistent, then it represents that product is static；The center in bounds that simultaneously monitors is capped, and border is unobstructed, then it represents that product is positioned at shooting area center, and image mechanism is converted to high-resolution by low resolution, starts shooting, otherwise does not starts.Low resolution and high-resolution conversion are also one and are changed to identification locally by entirety, and when detecting product area more than detection region, the local first carrying out detection product with high-resolution is shot and takes phase by photographic unit.
S2, image procossing and step display, with photographic unit carry out the processing mechanism that is electrically connected with image is fitted or swing become a full member to process, and the view data transmission after processing is to display screen, the product image that display photographs；Described matching can be characterized the matching of matching a little or Gray Correlation.
In images above step display, the display of image can also be only displayed as product required detection position, comprises the steps:
The view data that photographic unit is photographed by S21, processing mechanism rotates to forward；
S22, use LUT search and then find color space territory；
S23, spatial domain is marked, and uses morphologic method to filter；
S24, the region after label is further carried out the contrast of area or shape, remove incongruent point in label, obtain the image of required detection position.
S3, analytical procedure, whether processing mechanism meets the characteristic information in data base to view data Fitting Analysis, it is judged that does not meets, is marked, and marking signal transmission is shown to display screen.If meeting, then carry out filtering not labelling, and carry out the preservation of image；Described data base is the custom data standard carrying out in advance storing；Described data matching includes global feature matching and local feature matching；Described data standard includes the shape data parameter of product, color data parameter.In concrete application, described supplemental characteristic can embody the defective features such as the scuffing of circuit board, oxidation accordingly.The parameter that feature in described data base relates to includes the shape data parameter of product, color data parameter.
As long as the characteristic desired parameters of the collection of product in analytical procedure is different from labeled data or not in the admissible margin of tolerance, will carry out the labelling of color.
Overall fit and local fit also can more improve the Detection accuracy of product.
Owing to actual product is under the influence of preparation and external environment, inevitable gap can be produced with data sample, now, when the spacing of self-defining characteristic point is not inconsistent with the spacing of the characteristic point being extracted detection, then can carry out the positional information matching comparison of local further, last processing mechanism judges whether to make corresponding labelling again.In the present invention, fit procedure effectively achieves the type automatically detecting product.
Certainly, another important distinguishing characteristics of the present invention and prior art is, the data base in the present invention can be obtained by computer disposal mechanism autonomic learning, and concrete step includes:
The foundation of described data base comprises the steps:
First, the number of times lowest limit that collecting sample is set and the range of error allowed, proceed by the collection of sample；
Then, collected by camera sample product, first sample is acquired carrying out the extraction of characteristic, and identifies shape of product corresponding to this feature data and color；
Then, proceed after the collection of a sample, and the feature of a rear sample is mated with the feature of front sample, calculating meansigma methods；
Repeating previous step, collection and meansigma methods to data carry out matching primitives, and judge that the change of this meansigma methods, whether in required value range, if exceeding required value range, then rejects corresponding data；If meeting, the most continual it is added until completing the lowest limit number arranged, and carries out the preservation of data, form data base during detection.
Such as: setting and need the minimum standards of sampling as 100 times, the scope of defect is within 3%, and camera first gathers first sample, and transmission is to processing mechanism, and processing mechanism carries out the feature analysis of data, confirms the shape of this sample, color.Using its data as first group, carry out the data acquisition of second sample afterwards, and by the data of sample and first group of mean value calculation carrying out data, carry out the preservation of data；Proceed the three, the 4th .... the data acquisition of equal samples, and the calculating of the value that is constantly averaged, described meansigma methods compares with each cell mean data before, if scope of data difference is more than 3%, then carries out the rejecting of data；If in the range of, then carry out the preservation of data；Minimum collecting sample standard is 100 times, when the sample of the 100th time and when not meeting data, then needs to continue to add sample, to meet minimum standards, finally meets after requiring, carries out data preserving and forms follow-up database standard.
S4, confirms step, and the mark part of product carries out the confirmation of excellent product, and the operability carrying out being correlated with processes, if confirming as eliminating defect, then eliminates.
S5, analytical procedure again, image mechanism gathers the product information after confirming again, and data message and the characteristic information in data base are fitted whether comparison meets by processing mechanism again, if judging not meet, then labelling does not eliminates, and change color, if meeting, then labelling eliminates, data after reaffirming carry out class definition storage, and carry out the preservation of image.Whether in order to distinguish it has been acknowledged that mistake, the color of labelling can be converted by processing mechanism accordingly.
During analyzing at this, the data message being finally identified to can be carried out whether the most qualified judgement and classification by processing mechanism, and data is stored and memory module, carries out data basis for follow-up again examining.
S6, confirm step again, the product after analyzing again being reaffirmed the erasability of its defect, if confirming to eliminate, then carrying out classifying step, if can eliminate, then carry out the analytical procedure again of S5 after eliminating；
S7, classifying step, carry out the sorting placement of excellent product, be respectively placed in different boxes the product after detection；
S8, step is examined in classification, it is arranged at the photographic head above box and box is carried out location positioning, and the product in box is shot, and send data to processing module, it is analyzed comparison with the class definition data of preservation in S5, examines and whether place mistake, if not meeting, mistake, start and report to the police.Corresponding detection is done in the behavior of effectively can classifying operator of this step, prevents the classification of mistake.
Concrete, described S7 comprises the steps:
S71, extraction to box boundary characteristic, and predefined box data message compares with in data base by extraction feature；If meeting, then it represents that there is box；The data parameters of box generally comprises the setting of right-angle side, length and width and ratio thereof etc..Actual shape is by being specifically defined.So being also convenient for the placement operation step of operator, the operating habit making box can follow people carries out corresponding position adjustment.In addition, can define the color of different boxes accordingly, as green represents qualified frame, red expression does not conforms to gridiron yet.
S72, to carrying out region division, and the determination of boundary in box border；
S72, detect that boundary is blocked when photographic head, then it represents that have product to enter, and carry out placing the shooting of product, and contrasted with the defining classification data of storage in S5 by image data, detect whether identical.
The detection data of the present invention and picture etc. all will be stored in the data base on backstage, and when needs extract, data base therein can be carried out extraction checks, can have the tracing function of data and picture.By LAN or electrical connection, can very easily data be checked.
The approximating method being applied in the present invention include the one in distinguished point based matching or Gray Correlation matching, and combinations thereof matching.
Described feature based approximating method comprises the steps:
Step one, self-defined product design parameter and the shape data parameter of characteristic point, color data parameter are as normal data；
Step 2, formal parameter to detection product extract and with normal data comparison, to product characteristic point shape, color is maximized with normal data overlaps.
The present invention still has multiple specific embodiment, all employing equivalents or equivalent transformation and all technical schemes of being formed, within all falling within the scope of protection of present invention.
Claims (10)
1. circuit board detecting method based on artificial intelligence, it is characterised in that: comprise the steps,
S1, detection steps, by being arranged at whether the photographic unit above table top detects with the presence of product on table top；When there is product, it is judged that its most static being placed in detects regional center, the most then carry out shooting and take phase, and send data to processing mechanism, otherwise, do not shoot；
Described photographic unit includes two groups of cameras, often organize that described camera is respectively disposed adjacent for shooting the overall IP Camera group of product and for shooting the area array cameras group of product topography further, described area array cameras group shooting figure carries out the labelling of position in overall figure；
S2, image procossing and step display, with photographic unit carry out the processing mechanism that is electrically connected with image is fitted or swing become a full member to process, and the view data transmission after processing is to display screen, the product image that display photographs；
S3, analytical procedure；Setting up data base, processing mechanism analyses whether the characteristic information meeting in data base to view data detection, it is judged that does not meets, is marked, and marking signal transmission is shown to display screen, if meeting, then carries out filtering not labelling, and carries out the preservation of image；Described data base is established as self-defined setting or intelligent extraction, and described intelligent extraction comprises the steps；
S31, the number of times lowest limit arranging collecting sample and the range of error allowed, proceed by the collection of sample；
S32, collected by camera sample product, be acquired carrying out the extraction of characteristic to first sample, and identify shape of product corresponding to this feature data and color；
S33, proceed after the collection of a sample, and the feature of a rear sample is mated with the feature of front sample, calculating meansigma methods；
S34, repetition S33 step, collection and meansigma methods to data carry out matching primitives, and judge that the change of this meansigma methods, whether in required value range, if exceeding required value range, then rejects corresponding data；If meeting, the most continual it is added until completing the lowest limit number arranged, and carries out the preservation of data, form data base during detection；
S4, confirms step, and mark product is made whether the confirmation of erasability defect；
S5, analytical procedure again, image mechanism gathers the product information after confirming again, and data message is again carried out detecting whether to meet by processing mechanism with the characteristic information in data base, if judging not meet, then labelling does not eliminates, and change color, if meeting, then labelling eliminates, data after reaffirming carry out class definition storage, and carry out the preservation of respective image；
S6, confirm step again, the product after analyzing again being reaffirmed the erasability of its defect, if confirming to eliminate, then carrying out classifying step, if can eliminate, then carry out the analytical procedure again of S5 after eliminating；
S7, classifying step, the product after detection is carried out the sorting placement of last excellent product, is respectively placed in different boxes；
S8, step is examined in classification, it is arranged at the photographic head above box and box is carried out location positioning, and the product in box is shot, and send data to processing mechanism, it is analyzed comparison with the class definition data of preservation in S5, examine and whether place mistake, if not meeting, then mistake, start the alarm mechanism being electrically connected with processing mechanism.
Circuit board detecting method based on artificial intelligence the most according to claim 1, it is characterised in that: described S1 specifically includes following steps,
S11, divides operating table surface, and coverage is defined as monitored area, monitored area is made the boundary definition of scope, and confirms center in bounds；
S12, when the external boundary monitoring monitored area has covering, then it represents that have product to enter；
S13, before and after the image in contrastive detection region, frame is the most consistent again, when before and after image, frame is consistent, then it represents that product is static；The center in bounds that simultaneously monitors is capped, and border is unobstructed, then it represents that product is positioned at shooting area center, and image mechanism is converted to high-resolution by low resolution, starts shooting, otherwise does not starts.
Circuit board detecting method based on artificial intelligence the most according to claim 1, it is characterised in that: in described S2 image display step, image is the display of product required detection position, comprises the steps:
The view data that photographic unit is photographed by S21, processing mechanism carries out center of gravity or the principal axis of inertia processes to image forward；
S22, use LUT search and then find color space territory；
S23, spatial domain is marked, and uses morphologic method to filter；
S24, the region after label is further carried out the contrast of area or shape, remove incongruent point in label, obtain the image of required detection position.
Circuit board detecting method based on artificial intelligence the most according to claim 1, it is characterised in that: one during approximating method includes feature based matching or Gray Correlation matching in described S2, and combinations thereof matching.
Circuit board detecting method based on artificial intelligence the most according to claim 4, it is characterised in that: described feature based approximating method comprises the steps:
Step one, in data base, the shape data parameter of formal parameter and characteristic point, color data parameter are as normal data；
Step 2, formal parameter to detection product extract and with normal data comparison, to product characteristic point shape, color is maximized with normal data overlaps.
Circuit board detecting method based on artificial intelligence the most according to claim 1, it is characterised in that: in described S3 and S5 graphical analysis include Threshold segmentation, characteristic area compare, color displacement and shape compare in one or more combinations.
Circuit board detecting method based on artificial intelligence the most according to claim 1, it is characterised in that: the parameter that in described S3 and S5, the feature in data base relates to includes the shape data parameter of product, color data parameter.
Circuit board detecting method based on artificial intelligence the most according to claim 1, it is characterised in that: described S8 specifically includes following steps:
S81, extraction to box boundary characteristic, and predefined box data message compares with in data base by extraction feature；If meeting, then it represents that there is box；
S82, dividing carrying out subregion and color in box border, and carry out the determination of boundary, different colours represents the classification that whether qualified corresponding product is；
S83, detect that boundary is blocked when photographic head, then it represents that have product to enter, and carry out placing the shooting of product, and contrasted with the defining classification data of storage in S5 by image data, detect whether identical.
9. a circuit board detection device based on artificial intelligence, it is characterised in that: described detection device includes,
Operating table surface, is used for placing product to be detected, and described operating table surface divides and is provided with operating area；
Photographic unit, described image mechanism is placed on above described operating area by propping up, described photographic unit includes the adjacent two groups of cameras being fixedly installed, being respectively for shooting the overall IP Camera group of product and for shooting the area array cameras group of product topography further, the shooting image of described area array cameras group is labeled on general image；
Processing mechanism, described processing mechanism is electrically connected with described photographic unit；
Display screen, described display screen is electrically connected with described processing mechanism；
Box, described box is placed in the side of operating table surface, and described box is at least provided with qualified district and defective district；
Photographic head, described photographic head is placed in the top of box, and described photographic head is electrically connected with described processing mechanism.
10. circuit board detection device based on artificial intelligence as claimed in claim 9, it is characterised in that: described device also includes sound-controlled apparatus or the body-sensing assembly being electrically connected with processing mechanism.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510188433.9A CN106153639B (en) | 2015-04-21 | 2015-04-21 | Circuit board detecting method based on artificial intelligence and its detection device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510188433.9A CN106153639B (en) | 2015-04-21 | 2015-04-21 | Circuit board detecting method based on artificial intelligence and its detection device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106153639A true CN106153639A (en) | 2016-11-23 |
| CN106153639B CN106153639B (en) | 2018-10-19 |
Family
ID=58058769
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510188433.9A Active CN106153639B (en) | 2015-04-21 | 2015-04-21 | Circuit board detecting method based on artificial intelligence and its detection device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106153639B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106483145A (en) * | 2016-11-29 | 2017-03-08 | 深圳宜美智科技有限公司 | Pcb board automatic detection device and automatic testing method |
| CN106776877A (en) * | 2016-11-28 | 2017-05-31 | 常州市星网计算机技术有限公司 | Motorcycle parts model automatic identifying method |
| CN106841217A (en) * | 2017-03-01 | 2017-06-13 | 苏州光照精密仪器有限公司 | A kind of flaw detection method, device and system |
| CN107275245A (en) * | 2017-05-22 | 2017-10-20 | 茆胜 | A kind of OLED minitype displayer Rapid checking device and its method for quickly detecting |
| WO2018187989A1 (en) * | 2017-04-12 | 2018-10-18 | 意力(广州)电子科技有限公司 | Detection method and apparatus for use with display screen |
| CN109239076A (en) * | 2018-08-29 | 2019-01-18 | 西安理工大学 | A kind of sewing thread trace defect inspection method based on machine vision |
| CN109916601A (en) * | 2019-04-11 | 2019-06-21 | 北极光电(深圳)有限公司 | The system and detection method whether Thin Film Filter finds fault with people are checked based on machine vision |
| CN110490955A (en) * | 2019-08-26 | 2019-11-22 | 成都步科智能有限公司 | A kind of mark image acquiring method and device |
| CN114018950A (en) * | 2021-09-29 | 2022-02-08 | 浪潮(山东)计算机科技有限公司 | Circuit board divides board detecting system |
| US11935221B2 (en) | 2019-11-27 | 2024-03-19 | AT&S (Chongqing) Company Limited | User interface for judgment concerning quality classification of displayed arrays of component carriers |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH04357444A (en) * | 1991-06-03 | 1992-12-10 | Sharp Corp | Apparatus for inspecting printed circuit board |
| CN1280691A (en) * | 1997-10-09 | 2001-01-17 | 约瑟夫·L·比莱利亚 | Electronic assembly image inspection system |
| CN201079781Y (en) * | 2007-07-02 | 2008-07-02 | 比亚迪股份有限公司 | On-line accessory appearance detecting sorting machine |
| CN102186307A (en) * | 2010-01-11 | 2011-09-14 | 卡姆特有限公司 | Optical inspection during PCB production process |
| CN102680495A (en) * | 2011-03-15 | 2012-09-19 | 上海赫立电子科技有限公司 | Device and method for automatic optical detection |
| CN203758922U (en) * | 2014-02-18 | 2014-08-06 | 王相臣 | Image information extraction inspecting instrument for printed documents |
-
2015
- 2015-04-21 CN CN201510188433.9A patent/CN106153639B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH04357444A (en) * | 1991-06-03 | 1992-12-10 | Sharp Corp | Apparatus for inspecting printed circuit board |
| CN1280691A (en) * | 1997-10-09 | 2001-01-17 | 约瑟夫·L·比莱利亚 | Electronic assembly image inspection system |
| CN201079781Y (en) * | 2007-07-02 | 2008-07-02 | 比亚迪股份有限公司 | On-line accessory appearance detecting sorting machine |
| CN102186307A (en) * | 2010-01-11 | 2011-09-14 | 卡姆特有限公司 | Optical inspection during PCB production process |
| CN102680495A (en) * | 2011-03-15 | 2012-09-19 | 上海赫立电子科技有限公司 | Device and method for automatic optical detection |
| CN203758922U (en) * | 2014-02-18 | 2014-08-06 | 王相臣 | Image information extraction inspecting instrument for printed documents |
Non-Patent Citations (1)
| Title |
|---|
| 王盛婷: "基于AOI的FPC缺陷检测系统的研究与设计", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106776877A (en) * | 2016-11-28 | 2017-05-31 | 常州市星网计算机技术有限公司 | Motorcycle parts model automatic identifying method |
| CN106776877B (en) * | 2016-11-28 | 2020-05-01 | 常州市星网计算机技术有限公司 | Automatic identification method for motorcycle part models |
| CN106483145A (en) * | 2016-11-29 | 2017-03-08 | 深圳宜美智科技有限公司 | Pcb board automatic detection device and automatic testing method |
| CN106483145B (en) * | 2016-11-29 | 2017-12-26 | 深圳宜美智科技有限公司 | Pcb board automatic detection device and automatic testing method |
| CN106841217A (en) * | 2017-03-01 | 2017-06-13 | 苏州光照精密仪器有限公司 | A kind of flaw detection method, device and system |
| WO2018187989A1 (en) * | 2017-04-12 | 2018-10-18 | 意力(广州)电子科技有限公司 | Detection method and apparatus for use with display screen |
| CN107275245A (en) * | 2017-05-22 | 2017-10-20 | 茆胜 | A kind of OLED minitype displayer Rapid checking device and its method for quickly detecting |
| CN109239076A (en) * | 2018-08-29 | 2019-01-18 | 西安理工大学 | A kind of sewing thread trace defect inspection method based on machine vision |
| CN109916601A (en) * | 2019-04-11 | 2019-06-21 | 北极光电(深圳)有限公司 | The system and detection method whether Thin Film Filter finds fault with people are checked based on machine vision |
| CN110490955A (en) * | 2019-08-26 | 2019-11-22 | 成都步科智能有限公司 | A kind of mark image acquiring method and device |
| US11935221B2 (en) | 2019-11-27 | 2024-03-19 | AT&S (Chongqing) Company Limited | User interface for judgment concerning quality classification of displayed arrays of component carriers |
| CN114018950A (en) * | 2021-09-29 | 2022-02-08 | 浪潮(山东)计算机科技有限公司 | Circuit board divides board detecting system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106153639B (en) | 2018-10-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106153639A (en) | circuit board detecting method based on artificial intelligence and detection device thereof | |
| CN112053318B (en) | Two-dimensional PCB defect real-time automatic detection and classification device based on deep learning | |
| CN111626139B (en) | Accurate detection method for fault information of IT equipment in machine room | |
| CN103675588B (en) | The machine vision detection method of printed component part polarity and equipment | |
| CN102110332B (en) | Book registering and managing device based on computer vision and radio frequency identification technology | |
| CN103091331B (en) | System and method for visual inspection on burrs and stain defects of radio frequency identification (RFID) antennae | |
| CN109886298A (en) | A kind of detection method for quality of welding line based on convolutional neural networks | |
| AU2019260856A1 (en) | System and method for performing automated analysis of air samples | |
| CN106370671A (en) | PCB (printed circuit board) component detection system and method based on machine vision | |
| WO2017092427A1 (en) | Electronic element positioning method and apparatus | |
| CN106127746A (en) | Circuit board element missing part detection method and system | |
| CN105510348A (en) | Flaw detection method and device of printed circuit board and detection equipment | |
| CN103886294B (en) | Method and system for recognizing number of movement times of arrester for convertor station based on images | |
| KR20130064853A (en) | Secondary battery eloctrode panel vision detection method | |
| CN111312023B (en) | Device and method for automatically drawing middle school physics circuit experiment circuit diagram | |
| CN106651849A (en) | Area-array camera-based PCB bare board defect detection method | |
| CN116363573A (en) | Transformer substation equipment state anomaly identification method and system | |
| CN113723325A (en) | Tool defect detection system for prefabricated parts | |
| CN108156762A (en) | A kind of PCB production methods | |
| CN204613128U (en) | Circuit board detection device | |
| CN113794277A (en) | Method and system for identifying pressing plate state image | |
| CN106872488A (en) | A kind of double surface defect visible detection methods of rapid large-area transparent substrate and device | |
| CN107084989B (en) | Method and system for adding AOI device database | |
| CN114581760B (en) | Equipment fault detection method and system for machine room inspection | |
| CN114518526A (en) | Automatic testing machine control system suitable for PCB board ICT |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20170710 Address after: Tokyo City, Japan, and bin Du District, Tian marsh two Ding mu, 3 times 9 Applicant after: International Technology Development Corp Address before: 215021, 1913-1, international science and Technology Park, 1355 Jinji Lake Road, Jiangsu, Suzhou Applicant before: Kejke precision electronic technology development (Suzhou) Co., Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |