CN108896567A - A kind of pair of button cell surface weighs the method that defect is detected wounded - Google Patents
A kind of pair of button cell surface weighs the method that defect is detected wounded Download PDFInfo
- Publication number
- CN108896567A CN108896567A CN201810885589.6A CN201810885589A CN108896567A CN 108896567 A CN108896567 A CN 108896567A CN 201810885589 A CN201810885589 A CN 201810885589A CN 108896567 A CN108896567 A CN 108896567A
- Authority
- CN
- China
- Prior art keywords
- image
- defect
- wounded
- button cell
- weighs
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 50
- 210000004027 cell Anatomy 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 29
- 230000003287 optical effect Effects 0.000 claims abstract description 15
- 230000008569 process Effects 0.000 claims abstract description 12
- 239000011159 matrix material Substances 0.000 claims abstract description 9
- 238000005530 etching Methods 0.000 claims abstract description 5
- 238000013527 convolutional neural network Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 20
- 238000013528 artificial neural network Methods 0.000 claims description 18
- 238000005303 weighing Methods 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 4
- 230000002950 deficient Effects 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims description 2
- 230000001537 neural effect Effects 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 15
- 238000004519 manufacturing process Methods 0.000 abstract description 10
- 238000003384 imaging method Methods 0.000 abstract description 7
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 2
- 239000000428 dust Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000003908 quality control method Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 101100117236 Drosophila melanogaster speck gene Proteins 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 210000004218 nerve net Anatomy 0.000 description 1
- 238000005498 polishing Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of pair of button cell surfaces to weigh the method that defect is detected wounded, specifically includes following steps：S1, optical imagery first is carried out to product surface：Using profession customization without shadow area source, there is the dot matrix of laser-induced thermal etching on light source surface, dot matrix projects to the pattern that button cell positive electrode surface forms similar lattice-shaped, if battery surface, which exists, weighs defect wounded, certain optical deformation can occur for comb mesh pattern, even slight weighs wounded, this optical deformation is still had, and is related to battery surface detection technique field.This weighs the method that defect is detected wounded to button cell surface, greatly improve the accuracy and efficiency of product defects detection, complete the imaging and detection that button cell positive electrode surface weighs wounded, the uncontrollable factors such as fail to judge and judge by accident instead of human factor bring in traditional artificial detection process, it is high-efficient, accuracy rate is high, is greatly saved the cost of production, improves product quality.
Description
Technical field
The present invention relates to battery surface detection technique field, specially a kind of pair of button cell surface weighs defect wounded and examines
The method of survey.
Background technique
Button cell is also referred to as button cell, refers to that outer dimension as the battery of a small button, is in general relatively large in diameter,
Thinner thickness, button cell are to divide from shape battery, and same corresponding cell classification has column battery, a square electric
Pond, Special-shaped battery, button cell are widely used in various miniature electronic products because the bodily form is smaller, diameter from
4.8mm to 30mm, thickness are differed from 1.0mm to 7.7mm, are generally used for the backup power supply of each electronic product, as computer main board,
Electronic watch, electronic dictionary, electronic scale, memory card, remote controler, electronic toy, pacemaker, electronic hearing aid, counter and
Camera etc., button cell are also classified into chemical cell and physical battery two major classes, and chemical cell application is the most universal, they by
The compositions such as anode (anode) agent, cathode (cathode) agent and its electricity touching liquid, in process of production, surface is easy to appear pressure to button cell
Hurt defect, currently, each production plant is all to be detected by worker to every battery surface, due to its defect geometry feature
Randomness and high anti-espionage to illumination, worker need battery transformation all angles to be observed in detail, if there is weighing wounded,
Anode surface will appear the features such as blackening or speck, but manually be held during identifying defect due to long-time eye
It easily fails to judge, the problems such as erroneous judgement, causes defective products to flow to client, customer complaint is caused to be returned goods.
The accuracy and efficiency of current button cell product defects detection is lower, can not achieve to button cell anode table
The imaging and detection of face pressure wound, can not fail to judge and judge by accident instead of human factor bring in traditional artificial detection process etc. can not
Control factor, efficiency and accuracy rate are low, greatly waste the cost of production, can not improve the quality of product, existing defect inspection
Survey during can not rule out interference caused by the other factors such as dust, can not achieve extract a small amount of sample in process of production can also
It to obtain outstanding network model, gets a desired effect, so that solution quality control standard in process of production cannot be reached
Disunity, the purpose failed to judge and judge serious problem by accident.
Summary of the invention
(1) the technical issues of solving
In view of the deficiencies of the prior art, the present invention provides a kind of pair of button cell surfaces to weigh the side that defect is detected wounded
Method, the accuracy and efficiency for solving existing button cell product defects detection is lower, can not achieve to button cell anode
The problem of imaging and detection that surface weighs wounded, efficiency and accuracy rate are low, greatly waste production cost.
(2) technical solution
In order to achieve the above object, the present invention is achieved by the following technical programs：A kind of pair of button cell surface weighs wounded
The method that defect is detected, specifically includes following steps：
S1, optical imagery first is carried out to product surface：Using profession customization without shadow area source, there is laser-induced thermal etching on light source surface
Dot matrix, dot matrix projects to the pattern that button cell positive electrode surface forms similar lattice-shaped, if battery surface exist weigh wounded it is scarce
It falls into, certain optical deformation can occur for comb mesh pattern, even slight weighs wounded, this optical deformation is still had；
S2, training sample and test sample are formed；The a large amount of positive button cell sample with defect is collected, high-resolution is used
Rate industrial CCD carries out Image Acquisition, forms training sample and test sample image library；
S3, the defect characteristic on training sample surface is manually marked；Using the picture for marking defect in S2, using certainly
The image labeling tool of main exploitation is labeled, and OK product will participate in marking with NG product, is weighed defective locations wounded and is needed with one
Kind color paintbrush is marked according to pixel region, and the mark image of generation is in addition to marked region has color, and other parts are all
The picture of black and white is presented；
S4, the image after mark is enhanced, and utilizes convolutional neural networks training；It is handle to the image increase in S3
Image-region after mark carries out change of scale, such as rotation, scaling, mirror image, contrast variation's one or more group
It closes, expands training radix, form new training sample database, the sample database after increase is inputted in convolutional neural networks and is instructed
Practice, until the accuracy rate that stable convergence is reached by successive ignition, can also be obtained in the case where sample is less so outstanding
Convolutional neural networks model；
S5, input test sample prediction result into neural network；Save the network model that S4 is generated, the test to input
Sample image is predicted that each test sample image generates N characteristic patterns, each picture in the N characteristic patterns by prediction
The feature that element represents in original image on the location of pixels belongs to one of (N-1) kind defect probability score, that is, produces one
Network model；
S6, the target information extracted and needed is analyzed with BLOB；One is arranged to the probability score for weighing defect in step S5 wounded
Threshold value generates cluster index figure, and the mode with BLOB analysis area filtering is the location information where extractable defect area.
Preferably, the convolutional neural networks in the step S5 are a kind of feedforward neural networks, and artificial neuron can ring
Surrounding cells are answered, large-scale image procossing can be carried out, convolutional neural networks include convolutional layer and pond layer, convolutional neural networks packet
One-dimensional convolutional neural networks, two-dimensional convolution neural network and Three dimensional convolution neural network are included, one-dimensional convolutional neural networks are often answered
Data processing for sequence class；Two-dimensional convolution neural network is commonly applied to the identification of image class text；Three dimensional convolution nerve net
Network is mainly used in medical image and video class data identification.
Preferably, the BLOB in the step S6, which refers in computer vision in image, has Similar color, texture
Etc. one piece of connected region composed by features, BLOB analysis is that image is carried out to binaryzation, and segmentation obtains foreground and background, then
Connected region detection is carried out, to obtain the process of BLOB region unit.
(3) beneficial effect
The present invention provides a kind of pair of button cell surfaces to weigh the method that defect is detected wounded.Have compared with prior art
Standby following beneficial effect：This weighs the method that defect is detected wounded to button cell surface, specifically includes following steps：S1, elder generation
Optical imagery is carried out to product surface：Using profession customization without shadow area source, there is the dot matrix of laser-induced thermal etching on light source surface, and dot matrix is thrown
Shadow forms the pattern of similar lattice-shaped to button cell positive electrode surface, S2, forms training sample and test sample；It collects a large amount of
Button cell sample of the anode with defect carries out Image Acquisition with high resolution industrial CCD, forms training sample and test sample
S3, image library manually marks the defect characteristic on training sample surface；Using the picture for marking defect in S2, using certainly
The image labeling tool of main exploitation is labeled, and OK product will participate in marking with NG product, S4, carries out the image after mark
Enhancing, and utilize convolutional neural networks training；It is that the image-region after mark is carried out change of scale to the image increase in S3,
S5, input test sample prediction result into neural network；The network model that S4 is generated is saved, to the test sample image of input
It is predicted, each test sample image generates N characteristic patterns by prediction, and the target extracted and needed S6, is analyzed with BLOB
Information；One threshold value is arranged to the probability score for weighing defect in step S5 wounded, generates cluster index figure, analyzes area mistake with BLOB
The mode of filter is the location information where extractable defect area, realize using special polishing mode to positive electrode surface carry out at
Picture is analyzed defect image using artificial intelligence deep learning algorithm, to carry out intellectual determination to surface defects of products
Method, greatly improve product defects detection accuracy and efficiency, using profession customization without shadow area source to button cell just
Pole surface is imaged, and is weighed defect wounded and is shown as imaging surface generation optical deformation, it is special to breach high light-reflecting product surface crater
The technological difficulties for levying imaging, also eliminate interference caused by the other factors such as dust, realize and are increased to training sample using data
Add function, a small amount of sample can also obtain outstanding network model in process of production, get a desired effect, while using deep
Degree learning algorithm identifies the optical deformation feature in imaging, breaches conventional machines vision algorithm in appearance defect analysis
The inadequate bottleneck of middle recall rate, while substituting desk checking, solve quality control standard disunity in production process, fail to judge and
Serious problem is judged by accident, so as to complete imaging and detection that button cell positive electrode surface weighs wounded, instead of traditional artificial detection
Human factor bring such as fails to judge and judges by accident at the uncontrollable factors in the process, high-efficient, and accuracy rate is high, is greatly saved production
Cost improves product quality.
Specific embodiment
Below in conjunction with the embodiment of the present invention, technical scheme in the embodiment of the invention is clearly and completely described,
Obviously, described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Based in the present invention
Embodiment, every other embodiment obtained by those of ordinary skill in the art without making creative efforts, all
Belong to the scope of protection of the invention.
The present invention provides a kind of technical solution：A kind of pair of button cell surface weighs the method that defect is detected wounded, specifically
Include the following steps：
S1, optical imagery first is carried out to product surface：Using profession customization without shadow area source, there is laser-induced thermal etching on light source surface
Dot matrix, dot matrix projects to the pattern that button cell positive electrode surface forms similar lattice-shaped, if battery surface exist weigh wounded it is scarce
It falls into, certain optical deformation can occur for comb mesh pattern, even slight weighs wounded, this optical deformation is still had；
S2, training sample and test sample are formed；The a large amount of positive button cell sample with defect is collected, high-resolution is used
Rate industrial CCD carries out Image Acquisition, forms training sample and test sample image library；
S3, the defect characteristic on training sample surface is manually marked；Using the picture for marking defect in S2, using certainly
The image labeling tool of main exploitation is labeled, and OK product will participate in marking with NG product, is weighed defective locations wounded and is needed with one
Kind color paintbrush is marked according to pixel region, and the mark image of generation is in addition to marked region has color, and other parts are all
The picture of black and white is presented；
S4, the image after mark is enhanced, and utilizes convolutional neural networks training；It is handle to the image increase in S3
Image-region after mark carries out change of scale, such as rotation, scaling, mirror image, contrast variation's one or more group
It closes, expands training radix, form new training sample database, the sample database after increase is inputted in convolutional neural networks and is instructed
Practice, until the accuracy rate that stable convergence is reached by successive ignition, can also be obtained in the case where sample is less so outstanding
Convolutional neural networks model；
S5, input test sample prediction result into neural network；Save the network model that S4 is generated, the test to input
Sample image is predicted that each test sample image generates N characteristic patterns, each picture in the N characteristic patterns by prediction
The feature that element represents in original image on the location of pixels belongs to one of (N-1) kind defect probability score, that is, produces one
Network model；
S6, the target information extracted and needed is analyzed with BLOB；One is arranged to the probability score for weighing defect in step S5 wounded
Threshold value generates cluster index figure, and the mode with BLOB analysis area filtering is the location information where extractable defect area.
In the present invention, the convolutional neural networks in step S5 are a kind of feedforward neural networks, and artificial neuron can respond
Surrounding cells can carry out large-scale image procossing, and convolutional neural networks include convolutional layer and pond layer, and convolutional neural networks include
One-dimensional convolutional neural networks, two-dimensional convolution neural network and Three dimensional convolution neural network, one-dimensional convolutional neural networks are often applied
In the data processing of sequence class；Two-dimensional convolution neural network is commonly applied to the identification of image class text；Three dimensional convolution neural network
It is mainly used in medical image and video class data identification.
In the present invention, the BLOB in step S6, which refers in computer vision in image, has Similar color, texture etc.
One piece of connected region composed by feature, BLOB analysis be that image is carried out to binaryzation, segmentation obtain foreground and background, then into
Row connected region detection, to obtain the process of BLOB region unit.
It should be noted that, in this document, relational terms such as first and second and the like are used merely to a reality
Body or operation are distinguished with another entity or operation, are deposited without necessarily requiring or implying between these entities or operation
In any actual relationship or order or sequence.Moreover, the terms "include", "comprise" or its any other variant are intended to
Non-exclusive inclusion, so that the process, method, article or equipment including a series of elements is not only wanted including those
Element, but also including other elements that are not explicitly listed, or further include for this process, method, article or equipment
Intrinsic element.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of variations, modification, replacement can be carried out to these embodiments without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (3)
1. a kind of pair of button cell surface weighs the method that defect is detected wounded, it is characterised in that：Specifically include following steps：
S1, optical imagery first is carried out to product surface：Using profession customization without shadow area source, there is the point of laser-induced thermal etching on light source surface
Battle array, dot matrix projects to the pattern that button cell positive electrode surface forms similar lattice-shaped, if battery surface, which exists, weighs defect, grid wounded
Certain optical deformation can occur for grid pattern, even slight weighs wounded, this optical deformation is still had；
S2, training sample and test sample are formed；The a large amount of positive button cell sample with defect is collected, with high-resolution work
Industry CCD carries out Image Acquisition, forms training sample and test sample image library；
S3, the defect characteristic on training sample surface is manually marked；Using the picture for marking defect in S2, using independently opening
The image labeling tool of hair is labeled, and OK product will participate in marking with NG product, is weighed defective locations wounded and is needed with a kind of face
Color paintbrush is marked according to pixel region, and the mark image of generation is in addition to marked region has color, and other parts are all presented
The picture of black and white；
S4, the image after mark is enhanced, and utilizes convolutional neural networks training；It is mark to the image increase in S3
Image-region afterwards carries out change of scale, such as rotation, scaling, mirror image, contrast variation's a combination of one or more, expands
Big training radix, forms new training sample database, being trained in the sample database input convolutional neural networks after increase, passes through
Until successive ignition reaches the accuracy rate of stable convergence, outstanding convolutional Neural can be also obtained in the case where sample is less in this way
Network model；
S5, input test sample prediction result into neural network；The network model that S4 is generated is saved, to the test sample of input
Image is predicted that each test sample image generates N characteristic patterns, each pixel generation in the N characteristic patterns by prediction
Feature of the table in original image on the location of pixels belongs to one of (N-1) kind defect probability score, that is, produces a network
Model；
S6, the target information extracted and needed is analyzed with BLOB；One threshold value is arranged to the probability score for weighing defect in step S5 wounded,
Cluster index figure is generated, the mode with BLOB analysis area filtering is the location information where extractable defect area.
2. a kind of pair of button cell surface according to claim 1 weighs the method that defect is detected wounded, it is characterised in that：
Convolutional neural networks in the step S5 are a kind of feedforward neural networks, and artificial neuron can respond surrounding cells, can be with
Large-scale image procossing is carried out, convolutional neural networks include convolutional layer and pond layer, and convolutional neural networks include one-dimensional convolutional Neural
Network, two-dimensional convolution neural network and Three dimensional convolution neural network, one-dimensional convolutional neural networks are commonly applied to the number of sequence class
According to processing；Two-dimensional convolution neural network is commonly applied to the identification of image class text；Three dimensional convolution neural network is mainly used in doctor
Learn image and video class data identification.
3. a kind of pair of button cell surface according to claim 1 weighs the method that defect is detected wounded, it is characterised in that：
BLOB in the step S6 refers to having one composed by Similar color, Texture eigenvalue in image in computer vision
Block connected region, BLOB analysis is that image is carried out to binaryzation, and segmentation obtains foreground and background, then carries out connected region inspection
It surveys, to obtain the process of BLOB region unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810885589.6A CN108896567B (en) | 2018-08-06 | 2018-08-06 | Method for detecting crush defect of surface of button cell |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810885589.6A CN108896567B (en) | 2018-08-06 | 2018-08-06 | Method for detecting crush defect of surface of button cell |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108896567A true CN108896567A (en) | 2018-11-27 |
| CN108896567B CN108896567B (en) | 2024-06-04 |
Family
ID=64353462
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810885589.6A Active CN108896567B (en) | 2018-08-06 | 2018-08-06 | Method for detecting crush defect of surface of button cell |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108896567B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829483A (en) * | 2019-01-07 | 2019-05-31 | 鲁班嫡系机器人（深圳）有限公司 | Defect recognition model training method, device, computer equipment and storage medium |
| CN110108710A (en) * | 2019-01-22 | 2019-08-09 | 镇江苏仪德科技有限公司 | One kind being used for button cell anode concave point detection device and method |
| CN110232404A (en) * | 2019-05-21 | 2019-09-13 | 江苏理工学院 | A kind of recognition methods of industrial products surface blemish and device based on machine learning |
| CN111458351A (en) * | 2019-01-18 | 2020-07-28 | 无锡先导智能装备股份有限公司 | Battery detection system and battery detection method |
| CN112024422A (en) * | 2020-07-15 | 2020-12-04 | 镇江苏仪德科技有限公司 | Button cell surface defect detecting system |
| CN113591404A (en) * | 2021-09-29 | 2021-11-02 | 杭州宇谷科技有限公司 | Battery abnormity detection system and method based on deep learning |
| CN113884016A (en) * | 2021-06-16 | 2022-01-04 | 成都新锐科技发展有限责任公司 | Battery piece warping degree detection method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107543828A (en) * | 2017-08-25 | 2018-01-05 | 广东工业大学 | A kind of Surface Flaw Detection method and system |
| CN108038853A (en) * | 2017-12-18 | 2018-05-15 | 浙江工业大学 | Ceramic tile surface defect identification method based on convolutional neural network and active learning |
| CN108171707A (en) * | 2018-01-23 | 2018-06-15 | 武汉精测电子集团股份有限公司 | A kind of Mura defects level evaluation method and device based on deep learning |
| CN108292366A (en) * | 2015-09-10 | 2018-07-17 | 美基蒂克艾尔有限公司 | The system and method that suspect tissue region is detected in endoscopic surgery |
| CN207649553U (en) * | 2017-11-09 | 2018-07-24 | 彩虹集团（邵阳）特种玻璃有限公司 | A kind of online plate glass check device |
| CN208654046U (en) * | 2018-08-06 | 2019-03-26 | 慧泉智能科技（苏州）有限公司 | A kind of button cell surface weighs the device that defect is detected wounded |
-
2018
- 2018-08-06 CN CN201810885589.6A patent/CN108896567B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108292366A (en) * | 2015-09-10 | 2018-07-17 | 美基蒂克艾尔有限公司 | The system and method that suspect tissue region is detected in endoscopic surgery |
| CN107543828A (en) * | 2017-08-25 | 2018-01-05 | 广东工业大学 | A kind of Surface Flaw Detection method and system |
| CN207649553U (en) * | 2017-11-09 | 2018-07-24 | 彩虹集团（邵阳）特种玻璃有限公司 | A kind of online plate glass check device |
| CN108038853A (en) * | 2017-12-18 | 2018-05-15 | 浙江工业大学 | Ceramic tile surface defect identification method based on convolutional neural network and active learning |
| CN108171707A (en) * | 2018-01-23 | 2018-06-15 | 武汉精测电子集团股份有限公司 | A kind of Mura defects level evaluation method and device based on deep learning |
| CN208654046U (en) * | 2018-08-06 | 2019-03-26 | 慧泉智能科技（苏州）有限公司 | A kind of button cell surface weighs the device that defect is detected wounded |
Non-Patent Citations (2)
| Title |
|---|
| 陈宁纪: "基于视觉感知的带钢表面缺陷检测与识别", 中国优秀硕士学位论文全文数据库工程科技Ⅰ辑, no. 4 * |
| 鲁少辉: "基于主动式全景视觉的管道内表面缺陷检测技术的研究", 中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑, no. 4 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829483A (en) * | 2019-01-07 | 2019-05-31 | 鲁班嫡系机器人（深圳）有限公司 | Defect recognition model training method, device, computer equipment and storage medium |
| CN109829483B (en) * | 2019-01-07 | 2021-05-18 | 鲁班嫡系机器人（深圳）有限公司 | Defect recognition model training method and device, computer equipment and storage medium |
| CN111458351A (en) * | 2019-01-18 | 2020-07-28 | 无锡先导智能装备股份有限公司 | Battery detection system and battery detection method |
| CN110108710A (en) * | 2019-01-22 | 2019-08-09 | 镇江苏仪德科技有限公司 | One kind being used for button cell anode concave point detection device and method |
| CN110232404A (en) * | 2019-05-21 | 2019-09-13 | 江苏理工学院 | A kind of recognition methods of industrial products surface blemish and device based on machine learning |
| CN112024422A (en) * | 2020-07-15 | 2020-12-04 | 镇江苏仪德科技有限公司 | Button cell surface defect detecting system |
| CN113884016A (en) * | 2021-06-16 | 2022-01-04 | 成都新锐科技发展有限责任公司 | Battery piece warping degree detection method |
| CN113884016B (en) * | 2021-06-16 | 2024-02-13 | 成都新锐科技发展有限责任公司 | Method for detecting warping degree of battery piece |
| CN113591404A (en) * | 2021-09-29 | 2021-11-02 | 杭州宇谷科技有限公司 | Battery abnormity detection system and method based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108896567B (en) | 2024-06-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108896567A (en) | A kind of pair of button cell surface weighs the method that defect is detected wounded | |
| CN111028204B (en) | Cloth defect detection method based on multi-mode fusion deep learning | |
| CN109559310B (en) | Power transmission and transformation inspection image quality evaluation method and system based on significance detection | |
| CN114240939B (en) | Method, system, equipment and medium for detecting appearance defects of mainboard components | |
| CN111260788B (en) | Power distribution cabinet switch state identification method based on binocular vision | |
| CN109858480A (en) | Digital instrument identification method | |
| CN114418899B (en) | Self-adaptive repairing method and system for self-color printing and readable storage medium | |
| CN112215821A (en) | Photovoltaic cell panel grid line defect detection method and system based on artificial intelligence | |
| CN112396635B (en) | Multi-target detection method based on multiple devices in complex environment | |
| CN107356213A (en) | Optical filter concentricity measuring method and terminal device | |
| CN109916913A (en) | A kind of intelligence manufacture product identification and detection method based on machine vision | |
| CN115631350B (en) | Method and device for identifying colors of canned image | |
| CN109886391A (en) | A kind of neural network compression method based on the positive and negative diagonal convolution in space | |
| Anis et al. | Digital electric meter reading recognition based on horizontal and vertical binary pattern | |
| CN114120051A (en) | Method for screening edge points of adhesive tape | |
| CN110084746A (en) | A kind of knowledge figure pattern splicing method of mosaic machine people | |
| CN113205136A (en) | Real-time high-precision detection method for appearance defects of power adapter | |
| CN108198226A (en) | A kind of ceramic color recognition methods, electronic equipment, storage medium and device | |
| CN109350018B (en) | The palm detection method based on image applied to palm bleb detection system | |
| CN115841731A (en) | Infrared-monitoring park fire early warning method | |
| CN114066993B (en) | Power distribution cabinet control panel segmentation method based on machine vision | |
| CN108827974A (en) | A kind of ceramic tile defect inspection method and system | |
| CN112818836B (en) | Method and system for detecting personnel target of transformer substation scene | |
| CN114783005A (en) | Palm print recognition equipment and method based on deep learning | |
| CN112949666B (en) | Image recognition method for fault cathode plate of zinc stripping machine |
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 |