CN109242830A - A kind of machine vision technique detection method based on deep learning - Google Patents
A kind of machine vision technique detection method based on deep learning Download PDFInfo
- Publication number
- CN109242830A CN109242830A CN201810944235.4A CN201810944235A CN109242830A CN 109242830 A CN109242830 A CN 109242830A CN 201810944235 A CN201810944235 A CN 201810944235A CN 109242830 A CN109242830 A CN 109242830A
- Authority
- CN
- China
- Prior art keywords
- deep learning
- industrial
- network model
- learning network
- detection
- 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.)
- Pending
Links
- 238000013135 deep learning Methods 0.000 title claims abstract description 30
- 238000001514 detection method Methods 0.000 title claims abstract description 23
- 238000000034 method Methods 0.000 title claims abstract description 19
- 230000007547 defect Effects 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 11
- 230000002950 deficient Effects 0.000 claims abstract description 8
- 238000007689 inspection Methods 0.000 claims abstract description 8
- 239000000463 material Substances 0.000 claims abstract description 4
- 238000011017 operating method Methods 0.000 claims description 2
- 238000010801 machine learning Methods 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000008921 facial expression Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008450 motivation Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/0008—Industrial image inspection checking presence/absence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Software Systems (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
A kind of machine vision technique detection method based on deep learning, the present invention relates to industrial detection technical fields；A large amount of defective product picture is acquired as sample under good industrial light source by industrial camera；These samples pictures are demarcated by software, mark the fault location position of care, and mark defect kind；Using the above-mentioned samples pictures for marking defect kind as material hook deep learning network model, and carry out the training of deep learning network model；Trained deep learning network model is imported into NI Vision Builder for Automated Inspection, to identify the various defects at scene, cooperation industrial automation equipment completes the sorting of faulty goods.Missing inspection erroneous detection is avoided, can also learn new sample and accuracy is continuously improved, adaptability and stability are good, realize expert system to meet the industrial detection demand of various complex scenes.
Description
Technical field
The present invention relates to industrial detection technical fields, and in particular to a kind of machine vision technique detection based on deep learning
Method.
Background technique
The concept of deep learning is derived from the research of artificial neural network.Multilayer perceptron containing more hidden layers is exactly a kind of depth
Learning structure.Deep learning, which forms more abstract high level by combination low-level feature, indicates attribute classification or feature, with discovery
The distributed nature of data indicates.Deep learning is a kind of method based on to data progress representative learning in machine learning,
It is a new field in machine learning research, motivation is to establish, simulate the neural network that human brain carries out analytic learning,
It imitates the mechanism of human brain to explain data, such as image, sound and text.Observation (such as piece image) can be used more
Kind of mode indicates, such as vector of each pixel intensity value, or is more abstractively expressed as a series of region on sides, specific shape
Deng.And certain specific representation methods is used to be easier from example learning tasks (for example, recognition of face or facial expression are known
Not).The benefit of deep learning is that the feature learning and layered characteristic with non-supervisory formula or Semi-supervised extract highly effective algorithm to replace
In generation, obtains feature by hand.
Deep learning has two big main advantages relative to traditional machine learning method:
1, excellent result: the model by deep learning training in all kinds of artificial intelligence fields, (know by computer vision, voice
Not, natural language processing etc.) effect promoted than original machine learning method it is obvious.Typical example is exactly [31] ImageNet
Contest in 2012, the method based on deep learning not only obtain the achievement of first place, even more have dropped error rate close
Half；
2, easy to use: traditional machine learning method is larger to the feature dependence of model, and these features generally require
Preferable result could be obtained by constantly being debugged by the personnel of profession；Deep learning is a kind of feature learning method, is not required to very important person
Work writes feature, it can the automatic learning characteristic from initial data (image, audio, text), and these features are following
It can input again and go to be predicted in classifier.Therefore, using deep learning method, machine learning can significantly be reduced at this
The threshold applied in a little fields.
The disadvantage of deep learning is that training burden is big, generally requires the training time more much more than conventional method.But this problem
It is addressed with the optimization of Parallel Algorithm and popularizing for graphics acceleration card (GPU) and FPGA, the training time is no longer system
The bottleneck about used, therefore further improve the universality that deep learning uses.Since deep learning compares traditional machine
These of device study are dramatically different, therefore traditional machine learning method is also known as " shallow-layer study " (Shallow
Learning), among these including SVM, boosting, KNN etc..
Machine learning method generally use in terms of industrial detection at present or traditional, has as a drawback that and limits to
Property:
1, artificial extraction feature is needed, different classes of differentiation, difference numerous for defect kind can not be completely expressed
Unconspicuous situation, classification accuracy is low, is easy erroneous detection and missing inspection；
2, by light, material, the operating conditions such as ambient noise are affected, and adaptability and stability are poor；
3, it needs first to detect fault location, then classifies, can not carry out simultaneously, low efficiency；
4, can not be to defect and quality the problems such as, is accurately given a mark and is graded.
Summary of the invention
In view of the defects and deficiencies of the prior art, the present invention intends to provide a kind of structure is simple, design rationally, make
With the machine vision technique detection method easily based on deep learning, missing inspection erroneous detection is avoided, it is continuous can also to learn new sample
Accuracy is improved, adaptability and stability are good, realize expert system to meet the industrial detection demand of various complex scenes.
To achieve the above object, the technical solution adopted by the present invention is that: its operating procedure is as follows:
1, a large amount of defective product picture is acquired as sample, such as hole under good industrial light source by industrial camera
The problems such as hole, abrasion, burr, recess；
2, these samples pictures are demarcated by software, marks the fault location position of care, and mark defect kind；
3, using the above-mentioned samples pictures for marking defect kind as material hook deep learning network model, and depth is carried out
Spend the training of learning network model；
4, trained deep learning network model is imported into NI Vision Builder for Automated Inspection, to identify that the various of scene lack
It falls into, cooperation industrial automation equipment completes the sorting of faulty goods.
The working principle of the invention: the present invention utilizes the main characteristics of deep learning, that is, layer-by-layer feature learning.By more
The neural network of layer, deep learning is constantly drawn into further feature from the characterization of preceding layer, to reduce data
Dimension, it is easier to find substantive characteristics, by taking the deep neural network of four layers of hidden layer as an example, first layer is from original image
The feature of edge (edge) is arrived in middle study, and for the second layer on the edge of first layer, the feature of pattern (motifs) is arrived in study, the
Three layers on the basis of pattern, learn the feature to composition (part), and the 4th layer on the composition of third layer, learns to arrive target
(object) feature of feature, ideal inputs classifier, goes to complete identification or Detection task.In the framework of whole network
In, it is responsible for for subsequent every layer being abstracted upper one layer of input, more advanced content is arrived in constantly study.Deep learning it is this
Method, the part for being easy to keep model learning most condensed into data, so as to improve classification and recognition effect.
After above-mentioned steps, the invention has the following beneficial effects:
1, feature is not extracted manually, it is different classes of subtleer than more fully expressing with the mathematic(al) representation of higher dimensional space
Feature avoids missing inspection erroneous detection, can also learn new sample and accuracy is continuously improved；
2, by the interference to training sample scrambled algorithm simulation difference light and noise, the model after can training can
To overcome these interference problems, adaptability and stability are good；
3, it can be completed at the same time the multi-tasks such as defects detection and classification end to end, efficiency is very high；
4, can be with the experience of learning simulation expert, to defect and quality the problems such as is accurately given a mark and is graded, and is realized
Expert system meets the industrial detection demands of various complex scenes.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is work flow diagram of the invention.
Fig. 2 is depth convolutional neural networks illustraton of model in specific embodiment.
Specific embodiment
The present invention will be further described below with reference to the drawings.
Referring to as depicted in figs. 1 and 2, present embodiment is will be based on the machine vision technique detection side of deep learning
Method is applied to the identification of Defect Capacitance, the technical solution adopted is that:
1, actual capacitance defective data collection Dataset1 is constructed, sorts automated system phase using with final detection as far as possible
Same industrial camera and light source shoots Defect Capacitance, chooses about 6000 Defect Capacitance images, to its electricity of each width image labeling
Hold defective locations and attribute (attribute value of Defect Capacitance includes Defect Capacitance type and severity values)；
2, depth convolutional neural networks model is constructed, network architecture is by several convolutional layers, pond layer, full connection
Layer and an output layer are constituted, the defect that wherein output of the last one convolutional layer of core network is detected by processing
Position and attribute；Network model network is made of 24 convolutional layers, 5 pond layers, 2 full articulamentums；Wherein, convolutional layer uses
Certain amount and a certain size convolution kernel carry out convolution operation and obtain, convolution operation is defined as:
Wherein f is convolution kernel, and r is the size of convolution kernel, and f (i, j) indicates value of the convolution kernel at coordinate (i, j), I
For convolution input, commonly referred to as characteristic pattern, g (s, t) indicates carrying out the output valve that convolution operation obtains at point (s, t) to I；
Further, each convolutional layer uses different number convolution kernels, and the size of convolution kernel is 3 or 1；There are 5 ponds
Change layer, so-called pond refers to and carries out down-sampling to each characteristic pattern, and one value of selection, which is used as, in certain contiguous range adopts
Sample value, it is preferable that pondization uses the maximum pond method of 2 × 2 neighborhoods；Preferably, the activation letter of constructed convolutional neural networks
Number uses Rectified Linear Units function；
Different dimensions tensor is respectively adopted in three full articulamentums, wherein the corresponding output of the last one full articulamentum one
Amount, dimension are as follows: S*S* (B*5+C) S*S* (B*5+C), wherein S is grid division number, and B is that each grid is responsible for target number, C
For classification number；The expression formula meaning are as follows:
(1) each small lattice can correspond to B bounding box, and the wide high scope of bounding box is full figure, indicate centered on the small lattice
Find the bounding box position of defect；
(2) the corresponding score value of each bounding box, represent at this whether defective and positional accuracy；
(3) each small lattice can correspond to C probability value, find out the corresponding defect classification of maximum probability, and think to wrap in small lattice
A part containing the defect or the defect；
3, using the emulation Defect Capacitance data set Dataset1 constructed in step 1 to depth convolution constructed by step 2
Neural network model is trained, and the condition that training process stops is two kinds, and one is the thresholds that the value of loss function is less than setting
Value T, another kind are that training reaches certain times N, save trained model and parameter, wherein T value 0.01, N value
100000；It finally selects suitable loss function and classifier finally to determine defective locations and attribute, filters out inappropriate
Prediction；
Further, the loss function of whole network is defined as Softmax Loss, and is determined with Softmax classifier
Final defective locations and attribute；
4, the depth convolutional neural networks model and the resulting parameter of step 3 constructed using step 2 is to image to be identified
Defect Capacitance identification is carried out, the position as Defect Capacitance and attribute are exported.
After above-mentioned steps, present embodiment has the beneficial effect that one kind described in present embodiment is based on
The machine vision technique detection method of deep learning avoids missing inspection erroneous detection, can also learn new sample and accuracy is continuously improved, fit
Answering property and stability are good, realize expert system to meet the industrial detection demand of various complex scenes.
The above is only used to illustrate the technical scheme of the present invention and not to limit it, and those of ordinary skill in the art are to this hair
The other modifications or equivalent replacement that bright technical solution is made, as long as it does not depart from the spirit and scope of the technical scheme of the present invention,
It is intended to be within the scope of the claims of the invention.
Claims (1)
1. a kind of machine vision technique detection method based on deep learning, it is characterised in that: its operating procedure is as follows:
(1), a large amount of defective product picture is acquired as sample under good industrial light source by industrial camera；
(2), these samples pictures are demarcated by software, marks the fault location position of care, and mark defect kind；
(3), using the above-mentioned samples pictures for marking defect kind as material hook deep learning network model, and depth is carried out
The training of learning network model；
(4), trained deep learning network model is imported into NI Vision Builder for Automated Inspection, thus identify the various defects at scene,
Industrial automation equipment is cooperated to complete the sorting of faulty goods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810944235.4A CN109242830A (en) | 2018-08-18 | 2018-08-18 | A kind of machine vision technique detection method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810944235.4A CN109242830A (en) | 2018-08-18 | 2018-08-18 | A kind of machine vision technique detection method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109242830A true CN109242830A (en) | 2019-01-18 |
Family
ID=65071431
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810944235.4A Pending CN109242830A (en) | 2018-08-18 | 2018-08-18 | A kind of machine vision technique detection method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109242830A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110223266A (en) * | 2019-03-08 | 2019-09-10 | 湖南工业大学 | A kind of Railway wheelset tread damage method for diagnosing faults based on depth convolutional neural networks |
| CN110335239A (en) * | 2019-05-09 | 2019-10-15 | 菲特(天津)智能科技有限公司 | Defects detection training machine and its application method based on deep learning |
| CN110530875A (en) * | 2019-08-29 | 2019-12-03 | 珠海博达创意科技有限公司 | A kind of FPCB open defect automatic detection algorithm based on deep learning |
| CN111982910A (en) * | 2020-07-06 | 2020-11-24 | 华南理工大学 | Weak supervision machine vision detection method and system based on artificial defect simulation |
| CN112083002A (en) * | 2020-08-26 | 2020-12-15 | 苏州中科全象智能科技有限公司 | Capacitance appearance detection device and method based on artificial intelligence technology |
| CN113065520A (en) * | 2021-04-25 | 2021-07-02 | 江南大学 | Multi-modal data-oriented remote sensing image classification method |
| CN113139520A (en) * | 2021-05-14 | 2021-07-20 | 杭州旭颜科技有限公司 | Equipment diaphragm performance monitoring method for industrial Internet |
| CN113538417A (en) * | 2021-08-24 | 2021-10-22 | 安徽顺鼎阿泰克科技有限公司 | Transparent container defect detection method and device based on multi-angle and target detection |
| CN114491017A (en) * | 2021-12-22 | 2022-05-13 | 湖南大学 | Deep learning-based software requirements text defect detection method and system |
| WO2022120665A1 (en) * | 2020-12-09 | 2022-06-16 | 电子科技大学 | Capacitance defect intelligent detection method based on deep learning |
| CN114994046A (en) * | 2022-04-19 | 2022-09-02 | 深圳格芯集成电路装备有限公司 | Defect detection system based on deep learning model |
| CN116580030A (en) * | 2023-07-13 | 2023-08-11 | 厦门微图软件科技有限公司 | Welding quality anomaly detection method based on anomaly simulation |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106875381A (en) * | 2017-01-17 | 2017-06-20 | 同济大学 | A kind of phone housing defect inspection method based on deep learning |
| CN106952250A (en) * | 2017-02-28 | 2017-07-14 | 北京科技大学 | A metal strip surface defect detection method and device based on Faster R-CNN network |
| US20170323163A1 (en) * | 2016-05-06 | 2017-11-09 | City Of Long Beach | Sewer pipe inspection and diagnostic system and method |
| CN107833220A (en) * | 2017-11-28 | 2018-03-23 | 河海大学常州校区 | Fabric defect detection method based on depth convolutional neural networks and vision significance |
-
2018
- 2018-08-18 CN CN201810944235.4A patent/CN109242830A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170323163A1 (en) * | 2016-05-06 | 2017-11-09 | City Of Long Beach | Sewer pipe inspection and diagnostic system and method |
| CN106875381A (en) * | 2017-01-17 | 2017-06-20 | 同济大学 | A kind of phone housing defect inspection method based on deep learning |
| CN106952250A (en) * | 2017-02-28 | 2017-07-14 | 北京科技大学 | A metal strip surface defect detection method and device based on Faster R-CNN network |
| CN107833220A (en) * | 2017-11-28 | 2018-03-23 | 河海大学常州校区 | Fabric defect detection method based on depth convolutional neural networks and vision significance |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110223266A (en) * | 2019-03-08 | 2019-09-10 | 湖南工业大学 | A kind of Railway wheelset tread damage method for diagnosing faults based on depth convolutional neural networks |
| CN110335239A (en) * | 2019-05-09 | 2019-10-15 | 菲特(天津)智能科技有限公司 | Defects detection training machine and its application method based on deep learning |
| CN110530875A (en) * | 2019-08-29 | 2019-12-03 | 珠海博达创意科技有限公司 | A kind of FPCB open defect automatic detection algorithm based on deep learning |
| CN111982910A (en) * | 2020-07-06 | 2020-11-24 | 华南理工大学 | Weak supervision machine vision detection method and system based on artificial defect simulation |
| CN112083002A (en) * | 2020-08-26 | 2020-12-15 | 苏州中科全象智能科技有限公司 | Capacitance appearance detection device and method based on artificial intelligence technology |
| WO2022120665A1 (en) * | 2020-12-09 | 2022-06-16 | 电子科技大学 | Capacitance defect intelligent detection method based on deep learning |
| CN113065520A (en) * | 2021-04-25 | 2021-07-02 | 江南大学 | Multi-modal data-oriented remote sensing image classification method |
| CN113139520A (en) * | 2021-05-14 | 2021-07-20 | 杭州旭颜科技有限公司 | Equipment diaphragm performance monitoring method for industrial Internet |
| CN113139520B (en) * | 2021-05-14 | 2022-07-29 | 江苏中天互联科技有限公司 | Equipment diaphragm performance monitoring method for industrial Internet |
| CN113538417A (en) * | 2021-08-24 | 2021-10-22 | 安徽顺鼎阿泰克科技有限公司 | Transparent container defect detection method and device based on multi-angle and target detection |
| CN114491017A (en) * | 2021-12-22 | 2022-05-13 | 湖南大学 | Deep learning-based software requirements text defect detection method and system |
| CN114994046A (en) * | 2022-04-19 | 2022-09-02 | 深圳格芯集成电路装备有限公司 | Defect detection system based on deep learning model |
| CN116580030A (en) * | 2023-07-13 | 2023-08-11 | 厦门微图软件科技有限公司 | Welding quality anomaly detection method based on anomaly simulation |
| CN116580030B (en) * | 2023-07-13 | 2023-10-20 | 厦门微图软件科技有限公司 | Welding quality anomaly detection method based on anomaly simulation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109242830A (en) | A kind of machine vision technique detection method based on deep learning | |
| CN107133616B (en) | A segmentation-free character localization and recognition method based on deep learning | |
| CN109784204B (en) | Method for identifying and extracting main fruit stalks of stacked cluster fruits for parallel robot | |
| CN106971152A (en) | A kind of method of Bird's Nest in detection transmission line of electricity based on Aerial Images | |
| CN105930815B (en) | A kind of underwater biological detection method and system | |
| CN106951870B (en) | Intelligent detection and early warning method for active visual attention of significant events of surveillance video | |
| Dandavate et al. | CNN and data augmentation based fruit classification model | |
| CN107316058A (en) | Improve the method for target detection performance by improving target classification and positional accuracy | |
| CN107341506A (en) | A kind of Image emotional semantic classification method based on the expression of many-sided deep learning | |
| CN109389599A (en) | A kind of defect inspection method and device based on deep learning | |
| CN108564103A (en) | Data processing method and device | |
| CN115035082B (en) | A defect detection method for aircraft transparent parts based on YOLOv4 improved algorithm | |
| CN107862692A (en) | A kind of ribbon mark of break defect inspection method based on convolutional neural networks | |
| Bjørlykhaug et al. | Vision system for quality assessment of robotic cleaning of fish processing plants using CNN | |
| CN105389581A (en) | Germinated rice germ integrity intelligent identification system and identification method thereof | |
| CN114419372B (en) | Multi-scale point cloud classification method and system | |
| CN115564031A (en) | Detection network for glass defect detection | |
| CN111897333B (en) | Robot walking path planning method | |
| CN107742132A (en) | Potato Surface Defect Detection Method Based on Convolutional Neural Network | |
| Han et al. | Tomatoes maturity detection approach based on YOLOv5 and attention mechanisms | |
| Ramadhan et al. | Identification of cavendish banana maturity using convolutional neural networks | |
| CN113343773A (en) | Facial expression recognition system based on shallow convolutional neural network | |
| Dewi et al. | Automated fruit classification based on deep learning utilizing Yolov8 | |
| CN108986090A (en) | A kind of depth convolutional neural networks method applied to the detection of cabinet surface scratch | |
| CN110555401B (en) | An adaptive emotion expression system and method based on facial expression recognition |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190118 |
|
| WD01 | Invention patent application deemed withdrawn after publication |