CN108491867A - Image Matching based on artificial intelligence and recognition methods - Google Patents
Image Matching based on artificial intelligence and recognition methods Download PDFInfo
- Publication number
- CN108491867A CN108491867A CN201810198392.5A CN201810198392A CN108491867A CN 108491867 A CN108491867 A CN 108491867A CN 201810198392 A CN201810198392 A CN 201810198392A CN 108491867 A CN108491867 A CN 108491867A
- Authority
- CN
- China
- Prior art keywords
- image
- artificial intelligence
- product
- recognition methods
- image matching
- 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
Classifications
-
- 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
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to a kind of Image Matching based on artificial intelligence and recognition methods comprising following steps：Product model and original copy are collected, product image is obtained；Establish artificial intelligence image recognition database；Classified to data using supervised learning method, learnt；Carry out image recognition；Problem after being identified for image is counted, is overhauled.Thus, it is possible to eliminate Artifact to industrial products high degree, production efficiency is improved.It can classify to defect, classify to the rank of defect, convenient for investigation.The defects of specific region can be tracked, be conducive to product problem analysis, promote product processing quality.Maximum manpower-free's autonomous learning, improves artificial intelligence process efficiency.Have a wide range of application, may be used in most of field of industrial production.
Description
Technical field
The present invention relates to a kind of application of artificial intelligence more particularly to a kind of Image Matchings and identification based on artificial intelligence
Method.
Background technology
Currently, robot production is widely used in industrial production, but the quality control in production process still deposit it is big larger
The problem of, although being used in the means such as automatic optics inspection, efficiency and result cannot meet the needs produced greatly.
Specifically, manual confirmation is still needed for the defect of most of product, it is necessary for the statistic of classification of defect
It is carried out by artificial, and production process cannot be timely fed back to, especially the pith of product, subregion of having no idea
It detects simultaneously, has severely impacted product quality, improve the cost of production.
It is a kind of based on artificial intelligence to found in view of the above shortcomings, the designer, is actively subject to research and innovation
Image Matching and recognition methods make it with more the utility value in industry.
Invention content
In order to solve the above technical problems, the object of the present invention is to provide a kind of Image Matching based on artificial intelligence and identifications
Method.
The Image Matching based on artificial intelligence of the present invention and recognition methods comprising following steps：Step 1 collects production
Product model and original copy obtain product image.Step 2 establishes artificial intelligence image recognition database.Step 3 is learned using supervision
Habit method is classified to data, is learnt.Step 4 carries out image recognition.Step 5, the problem after being identified for image are united
Meter, maintenance.
Further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 1, production
The minimum pixel of product model and original copy is more than or equal to 10UM, and product image is obtained by line scan camera or 3D cameras,
Noise reduction process is carried out by adaptive grayscale value to product shadow.
Further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 2,
To having marked the training sample of result, prediction result is exported by model, passes through optimization so that the mistake of predicted value and actual value
Difference minimizes, and enables model that can classify to unknown data.
Further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 3,
Supervised learning method is established using convolutional neural networks, labeled data is trained,
The convolutional calculation of convolution kernel is crossed by S1, convolutional Neural Netcom, and image data is evolved into the data of convolutional layer；
S2 uses Relu functions by active coating, carries out nonlinear transformation；
S3, access pond layer reduce data scale；
S4, S1 constitute systemic circulation to S3, form depth network；
S5 accesses the recirculating network of full articulamentum and active coating composition；
S6, connection softmax functions calculate loss function value.
Further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 3
Learning process is, by BP algorithm, with chain method, then to the gradient of the parameter of each layer in neural network carry out transmit and more
Newly；
Parameter is learnt using SGD algorithms, adam algorithms, annealing algorithm；
Using Dropout methods, in the training process of every batch of, the neuron of each hidden layer of neural network is random loses
State living, it is random to retain another part neuron.
Further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 3
In learning process,
Classify according to product quality, categorical attribute measured by size and position,
Visualization quality class set is established according to the professional standard and Customer Standard of product.
Further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 4,
Image is annotated using training machine, is classified, product quality category set is established, in conjunction with database,
S1 selects detection zone to detect image and the inconsistent local conduct of original copy using product Image Matching original copy
Candidate imagery；Matching uses two methods, and one is graph outline matching, another kind is graphics skeleton matching；
S2 detects Image Matching inconsistent candidate imagery, carries out artificial intelligence image recognition.
Further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 5,
It according to the requirement of setting, carries out statistic of classification and uploads in real time, instruct industrial production, after being identified by artificial intelligence, machine is certainly
Dynamic type, the rank for distinguishing defect, and defect Producing reason and process are indicated according to label substance, machine by statistical number factually
Shi Shangchuan searches problem in time for administrative staff.
Still further, above-mentioned Image Matching and recognition methods based on artificial intelligence, wherein in the step 5,
If there are recoverable flaw, repaired automatically using artificial or full-automatic prosthetic appliance, for the recoverable defect in part, then
After inputting dependent instruction, whether manual confirmation has been repaired completely again.
According to the above aspect of the present invention, the present invention has at least the following advantages：
1, Artifact can be eliminated to industrial products high degree, improves production efficiency.
2, it can classify to defect, classify to the rank of defect, convenient for investigation.
3, the defects of specific region can be tracked, is conducive to product problem analysis, promotes product processing quality.
4, maximum manpower-free's autonomous learning, improves artificial intelligence process efficiency.
5, have a wide range of application, may be used in most of field of industrial production.
Above description is only the general introduction of technical solution of the present invention, in order to better understand the technical means of the present invention,
And can be implemented in accordance with the contents of the specification, it is described in detail as after with presently preferred embodiments of the present invention below.
Specific implementation mode
With reference to embodiment, the embodiment of the present invention is furthur described in detail.Following embodiment is used for
Illustrate the present invention, but is not limited to the scope of the present invention.
Image Matching based on artificial intelligence and recognition methods comprising following steps：
First, product model and original copy are collected, product image is obtained.In order to obtain preferable product image, convenient for follow-up
Artificial intelligence is learnt and is judged, the minimum pixel of product model and original copy is more than or equal to 10UM.It can control minimum image as a result,
The size of element is less than the 1/2 of minimum defect.It is convenient in order to implement, it can be completed by CAD, the various softwares such as ODB++, CAMDATA.
Meanwhile product image can be obtained by line scan camera or 3D cameras, product shadow is carried out by adaptive grayscale value
Noise reduction process.Specifically, for different industrial products, used camera is different.Such as PCB, generally using 8K or
16K line scan cameras.Product appearance uses 130,000,000 photographing cameras.The irregular product such as curve screen uses 3D cameras.
Then, artificial intelligence image recognition database is established.Classified to data using supervised learning method, learnt.It is right
The training sample for having marked result exports prediction result by model, passes through optimization so that the error of predicted value and actual value
It minimizes, enables model that can classify to unknown data.
A little effects are preferably learned in order to possess, from the point of view of a preferable embodiment of the invention, using convolutional Neural net
Network establishes supervised learning method, is trained to labeled data.Detailed process is as follows：
The convolutional calculation of convolution kernel is crossed by S1, convolutional Neural Netcom, and image data is evolved into the data of convolutional layer.
S2 uses Relu functions by active coating, carries out nonlinear transformation.
S3, access pond layer reduce data scale.When actual implementation, step S1, step S2 can be followed on demand
Ring.
S4, S1 constitute systemic circulation to S3, form depth network.
S5 accesses the recirculating network of full articulamentum and active coating composition.
S6, connection softmax functions calculate loss function value.
The present invention uses learning process for by BP algorithm, with chain method, then to the parameter of each layer in neural network
Gradient transmitted and updated.Meanwhile using SGD algorithms, adam algorithms, annealing algorithm scheduling algorithm to parameter
It practises.Since adam methods combine two kinds of optimization algorithms of Momentum and RMSprop, trained efficiency can be significantly improved.And
And Dropout methods can also be used, in the training process of every batch of, the neuron of each hidden layer of neural network is
Random inactivated state has certain probability not use the neuron, random to retain another part neuron.This method can
The problem of effective solution model overfitting.Deep learning can effectively be solved by being added BatchNorm layers before Relu layers
In common gradient disperse and gradient explosion issues.
In entire learning process, it can classify according to product quality, the reason is that, having relatively per class industrial products
The international standard and professional standard answered, machine learning Main Basiss above-mentioned standard.Meanwhile it can belong to by size and position metering classification
Property, establish visualization quality class set according to the professional standard and Customer Standard of product.Such as PCB professional standard IPC60, semiconductor
Encapsulation standard JEDEC, JEITA, SEMI G9-86, SEMI G10-86 etc..
Later, image recognition is carried out.During this period, in order to improve discrimination, training machine can be used, image is noted
It releases, classify, product quality category set is established, in conjunction with database.Specifically：First, detection zone is selected, product image is utilized
Original copy is matched, detects image with the inconsistent place of original copy as candidate imagery；Matching uses two methods, and one is figures
Outline, another kind are graphics skeleton matchings.Later, inconsistent candidate imagery is detected for Image Matching, is carried out artificial
Intelligent imaging recognition.In other words, it can be based on Feature, template matching modes carry out image analysing computer.
Problem after being identified finally, for image is counted, is overhauled.During this period, according to the requirement of setting, divided
Class is counted and is uploaded in real time, instructs industrial production, after being identified by artificial intelligence, type, the rank of machine automatic distinguishing defect,
And defect Producing reason and process are indicated according to label substance, machine uploads statistical data in real time, timely for administrative staff
It searches problem.Prevent lot size problem product from occurring as a result,.Meanwhile if there are recoverable flaw, using artificial or full-automatic
Prosthetic appliance is repaired automatically, and for the recoverable defect in part, then after inputting dependent instruction, completely whether manual confirmation again
It repairs.
In brief, implementation process of the invention is as follows：The product model and original copy of product to be checked are inputted first.From image
Acquisition terminal obtains the image of product to be checked.First time image analysis is carried out according to the feature of product.Later, according to product quality
The influence that standard is set up carries out similarity screening as database, to the result of first time image analysis, while also carrying out defect
And problem Producing reason is classified, and position mark is carried out to the defects of specific region.Finally, it is compiled according to image recognition
The defect of code retrieval specific region.
It can be seen that after applying the present invention by above-mentioned character express, gather around and have the following advantages：
1, Artifact can be eliminated to industrial products high degree, improves production efficiency.
2, it can classify to defect, classify to the rank of defect, convenient for investigation.
3, the defects of specific region can be tracked, is conducive to product problem analysis, promotes product processing quality.
4, maximum manpower-free's autonomous learning, improves artificial intelligence process efficiency.
5, have a wide range of application, may be used in most of field of industrial production.
The above is only a preferred embodiment of the present invention, it is not intended to restrict the invention, it is noted that for this skill
For the those of ordinary skill in art field, without departing from the technical principles of the invention, can also make it is several improvement and
Modification, these improvements and modifications also should be regarded as protection scope of the present invention.
Claims (9)
1. the Image Matching based on artificial intelligence and recognition methods, it is characterised in that include the following steps：
Step 1 collects product model and original copy, obtains product image；
Step 2 establishes artificial intelligence image recognition database；
Step 3 is classified to data using supervised learning method, is learnt；
Step 4 carries out image recognition；
Step 5, the problem after being identified for image are counted, are overhauled.
2. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
In one, the minimum pixel of product model and original copy is more than or equal to 10UM, and production is obtained by line scan camera or 3D cameras
Product image carries out noise reduction process to product shadow by adaptive grayscale value.
3. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
In two, to having marked the training sample of result, by model export prediction result, pass through optimization so that predicted value and really
Value minimizes the error, and enables model that can classify to unknown data.
4. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
In three, supervised learning method is established using convolutional neural networks, labeled data is trained,
The convolutional calculation of convolution kernel is crossed by S1, convolutional Neural Netcom, and image data is evolved into the data of convolutional layer；
S2 uses Relu functions by active coating, carries out nonlinear transformation；
S3, access pond layer reduce data scale；
S4, S1 constitute systemic circulation to S3, form depth network；
S5 accesses the recirculating network of full articulamentum and active coating composition；
S6, connection softmax functions calculate loss function value.
5. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
Learning process in three is, by BP algorithm, with chain method, is then passed to the gradient of the parameter of each layer in neural network
It passs and updates；
Parameter is learnt using SGD algorithms, adam algorithms, annealing algorithm；
Using Dropout methods, in the training process of every batch of, the neuron of each hidden layer of neural network is random inactivation shape
State, it is random to retain another part neuron.
6. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
In learning process in three,
Classify according to product quality, categorical attribute measured by size and position,
Visualization quality class set is established according to the professional standard and Customer Standard of product.
7. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
In four, image is annotated using training machine, is classified, product quality category set is established, in conjunction with database,
S1 selects detection zone, using product Image Matching original copy, detects the image place inconsistent with original copy as candidate
Image；Matching uses two methods, and one is graph outline matching, another kind is graphics skeleton matching；
S2 detects Image Matching inconsistent candidate imagery, carries out artificial intelligence image recognition.
8. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
In five, according to the requirement of setting, carries out statistic of classification and upload in real time, instruct industrial production, after being identified by artificial intelligence, machine
Type, the rank of device automatic distinguishing defect, and indicate defect Producing reason and process according to label substance, machine is by statistical number
It uploads when factually, searches problem in time for administrative staff.
9. Image Matching and recognition methods according to claim 1 based on artificial intelligence, it is characterised in that：The step
It in five, if there are recoverable flaw, is repaired using artificial or full-automatic prosthetic appliance, is lacked for part is recoverable automatically
It falls into, then after inputting dependent instruction, whether manual confirmation has been repaired completely again.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810198392.5A CN108491867A (en) | 2018-03-12 | 2018-03-12 | Image Matching based on artificial intelligence and recognition methods |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810198392.5A CN108491867A (en) | 2018-03-12 | 2018-03-12 | Image Matching based on artificial intelligence and recognition methods |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108491867A true CN108491867A (en) | 2018-09-04 |
Family
ID=63338274
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810198392.5A Pending CN108491867A (en) | 2018-03-12 | 2018-03-12 | Image Matching based on artificial intelligence and recognition methods |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108491867A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109934814A (en) * | 2019-03-15 | 2019-06-25 | 英业达科技有限公司 | Surface defect detecting system and its method |
| CN109977808A (en) * | 2019-03-11 | 2019-07-05 | 北京工业大学 | A kind of wafer surface defects mode detection and analysis method |
| CN110290203A (en) * | 2019-06-25 | 2019-09-27 | 东莞德福得精密五金制品有限公司 | Artificial intelligence cloud computing carries out Noninvasive production defect recognition and information communication to manufactured goods |
| CN110473170A (en) * | 2019-07-10 | 2019-11-19 | 苏州卓融新能源科技有限公司 | A kind of artificial intelligence detection method suitable for the true and false determining defects of pcb board |
| CN110544247A (en) * | 2019-09-03 | 2019-12-06 | 东莞德福得精密五金制品有限公司 | method for inspecting and monitoring operation defects of artificial intelligent cloud computing multi-path equipment system |
| CN110930350A (en) * | 2018-09-19 | 2020-03-27 | 鸿富锦精密电子(天津)有限公司 | Machine learning method and automatic optical detection equipment applying same |
| CN113177626A (en) * | 2021-06-29 | 2021-07-27 | 中国科学院自动化研究所 | Data-driven automatic learning method, system and equipment for complex system mechanism |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104850858A (en) * | 2015-05-15 | 2015-08-19 | 华中科技大学 | Injection-molded product defect detection and recognition method |
| CN105320945A (en) * | 2015-10-30 | 2016-02-10 | 小米科技有限责任公司 | Image classification method and apparatus |
| WO2017113232A1 (en) * | 2015-12-30 | 2017-07-06 | 中国科学院深圳先进技术研究院 | Product classification method and apparatus based on deep learning |
| CN107392896A (en) * | 2017-07-14 | 2017-11-24 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of Wood Defects Testing method and system based on deep learning |
| CN107481231A (en) * | 2017-08-17 | 2017-12-15 | 广东工业大学 | A kind of handware defect classifying identification method based on depth convolutional neural networks |
| CN107563999A (en) * | 2017-09-05 | 2018-01-09 | 华中科技大学 | A kind of chip defect recognition methods based on convolutional neural networks |
| CN107607554A (en) * | 2017-09-26 | 2018-01-19 | 天津工业大学 | A kind of Defect Detection and sorting technique of the zinc-plated stamping parts based on full convolutional neural networks |
-
2018
- 2018-03-12 CN CN201810198392.5A patent/CN108491867A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104850858A (en) * | 2015-05-15 | 2015-08-19 | 华中科技大学 | Injection-molded product defect detection and recognition method |
| CN105320945A (en) * | 2015-10-30 | 2016-02-10 | 小米科技有限责任公司 | Image classification method and apparatus |
| WO2017113232A1 (en) * | 2015-12-30 | 2017-07-06 | 中国科学院深圳先进技术研究院 | Product classification method and apparatus based on deep learning |
| CN107392896A (en) * | 2017-07-14 | 2017-11-24 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of Wood Defects Testing method and system based on deep learning |
| CN107481231A (en) * | 2017-08-17 | 2017-12-15 | 广东工业大学 | A kind of handware defect classifying identification method based on depth convolutional neural networks |
| CN107563999A (en) * | 2017-09-05 | 2018-01-09 | 华中科技大学 | A kind of chip defect recognition methods based on convolutional neural networks |
| CN107607554A (en) * | 2017-09-26 | 2018-01-19 | 天津工业大学 | A kind of Defect Detection and sorting technique of the zinc-plated stamping parts based on full convolutional neural networks |
Non-Patent Citations (1)
| Title |
|---|
| 赵轶: "深度学习在硅钢钢片缺陷识别中的应用", 《数字技术与应用》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI715051B (en) * | 2018-09-19 | 2021-01-01 | 鴻齡科技股份有限公司 | Machine learning method and automatic optical inspection device using the method thereof |
| CN110930350A (en) * | 2018-09-19 | 2020-03-27 | 鸿富锦精密电子(天津)有限公司 | Machine learning method and automatic optical detection equipment applying same |
| CN109977808A (en) * | 2019-03-11 | 2019-07-05 | 北京工业大学 | A kind of wafer surface defects mode detection and analysis method |
| CN109977808B (en) * | 2019-03-11 | 2020-10-27 | 北京工业大学 | Wafer surface defect mode detection and analysis method |
| CN109934814A (en) * | 2019-03-15 | 2019-06-25 | 英业达科技有限公司 | Surface defect detecting system and its method |
| CN109934814B (en) * | 2019-03-15 | 2022-02-01 | 英业达科技有限公司 | Surface defect detection system and method thereof |
| CN110290203A (en) * | 2019-06-25 | 2019-09-27 | 东莞德福得精密五金制品有限公司 | Artificial intelligence cloud computing carries out Noninvasive production defect recognition and information communication to manufactured goods |
| CN110473170A (en) * | 2019-07-10 | 2019-11-19 | 苏州卓融新能源科技有限公司 | A kind of artificial intelligence detection method suitable for the true and false determining defects of pcb board |
| CN110473170B (en) * | 2019-07-10 | 2023-04-07 | 苏州卓融新能源科技有限公司 | Artificial intelligence detection method suitable for judging true and false defects of PCB |
| CN110544247A (en) * | 2019-09-03 | 2019-12-06 | 东莞德福得精密五金制品有限公司 | method for inspecting and monitoring operation defects of artificial intelligent cloud computing multi-path equipment system |
| CN110544247B (en) * | 2019-09-03 | 2023-11-07 | 东莞德福得精密五金制品有限公司 | Method for checking and monitoring operation defects of artificial intelligent cloud computing multi-path equipment system |
| CN113177626A (en) * | 2021-06-29 | 2021-07-27 | 中国科学院自动化研究所 | Data-driven automatic learning method, system and equipment for complex system mechanism |
| CN113177626B (en) * | 2021-06-29 | 2021-09-14 | 中国科学院自动化研究所 | Data-driven automatic learning method, system and equipment for complex system mechanism |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108491867A (en) | Image Matching based on artificial intelligence and recognition methods | |
| CN111815564B (en) | Method and device for detecting silk ingots and silk ingot sorting system | |
| CN107451602A (en) | A kind of fruits and vegetables detection method based on deep learning | |
| Fuentes et al. | Coffee fruit recognition using artificial vision and neural networks | |
| CN105654469A (en) | Infant stool color automatic analysis method and system | |
| CN113077450B (en) | Cherry grading detection method and system based on deep convolutional neural network | |
| CN107886095A (en) | A kind of classifying identification method merged based on machine vision and olfactory characteristic | |
| CN112749675A (en) | Potato disease identification method based on convolutional neural network | |
| Margraf et al. | An evolutionary learning approach to self-configuring image pipelines in the context of carbon fiber fault detection | |
| Tan et al. | Surface defect identification of Citrus based on KF-2D-Renyi and ABC-SVM | |
| CN110935646A (en) | Full-automatic crab grading system based on image recognition | |
| CN117315380A (en) | Deep learning-based pneumonia CT image classification method and system | |
| Septiarini et al. | Oil palm leaf disease detection on natural background using convolutional neural networks | |
| Ramirez et al. | Agroindustrial plant for the classification of hass avocados in real-time with ResNet-18 architecture | |
| CN113033705B (en) | Intelligent judgment and verification method for copper converter blowing slagging period end point based on pattern recognition | |
| Tian et al. | Feature fusion–based preprocessing for steel plate surface defect recognition | |
| Surige et al. | IoT-based monitoring system for oyster mushroom farming | |
| Shukla et al. | A New Approach for Leaf Disease Detection using Multilayered Convolutional Neural Network | |
| CN110322443A (en) | A kind of quality determining method and device of intelligent appliance | |
| Zhang et al. | A novel defect diagnosis method for kyropoulos process-based sapphire growth | |
| Mohanapriya et al. | Identification of Fruit Disease Using Instance Segmentation | |
| Poorni et al. | Detection of rice leaf diseases using convolutional neural network | |
| Lin et al. | Wood polish classification for automated quality inspection based on AI vision | |
| Caguioa et al. | Duck Egg Quality Classification Based on its Shell Visual Property Through Transfer Learning Using ResNet-50 | |
| Paramanandham et al. | Enhancing Disease Detection with Weight Initialization and Residual Connections using LeafNet for Groundnut Leaf Diseases |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180904 |