CN103778445B - Cold-rolled strip steel surface defect reason analyzing method and system - Google Patents
Cold-rolled strip steel surface defect reason analyzing method and system Download PDFInfo
- Publication number
- CN103778445B CN103778445B CN201410026959.2A CN201410026959A CN103778445B CN 103778445 B CN103778445 B CN 103778445B CN 201410026959 A CN201410026959 A CN 201410026959A CN 103778445 B CN103778445 B CN 103778445B
- Authority
- CN
- China
- Prior art keywords
- defect
- unit
- feature
- cause
- defect image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a cold-rolled strip steel surface defect reason analyzing system. The system comprises a real-time defect image input unit, a defect reason matching unit, an input unit, a defect reason output unit and a defect reason characteristic knowledge base based on a learning vector quantization neural network. The real-time defect image input unit, the defect reason output unit and the defect reason characteristic knowledge base are all electrically connected with the defect reason matching unit, and the defect reason characteristic knowledge base, the defect reason output unit and the defect reason matching unit are all electrically connected with the input unit. According to the cold-rolled strip steel surface defect reason analyzing system, a defect reason characteristic library is established through experience-type defect reason knowledge hidden in the brain of an expert, and when real-time image information is input, a defect reason matching result can be displayed in real time. The system is high in operating speed, high in accuracy, high in efficiency, high in real-time performance, high in practicability and simple to operate, greatly improves the production efficiency, and lowers the defect rate and the production cost.
Description
Technical field
The present invention relates to a kind of Cold-strip Steel Surface Causes Analysis method and system.
Background technology
Produce towards lean, intelligent direction development with strip steel, very high requirement is proposed to belt steel product quality, such as
What analyzes belt steel surface defect cause in real time, and guides production to become particularly important.
Cold rolled strip steel production process includes " pickling-steel rolling-annealing-smooth-finishing " etc., because pickling is bad, roll table
Planar defect or roll design is unreasonable etc. causes a lot of surface defects, such as hole, Bian Lang, side are split, are mingled with, bubble etc., serious shadow
Ring product quality.Existing Causes Analysis method is dependence experience worker to the experience rolling in product or finished product mostly
Analysis, has hysteresis quality, and Personnel Dependence is it is impossible to real-time instruction produces and rolls a large amount of waste products;It cannot be taken into account up and down simultaneously
Relation between technique, has one-sidedness.
Steel strip surface defect reason knowledge is to be hidden in the implicit knowledge of expert's brains, and has uncertainty, a kind of scarce
Falling into have many factors to cause, and a kind of factor is likely to result in number of drawbacks again, how to find implicit, effective, valuable
, intelligible factor, obtain uncertain mapping relations between them, and by its structuring, domination, will be all that defect is former
Premise because of real-time analysis.
Defect cause is only mapped with defect image feature, sets up defect cause feature knowledge storehouse and is just easy to lack in real time
Sunken reason retrieval coupling, Image semantic classification, defect Target Segmentation, feature extraction all will be known, set up defect for obtaining characteristics of image
Reason feature database provides technical support.
Content of the invention
It is an object of the invention to provide a kind of Cold-strip Steel Surface Causes Analysis method and system, solve prior art
Present in the problems referred to above.
The technical scheme is that a kind of Cold-strip Steel Surface Causes Analysis side
Method, comprises the following steps:
Step 1: set up the defect cause feature knowledge storehouse based on learning vector quantization neutral net;
Step 2: defect image input block gathers real-time defect image in real time, and described real-time defect image is sent to
Defect cause matching unit;
Step 3: described defect cause matching unit is by described learning vector quantization neutral net in described defect cause
Match the defect cause corresponding with described real-time defect image in feature knowledge storehouse, and described defect cause is sent to scarce
Sunken reason output unit;
Step 4: described defect cause is shown by described defect cause output unit by input block.
Further, described step 1 specifically includes following steps:
Step 1.1: defect cause knowledge is input to by described defect cause feature knowledge storehouse by input block;
Step 1.2: defect image is input to by described defect cause feature knowledge storehouse by input block, and passes through defect
Image characteristics extraction method extracts the defect image characteristic vector of described defect image;
Step 1.3: defect cause features training unit passes through learning vector quantization neutral net and matlab instrument to defeated
The described defect cause knowledge entering and defect image characteristic vector are trained, and obtain described defect cause knowledge and described defect
The mapping relations of image feature vector, and described mapping relations are sent to the shared coding unit of reason feature;
Step 1.4: described reason feature is shared coding unit and described mapping relations are encoded, and coding result is sent out
Deliver to Database Unit;
Step 1.5: described Database Unit stores described coding result.
Further, described step 3 specifically includes following steps:
Step 3.1: defect cause matching unit extracts described real-time defect image by defect image feature extraction
Defect image characteristic vector;
Step 3.2: defect cause matching unit is special in described defect cause by described learning vector quantization neutral net
Levy and in knowledge base, match the defect cause corresponding with the defect image characteristic vector of described real-time defect image, and will be described
Defect cause is sent to defect cause output unit.
A kind of Cold-strip Steel Surface Causes Analysis system, including real-time defect image input block, defect cause
Join unit, input block, defect cause output unit and the defect cause feature knowledge based on learning vector quantization neutral net
Storehouse, described real-time defect image input block, defect cause output unit and defect cause feature knowledge storehouse all with described defect
Reason matching unit electrically connects, described defect cause feature knowledge storehouse, defect cause output unit and described defect cause coupling
Unit is all electrically connected with described input block;
Described real-time defect image input block is used for gathering real-time defect image, and described real-time defect image is sent
To defect cause matching unit;
Described defect cause matching unit is used for knowing in described defect cause feature by learning vector quantization neutral net
Know in storehouse and match the defect cause corresponding with described real-time defect image, and described defect cause is sent to defect cause
Output unit;
Described defect cause output unit is used for being shown described defect cause by described input block.
Further, described defect cause feature knowledge storehouse includes defect image feature extraction unit, defect cause feature instruction
Practice unit, reason feature shares coding unit database unit;
Described input block is additionally operable to for defect cause knowledge and defect image to be input to described defect cause feature knowledge
Storehouse;
Described defect image feature extraction unit is used for extracting described defect image by defect image feature extraction
Defect image characteristic vector, and described defect image characteristic vector is sent to described defect cause features training unit;
Described defect cause features training unit is used for by learning vector quantization neutral net and matlab instrument to defeated
The described defect cause knowledge entering and defect image characteristic vector are trained, and obtain described defect cause knowledge and described defect
The mapping relations of image feature vector, and described mapping relations are sent to the shared coding unit of described reason feature;
Described reason feature is shared coding unit and is used for described mapping relations are encoded, and coding result is sent to
Database Unit;
Described Database Unit is used for storing described coding result.
Further, described defect cause matching unit includes defect image characteristic extracting module and defect cause coupling mould
Block, described defect image characteristic extracting module is used for extracting described real-time defect image by defect image feature extraction
Defect image characteristic vector;Described defect cause matching module is used for lacking described by described learning vector quantization neutral net
The defect cause corresponding with the defect image characteristic vector of described real-time defect image is matched in sunken reason feature knowledge storehouse,
And described defect cause is sent to defect cause output unit.
Further, described defect image feature extraction specifically includes following steps:
Step 1: image pre-processing unit is smoothed and greyscale transformation successively to defect image, described greyscale transformation
Strengthen algorithm stretching contrast including time domain, frequency domain, to improve picture quality, and the defect image after processing is sent to defect
Object segmentation unit;
Step 2: described defect object segmentation unit enters rower to the region related to defect cause on described defect image
Note, and the described defect image after processing is sent to feature selection unit;
Step 3: feature selection unit chooses the characteristic vector related to defect cause, described characteristic vector include average,
Population variance, x direction variance, y direction variance, energy, entropy, x direction zero-crossing values and/or y direction zero-crossing values, and will process after lack
Sunken image sends to Feature Dimension Reduction unit;
Step 4: Feature Dimension Reduction unit removes the characteristic vector comprising redundancy, and by the defect image after processing to spy
Levy extraction unit to send;
Step 5: feature extraction unit extracts final defect image characteristic vector.
The invention has the beneficial effects as follows: a kind of present invention Cold-strip Steel Surface Causes Analysis method and system are passed through hidden
The empirical defect cause knowledge being hidden in expert's brains sets up defect cause feature database, when inputting real-time image information, energy
Display defect reason matching result in real time.System running speed is fast, realizes real-time defect image reason intelligent Matching, and intelligence is defeated
Go out matching result, and high precision, efficiency high, real-time, practical, simple to operate, substantially increase production efficiency, reduce
Ratio of defects and production cost, compared with prior art, have that the analysis of causes is real-time, high precision, easy to operate and produce effect
The high advantage of rate.
Brief description
Fig. 1 is a kind of structured flowchart of present invention Cold-strip Steel Surface Causes Analysis system;
Fig. 2 is a kind of structure in the defect cause feature knowledge storehouse of present invention Cold-strip Steel Surface Causes Analysis system
Block diagram.
Specific embodiment
Below in conjunction with accompanying drawing, the principle of the present invention and feature are described, example is served only for explaining the present invention, and
Non- for limiting the scope of the present invention.
A kind of Cold-strip Steel Surface Causes Analysis method, comprises the following steps:
Step 1: set up the defect cause feature knowledge storehouse based on learning vector quantization neutral net;
Step 2: defect image input block gathers real-time defect image in real time, and described real-time defect image is sent to
Defect cause matching unit;
Step 3: described defect cause matching unit is by described learning vector quantization neutral net in described defect cause
Match the defect cause corresponding with described real-time defect image in feature knowledge storehouse, and described defect cause is sent to scarce
Sunken reason output unit;
Step 4: described defect cause is shown by described defect cause output unit by input block.
Described step 1 specifically includes following steps:
Step 1.1: defect cause knowledge is input to by described defect cause feature knowledge storehouse by input block;
Step 1.2: defect image is input to by described defect cause feature knowledge storehouse by input block, and passes through defect
Image characteristics extraction method extracts the defect image characteristic vector of described defect image;
Step 1.3: defect cause features training unit passes through learning vector quantization neutral net and matlab instrument to defeated
The described defect cause knowledge entering and defect image characteristic vector are trained, and obtain described defect cause knowledge and described defect
The mapping relations of image feature vector, and described mapping relations are sent to the shared coding unit of reason feature;
Step 1.4: described reason feature is shared coding unit and described mapping relations are encoded, and coding result is sent out
Deliver to Database Unit;
Step 1.5: described Database Unit stores described coding result.
Described step 3 specifically includes following steps:
Step 3.1: defect cause matching unit extracts described real-time defect image by defect image feature extraction
Defect image characteristic vector;
Step 3.2: defect cause matching unit is special in described defect cause by described learning vector quantization neutral net
Levy and in knowledge base, match the defect cause corresponding with the defect image characteristic vector of described real-time defect image, and will be described
Defect cause is sent to defect cause output unit;
A kind of Cold-strip Steel Surface Causes Analysis system, including real-time defect image input block, defect cause
Join unit, input block, defect cause output unit and the defect cause feature knowledge based on learning vector quantization neutral net
Storehouse, described real-time defect image input block, defect cause output unit and defect cause feature knowledge storehouse all with described defect
Reason matching unit electrically connects, described defect cause feature knowledge storehouse, defect cause output unit and described defect cause coupling
Unit is all electrically connected with described input block;
Described real-time defect image input block is used for gathering real-time defect image, and described real-time defect image is sent
To defect cause matching unit;
Described defect cause matching unit is used for knowing in described defect cause feature by learning vector quantization neutral net
Know in storehouse and match the defect cause corresponding with described real-time defect image, and described defect cause is sent to defect cause
Output unit;
Described defect cause output unit is used for being shown described defect cause by described input block.
Described defect cause feature knowledge storehouse include defect image feature extraction unit, defect cause features training unit,
Reason feature shares coding unit database unit;
Described input block is additionally operable to for defect cause knowledge and defect image to be input to described defect cause feature knowledge
Storehouse;
Described defect image feature extraction unit is used for extracting described defect image by defect image feature extraction
Defect image characteristic vector, and described defect image characteristic vector is sent to described defect cause features training unit;
Described defect cause features training unit is used for by learning vector quantization neutral net and matlab instrument to defeated
The described defect cause knowledge entering and defect image characteristic vector are trained, and obtain described defect cause knowledge and described defect
The mapping relations of image feature vector, and described mapping relations are sent to the shared coding unit of described reason feature;
Described reason feature is shared coding unit and is used for described mapping relations are encoded, and coding result is sent to
Database Unit;
Described Database Unit is used for storing described coding result.
Described defect cause matching unit includes defect image characteristic extracting module and defect cause matching module, described scarce
Sunken image characteristics extraction module is used for extracting the defect image of described real-time defect image by defect image feature extraction
Characteristic vector;Described defect cause matching module is used for special in described defect cause by described learning vector quantization neutral net
Levy and in knowledge base, match the defect cause corresponding with the defect image characteristic vector of described real-time defect image, and will be described
Defect cause is sent to defect cause output unit.
Described defect image feature extraction specifically includes following steps:
Step 1: image pre-processing unit is smoothed and greyscale transformation successively to defect image, described greyscale transformation
Strengthen algorithm stretching contrast including time domain, frequency domain, to improve picture quality, and the defect image after processing is sent to defect
Object segmentation unit;The purpose of greyscale transformation is for preferably binaryzation;It is preferably, carried out at denoising by morphological method
Reason, the method has preferable inhibition to Gaussian noise, salt-pepper noise, and the speed of service is fast;
Step 2: described defect object segmentation unit adopts adaptivenon-uniform sampling algorithm to former with defect on described defect image
Because related region is labeled, and the described defect image after processing is sent to feature selection unit;
Step 3: feature selection unit utilizes the adaptivity of EVOLUTIONARY COMPUTATION to choose the characteristic vector related to defect cause,
Described characteristic vector includes average, population variance, x direction variance, y direction variance, energy, entropy, x direction zero-crossing values and/or y direction
Zero-crossing values, and the defect image after processing is sent to Feature Dimension Reduction unit;
Step 4: Feature Dimension Reduction unit removes the characteristic vector comprising redundancy, and by the defect image after processing to spy
Levy extraction unit to send;
Step 5: feature extraction unit extracts final defect image characteristic vector.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all spirit in the present invention and
Within principle, any modification, equivalent substitution and improvement made etc., should be included within the scope of the present invention.
Claims (4)
1. a kind of Cold-strip Steel Surface Causes Analysis method is it is characterised in that comprise the following steps:
Step 1: set up the defect cause feature knowledge storehouse based on learning vector quantization neutral net;
Step 2: defect image input block gathers real-time defect image in real time, and described real-time defect image is sent to defect
Reason matching unit;
Step 3: described defect cause matching unit extracts described real-time defect image by defect image feature extraction
Defect image characteristic vector, is then matched in described defect cause feature knowledge storehouse by learning vector quantization neutral net
The defect cause corresponding with the defect image characteristic vector of described real-time defect image, and described defect cause is sent to scarce
Sunken reason output unit;
Step 4: described defect cause is shown by described defect cause output unit by input block;
Described step 1 specifically includes following steps:
Step 1.1: defect cause knowledge is input to by described defect cause feature knowledge storehouse by input block;
Step 1.2: defect image is input to by described defect cause feature knowledge storehouse by input block, and passes through defect image
Feature extraction extracts the defect image characteristic vector of described defect image;
Step 1.3: defect cause features training unit passes through learning vector quantization neutral net and matlab instrument to input
Described defect cause knowledge and defect image characteristic vector are trained, and obtain described defect cause knowledge and described defect image
The mapping relations of characteristic vector, and described mapping relations are sent to the shared coding unit of reason feature;
Step 1.4: described reason feature is shared coding unit and described mapping relations are encoded, and coding result is sent to
Database Unit;
Step 1.5: described Database Unit stores described coding result.
2. a kind of Cold-strip Steel Surface Causes Analysis method according to claim 1 is it is characterised in that described defect
Image characteristics extraction method specifically includes following steps:
Step 1: image pre-processing unit is smoothed and greyscale transformation successively to defect image, described greyscale transformation includes
Time domain, frequency domain strengthen algorithm stretching contrast, to improve picture quality, and the defect image after processing are sent to defect target
Cutting unit;
Step 2: described defect object segmentation unit is labeled to the region related to defect cause on described defect image, and
Described defect image after processing is sent to feature selection unit;
Step 3: feature selection unit chooses the characteristic vector related to defect cause, described characteristic vector includes average, always side
Difference, x direction variance, y direction variance, energy, entropy, x direction zero-crossing values and/or y direction zero-crossing values, and by process after defect map
As sending to Feature Dimension Reduction unit;
Step 4: Feature Dimension Reduction unit removes the characteristic vector comprising redundancy, and the defect image after processing is carried to feature
Unit is taken to send;
Step 5: feature extraction unit extracts final defect image characteristic vector.
3. a kind of Cold-strip Steel Surface Causes Analysis system is it is characterised in that including real-time defect image input block, lacking
Sunken reason matching unit, input block, defect cause output unit and the defect cause based on learning vector quantization neutral net
Feature knowledge storehouse, described real-time defect image input block, defect cause output unit and defect cause feature knowledge storehouse all with
Described defect cause matching unit electrical connection, described defect cause feature knowledge storehouse, defect cause output unit and described defect
Reason matching unit is all electrically connected with described input block;
Described real-time defect image input block is used for gathering real-time defect image, and described real-time defect image is sent to scarce
Sunken reason matching unit;
Described defect cause matching unit includes defect image characteristic extracting module and defect cause matching module, described defect map
As characteristic extracting module is used for extracting the defect image feature of described real-time defect image by defect image feature extraction
Vector;Described defect cause matching module is used for knowing in described defect cause feature by described learning vector quantization neutral net
Know in storehouse and match the defect cause corresponding with the defect image characteristic vector of described real-time defect image, and by described defect
Reason is sent to defect cause output unit;Described defect cause output unit is used for described defect by described input block
Reason is shown;
Described defect cause feature knowledge storehouse includes defect image feature extraction unit, defect cause features training unit, reason
Feature shares coding unit database unit;
Described input block is additionally operable to for defect cause knowledge and defect image to be input to described defect cause feature knowledge storehouse;
Described defect image feature extraction unit is used for extracting lacking of described defect image by defect image feature extraction
Sunken image feature vector, and described defect image characteristic vector is sent to described defect cause features training unit;
Described defect cause features training unit is used for by learning vector quantization neutral net and matlab instrument to input
Described defect cause knowledge and defect image characteristic vector are trained, and obtain described defect cause knowledge and described defect image
The mapping relations of characteristic vector, and described mapping relations are sent to the shared coding unit of described reason feature;
Described reason feature is shared coding unit and is used for described mapping relations are encoded, and coding result is sent to data
Library unit;
Described Database Unit is used for storing described coding result.
4. a kind of Cold-strip Steel Surface Causes Analysis system according to claim 3 is it is characterised in that described defect
Image characteristics extraction method specifically includes following steps:
Step 1: image pre-processing unit is smoothed and greyscale transformation successively to defect image, described greyscale transformation includes
Time domain, frequency domain strengthen algorithm stretching contrast, to improve picture quality, and the defect image after processing are sent to defect target
Cutting unit;
Step 2: described defect object segmentation unit is labeled to the region related to defect cause on described defect image, and
Described defect image after processing is sent to feature selection unit;
Step 3: feature selection unit chooses the characteristic vector related to defect cause, described characteristic vector includes average, always side
Difference, x direction variance, y direction variance, energy, entropy, x direction zero-crossing values and/or y direction zero-crossing values, and by process after defect map
As sending to Feature Dimension Reduction unit;
Step 4: Feature Dimension Reduction unit removes the characteristic vector comprising redundancy, and the defect image after processing is carried to feature
Unit is taken to send;
Step 5: feature extraction unit extracts final defect image characteristic vector.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410026959.2A CN103778445B (en) | 2014-01-21 | 2014-01-21 | Cold-rolled strip steel surface defect reason analyzing method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410026959.2A CN103778445B (en) | 2014-01-21 | 2014-01-21 | Cold-rolled strip steel surface defect reason analyzing method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103778445A CN103778445A (en) | 2014-05-07 |
CN103778445B true CN103778445B (en) | 2017-02-01 |
Family
ID=50570658
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410026959.2A Active CN103778445B (en) | 2014-01-21 | 2014-01-21 | Cold-rolled strip steel surface defect reason analyzing method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103778445B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109692877B (en) * | 2018-12-29 | 2020-11-10 | 中冶南方工程技术有限公司 | Cold-rolled strip steel surface quality management system and method |
CN112775182A (en) * | 2020-12-29 | 2021-05-11 | 广州市荻亚机电设备有限公司 | Manufacturing equipment for producing cold-rolled steel plate for manufacturing high-end equipment |
CN115311350A (en) * | 2022-08-08 | 2022-11-08 | 北京远舢智能科技有限公司 | Method and device for determining position parameters of edge wave defects, electronic equipment and medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1150838A (en) * | 1995-06-12 | 1997-05-28 | 日本钢管株式会社 | Method of determining cause of defects on steel material surface |
CN103286081A (en) * | 2013-05-07 | 2013-09-11 | 浙江工业大学 | Monocular multi-perspective machine vision-based online automatic sorting device for steel ball surface defect |
-
2014
- 2014-01-21 CN CN201410026959.2A patent/CN103778445B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1150838A (en) * | 1995-06-12 | 1997-05-28 | 日本钢管株式会社 | Method of determining cause of defects on steel material surface |
CN103286081A (en) * | 2013-05-07 | 2013-09-11 | 浙江工业大学 | Monocular multi-perspective machine vision-based online automatic sorting device for steel ball surface defect |
Non-Patent Citations (2)
Title |
---|
A Measuring Method of Steel Plate Defect Uncertain Correlation Information based on Rough Set and Functional Dependency;Junwei Liu, Jianyi Kong, Xingdong Wang;《Sensors and Transducers》;20130530;第21卷;第101页倒数第1段,第103页左栏第4段,图1、3 * |
基于机器视觉的高速带钢孔洞检测***;何永辉,王康健,石桂芬;《应用光学》;20070531;第28卷(第3期);第346页倒数第1段,图1 * |
Also Published As
Publication number | Publication date |
---|---|
CN103778445A (en) | 2014-05-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107563385B (en) | License plate character recognition method based on depth convolution production confrontation network | |
CN109829443B (en) | Video behavior identification method based on image enhancement and 3D convolution neural network | |
CN105447851B (en) | The sound hole defect inspection method and system of a kind of glass panel | |
CN105095856B (en) | Face identification method is blocked based on mask | |
CN108629338B (en) | Face beauty prediction method based on LBP and convolutional neural network | |
CN102682287B (en) | Pedestrian detection method based on saliency information | |
CN103854016B (en) | Jointly there is human body behavior classifying identification method and the system of feature based on directivity | |
CN110348357A (en) | A kind of fast target detection method based on depth convolutional neural networks | |
CN104298981A (en) | Face microexpression recognition method | |
CN106980826A (en) | A kind of action identification method based on neutral net | |
CN103729842A (en) | Fabric defect detection method based on local statistical characteristics and overall significance analysis | |
CN105047194A (en) | Self-learning spectrogram feature extraction method for speech emotion recognition | |
CN104143091B (en) | Based on the single sample face recognition method for improving mLBP | |
CN109829924A (en) | A kind of image quality evaluating method based on body feature analysis | |
CN104156729B (en) | A kind of classroom demographic method | |
CN111597983B (en) | Method for realizing identification of generated false face image based on deep convolutional neural network | |
CN103971106A (en) | Multi-view human facial image gender identification method and device | |
CN103778445B (en) | Cold-rolled strip steel surface defect reason analyzing method and system | |
CN104834909B (en) | A kind of new image representation method based on Gabor comprehensive characteristics | |
CN106778529A (en) | A kind of face identification method based on improvement LDP | |
CN109543546A (en) | The gait age estimation method returned based on the distribution of depth sequence | |
CN109509299A (en) | A kind of automatic vending machine with recognition of face | |
CN103185731A (en) | Device for detecting beef tenderness based on color image textural features and method thereof | |
CN116434759B (en) | Speaker identification method based on SRS-CL network | |
CN110071884A (en) | A kind of Modulation Recognition of Communication Signal method based on improvement entropy cloud feature |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |