CN104376566B - Steel strip surface defect image characteristic extracting method based on local feature space length - Google Patents

Steel strip surface defect image characteristic extracting method based on local feature space length Download PDF

Info

Publication number
CN104376566B
CN104376566B CN201410695659.3A CN201410695659A CN104376566B CN 104376566 B CN104376566 B CN 104376566B CN 201410695659 A CN201410695659 A CN 201410695659A CN 104376566 B CN104376566 B CN 104376566B
Authority
CN
China
Prior art keywords
surface defect
data point
defect image
vector data
steel surface
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.)
Expired - Fee Related
Application number
CN201410695659.3A
Other languages
Chinese (zh)
Other versions
CN104376566A (en
Inventor
李波
胡洋
张晓龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University of Science and Engineering WUSE
Original Assignee
Wuhan University of Science and Engineering WUSE
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wuhan University of Science and Engineering WUSE filed Critical Wuhan University of Science and Engineering WUSE
Priority to CN201410695659.3A priority Critical patent/CN104376566B/en
Publication of CN104376566A publication Critical patent/CN104376566A/en
Application granted granted Critical
Publication of CN104376566B publication Critical patent/CN104376566B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
    • G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30168—Image quality inspection

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Multimedia (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
  • Image Processing (AREA)

Abstract

The present invention relates to a kind of steel strip surface defect image characteristic extracting method based on local feature space length.Its technical scheme is：For the pretreated vector data point X of a web steel surface defect imageiSelected and the pretreated vector data point X of a web steel surface defect image from the pretreated vector data point composition matrix data X of all web steel surface defect images respectivelyiK Neighbor Points of classification identical set up manifold local feature space S (Xi), then with manifold local feature space S (Xi) the distance between measurement multiple manifold divergence Js, on the basis of keeping manifold partial structurtes constant, maximize multiple manifold divergence to find low dimension projective matrix A, realize the differentiation feature extraction of steel strip surface defect image.The present invention is by maximizing multiple manifold divergence JsExtract steel strip surface defect image characteristic of division, with improve steel strip surface defect image recognition effect the characteristics of.

Description

Steel strip surface defect image characteristic extracting method based on local feature space length
Technical field
The invention belongs to steel strip surface defect image feature extraction techniques field.It is more particularly to a kind of to be based on local feature The steel strip surface defect image characteristic extracting method of space length.
Background technology
Strip is one of major product form of steel and iron industry, is the indispensable raw material such as Aero-Space, the manufacture of automobile steamer, It is related to the development of many manufacturings.In recent years, the demand of strip is continuously increased, and requires there is higher surface matter Amount.And in its operation of rolling, due to reasons such as continuous casting steel billet, rolling equipment and rolling mill practices, cause rolled plate surface to go out Showed crackle, iron scale, scab, roll marks, scratch, the defect such as hole and pit, these defects not only have impact on product appearance, It is prior to be reduction of the performances such as the corrosion resistance, wear resistence and fatigue strength of product, therefore buried peace to product Full hidden danger.Therefore the surface quality detection of strip is particularly important.
Strip surface quality detection experienced artificial range estimation detection, conventional lossless detection and detected based on image recognition three Individual developing stage.At present, the strip surface quality detection method based on image steganalysis is the focus of research, wherein strip table Planar defect image recognition is the key of Cold-strip Steel Surface defects detection, and steel strip surface defect image characteristics extraction is wherein most For important step, the quality of steel strip surface defect image characteristics extraction directly affects final Detection results.For strip Surface defect image feature extraction and classification, some experts and scholar have carried out more research to it：S.Arivazhagan and Wavelet transformation is applied to Texture classification by L.Ganesan, achieves preferable effect；XuechuanWang and Kuldip K.Paliwal is to the feature extracting method of standard and how to reduce the dimension of characteristic vector and is set forth；The profits such as H.zhang With computer simulation experiment, it was demonstrated that the different characteristic of image can be extracted well using wavelet transform function, image point is carried out Analysis.But it is due to that steel strip surface defect species is various and feature is complicated, the discrimination based on these traditional feature extracting methods It is still not ideal enough.
The content of the invention
It is contemplated that overcome prior art not enough, it is therefore an objective to which providing a kind of can improve steel strip surface defect image recognition The steel strip surface defect image characteristic extracting method based on local feature space length of effect.
To realize above-mentioned technical proposal, the technical solution adopted by the present invention is comprised the concrete steps that：
Step 1: carrying out gray processing processing, smoothing processing, normalized successively to the steel strip surface defect image of collection And vectorization, obtain the pretreated vector data point X of a web steel surface defect imagei, all web steel surface defect maps As pretreated vector data point Xi(i=1,2 ..., n) constitute matrix data X.Wherein：N represents that all belt steel surfaces lack Fall into the sum of image.
Step 2: being found and a web steel surface from the matrix data X after all steel strip surface defect image preprocessings The pretreated vector data point X of defect imageiEuclidean distance is minimum and classification identical k web steel surface defect images are located in advance Vector data point after reason, is constituted and the pretreated vector data point X of a web steel surface defect imageiRelated manifold office Portion feature space S (Xi), minimal linear represents error coefficient WjFor：
In formula (1) and (2)：
XiRepresent the pretreated vector data point of a web steel surface defect image；
XjRepresent and the pretreated vector data point X of a web steel surface defect imageiEuclidean distance minimum and classification phase Same pretreated j-th of vector data point of k web steel surface defect images；
GjtRepresent local Gtram matrix, Gjt=(Xi-Xj)·(Xi-Xt)；
XtRepresent and the pretreated vector data point X of a web steel surface defect imageiEuclidean distance minimum and classification phase Same pretreated t-th of vector data point of k web steel surface defect images；
I, j and t are natural number, i ≠ j ≠ t.
Step 3: for remaining any pretreated vector data point X of web steel surface defect imagei, repeat to walk Rapid two, obtain minimal linear and represent error coefficient matrix W.
Step 4: setting up the mathematical modeling J of multiple manifold divergences
In formula (3)：
S(Xi) represent and the pretreated vector data point X of a web steel surface defect imageiRelated manifold is local special Levy space；
S(Xm) represent and the pretreated vector data point X of another web steel surface defect imagemRelated manifold is local Feature space；
||S(Xi)-S(Xm) | | represent and the pretreated vector data point X of a web steel surface defect imageiRelated Manifold local feature space S (Xi) arrive and the pretreated vector data point X of another web steel surface defect imagemRelated stream Shape local feature space S (Xm) between Euclidean distance Dim；
HimRepresent the pretreated vector data point X of a web steel surface defect imageiWith it is another
The pretreated vector data point X of one web steel surface defect imagemBetween class
Other different information：
Step 5: with the pretreated vector data point X of a web steel surface defect imageiRelated manifold local feature Space S (Xi) arrive and the pretreated vector data point X of another web steel surface defect imagemRelated manifold local feature is empty Between S (Xm) between Euclidean distance DimFor：
In formula (5)：
fi (P)(Xm) represent the pretreated vector data point X of a web steel surface defect imageiWith another web steel table Vector data point X after planar defect image preprocessingmRelated manifold local feature space S (Xm) in projection；
fm (P)(Xi) represent another pretreated vector data point X of web steel surface defect imagemWith a web steel table Vector data point X after planar defect image preprocessingiRelated manifold local feature space S (Xi) in projection.
Step 6: building target function model
s.t ATX(I-W)(I-W)TXTA=I
In formula (6)：
A represents low dimension projective matrix；
T representing matrix transposition；
I represents unit matrix.
Generalized eigenvalue decomposition is carried out to formula (6)
In formula (7)：
λ represents characteristic value；
F represents characteristic vector.
The descending order of eigenvalue λ is arranged, the characteristic vector f composition low-dimensionals before taking corresponding to d characteristic value Projection matrix A, wherein d represent the dimension of extracted feature.
Step 7: for the pretreated vector data point X of a web steel surface defect imagei, after linear projection Low-dimensional characteristic YiFor：
Yi=ATXi (8)
In formula (8)：A represents low dimension projective matrix；
T representing matrix transposition.
The k is the natural number more than 2.
Due to using above-mentioned technical proposal, the beneficial effects of the invention are as follows：
The present invention extracts the feature of steel strip surface defect image using the method based on local feature space length, on the one hand Manifold local feature space S (X is set up using the classification information of vector datai), it is empty with different classes of manifold local feature Between the mathematical modeling J of multiple manifold divergence is set up apart from sums, on the other hand in manifold local feature space S (Xi) interior using local Linear expression represents error coefficient W to calculate minimal linearj, maintain the partial structurtes information of multiple manifold.Because multiple manifold dissipates What degree embodied is the dispersion degree between different classes of data, therefore maximization multiple manifold divergence can find optimal classification sky Between, realize that the low-dimensional of steel strip surface defect image differentiates effective extraction of feature, improve steel strip surface defect image recognition Effect.
Therefore, the present invention is by maximizing multiple manifold divergence JsThe characteristic of division of steel strip surface defect image is extracted, is had The characteristics of improving steel strip surface defect image recognition effect.
Embodiment
With reference to embodiment, the invention will be further described, not to the limitation of its protection domain.
Embodiment 1
A kind of steel strip surface defect image characteristic extracting method based on local feature space length.Side described in the present embodiment Method is comprised the concrete steps that：
Step 1: carrying out gray processing processing, smoothing processing, normalized successively to the steel strip surface defect image of collection And vectorization, obtain the pretreated vector data point X of a web steel surface defect imagei, all web steel surface defect maps As pretreated vector data point Xi(i=1,2 ..., n) constitute matrix data X.Wherein：N represents that all belt steel surfaces lack Fall into the sum of image.
Step 2: being found and a web steel surface from the matrix data X after all steel strip surface defect image preprocessings The pretreated vector data point X of defect imageiEuclidean distance is minimum and classification identical k web steel surface defect images are located in advance Vector data point after reason, is constituted and the pretreated vector data point X of a web steel surface defect imageiRelated manifold office Portion feature space S (Xi), minimal linear represents error coefficient WjFor：
In formula (1) and (2)：
XiRepresent the pretreated vector data point of a web steel surface defect image；
XjRepresent and the pretreated vector data point X of a web steel surface defect imageiEuclidean distance minimum and classification phase Same pretreated j-th of vector data point of k web steel surface defect images；
GjtRepresent local Gtram matrix, Gjt=(Xi-Xj)·(Xi-Xt)；
XtRepresent and the pretreated vector data point X of a web steel surface defect imageiEuclidean distance minimum and classification phase Same pretreated t-th of vector data point of k web steel surface defect images；
I, j and t are natural number, i ≠ j ≠ t.
Step 3: for remaining any pretreated vector data point X of web steel surface defect imagei, repeat to walk Rapid two, obtain minimal linear and represent error coefficient matrix W.
Step 4: setting up the mathematical modeling J of multiple manifold divergences
In formula (3)：
S(Xi) represent and the pretreated vector data point X of a web steel surface defect imageiRelated manifold is local special Levy space；
S(Xm) represent and the pretreated vector data point X of another web steel surface defect imagemRelated manifold is local Feature space；
||S(Xi)-S(Xm) | | represent and the pretreated vector data point X of a web steel surface defect imageiRelated Manifold local feature space S (Xi) arrive and the pretreated vector data point X of another web steel surface defect imagemRelated stream Shape local feature space S (Xm) between Euclidean distance Dim；
HimRepresent the pretreated vector data point X of a web steel surface defect imageiWith another width steel strip surface defect Vector data point X after image preprocessingmBetween uneven class size information：
Step 5: with the pretreated vector data point X of a web steel surface defect imageiRelated manifold local feature Space S (Xi) arrive and the pretreated vector data point X of another web steel surface defect imagemRelated manifold local feature is empty Between S (Xm) between Euclidean distance DimFor：
In formula (5)：
fi (P)(Xm) represent the pretreated vector data point X of a web steel surface defect imageiWith another web steel table Vector data point X after planar defect image preprocessingmRelated manifold local feature space S (Xm) in projection；
fm (P)(Xi) represent another pretreated vector data point X of web steel surface defect imagemWith a web steel table Vector data point X after planar defect image preprocessingiRelated manifold local feature space S (Xi) in projection.
Step 6: building target function model
s.t ATX(I-W)(I-W)TXTA=I
In formula (6)：
A represents low dimension projective matrix；
T representing matrix transposition；
I represents unit matrix.
Generalized eigenvalue decomposition is carried out to formula (6)
In formula (7)：
λ represents characteristic value；
F represents characteristic vector.
The descending order of eigenvalue λ is arranged, the characteristic vector f composition low-dimensionals before taking corresponding to d characteristic value Projection matrix A, wherein d represent the dimension of extracted feature.
Step 7: for the pretreated vector data point X of a web steel surface defect imagei, after linear projection Low-dimensional characteristic YiFor：
Yi=ATXi (8)
In formula (8)：A represents low dimension projective matrix；
T representing matrix transposition.
The k is the natural number more than 2.
The beneficial effect of present embodiment is：
Present embodiment extracts the spy of steel strip surface defect image using the method based on local feature space length Levy, manifold local feature space S (X is on the one hand set up using the classification information of vector datai), with different classes of manifold office Portion's feature space sets up the mathematical modeling J of multiple manifold divergence apart from sums, on the other hand in manifold local feature space S (Xi) in Minimal linear is calculated using local linear expression and represents error coefficient Wj, maintain the partial structurtes information of multiple manifold.Because What multiple manifold divergence embodied is dispersion degree between different classes of data, therefore maximizes multiple manifold divergence and can find most preferably Classification subspace, realizes that the low-dimensional of steel strip surface defect image differentiates effective extraction of feature, improves steel strip surface defect figure As the effect of identification.
Therefore, present embodiment is by maximizing multiple manifold divergence JsThe classification for extracting steel strip surface defect image is special Levy, with improve steel strip surface defect image recognition effect the characteristics of.

Claims (2)

1. a kind of steel strip surface defect image characteristic extracting method based on local feature space length, it is characterised in that the side Method is comprised the concrete steps that：
Step 1: the steel strip surface defect image of collection is carried out successively gray processing processing, smoothing processing, normalized and to Quantify, obtain the pretreated vector data point X of a web steel surface defect imagei, all web steel surface defect images are pre- Vector data point X after processingiMatrix data X is constituted, wherein：I=1,2 ..., n, n represent all steel strip surface defect images Sum；
Step 2: being found and a width steel strip surface defect from the matrix data X after all steel strip surface defect image preprocessings Vector data point X after image preprocessingiAfter the minimum pretreatment with classification identical k web steel surfaces defect image of Euclidean distance Vector data point, constitute with the pretreated vector data point X of a web steel surface defect imageiRelated manifold is local special Levy space S (Xi), minimal linear represents error coefficient WjFor：
In formula (1) and (2)：
XiThe pretreated vector data point of a web steel surface defect image is represented,
XjRepresent and the pretreated vector data point X of a web steel surface defect imageiEuclidean distance minimum and classification identical Pretreated j-th of vector data point of k web steel surface defect images,
GjtRepresent local Gtram matrix, Gjt=(Xi-Xj)·(Xi-Xt),
XtRepresent and the pretreated vector data point X of a web steel surface defect imageiEuclidean distance minimum and classification identical Pretreated t-th of vector data point of k web steel surface defect images,
I, j and t are natural number, i ≠ j ≠ t；
Step 3: for remaining any pretreated vector data point X of web steel surface defect imagei, repeat step two, Obtain minimal linear and represent error coefficient matrix W；
Step 4: setting up the mathematical modeling J of multiple manifold divergences
In formula (3)：
S(Xi) represent and the pretreated vector data point X of a web steel surface defect imageiRelated manifold local feature is empty Between,
S(Xm) represent and the pretreated vector data point X of another web steel surface defect imagemRelated manifold local feature Space,
||S(Xi)-S(Xm) | | represent and the pretreated vector data point X of a web steel surface defect imageiRelated manifold Local feature space S (Xi) arrive and the pretreated vector data point X of another web steel surface defect imagemRelated manifold office Portion feature space S (Xm) between Euclidean distance Dim,
HimRepresent the pretreated vector data point X of a web steel surface defect imageiWith another web steel surface defect image Pretreated vector data point XmBetween uneven class size information：
Step 5: with the pretreated vector data point X of a web steel surface defect imageiRelated manifold local feature space S (Xi) arrive and the pretreated vector data point X of another web steel surface defect imagemRelated manifold local feature space S (Xm) between Euclidean distance DimFor：
In formula (5)：
Represent the pretreated vector data point X of a web steel surface defect imageiLack with another web steel surface The vector data point X fallen into after image preprocessingmRelated manifold local feature space S (Xm) in projection,
Represent the pretreated vector data point X of another web steel surface defect imagemLack with a web steel surface The vector data point X fallen into after image preprocessingiRelated manifold local feature space S (Xi) in projection；
Step 6: building target function model
In formula (6)：
A represents low dimension projective matrix,
T representing matrix transposition,
I represents unit matrix；
Generalized eigenvalue decomposition is carried out to formula (6)
In formula (7)：
λ represents characteristic value,
F represents characteristic vector；
The descending order of eigenvalue λ is arranged, the characteristic vector f composition low dimension projectives before taking corresponding to d characteristic value Matrix A, wherein d represent the dimension of extracted feature；
Step 7: for the pretreated vector data point X of a web steel surface defect imagei, the low-dimensional after linear projection Characteristic YiFor：
Yi=ATXi (8)
In formula (8)：A represents low dimension projective matrix,
T representing matrix transposition.
2. the steel strip surface defect image characteristic extracting method according to claim 1 based on local feature space length, It is characterized in that the k is the natural number more than 2.
CN201410695659.3A 2014-11-26 2014-11-26 Steel strip surface defect image characteristic extracting method based on local feature space length Expired - Fee Related CN104376566B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410695659.3A CN104376566B (en) 2014-11-26 2014-11-26 Steel strip surface defect image characteristic extracting method based on local feature space length

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410695659.3A CN104376566B (en) 2014-11-26 2014-11-26 Steel strip surface defect image characteristic extracting method based on local feature space length

Publications (2)

Publication Number Publication Date
CN104376566A CN104376566A (en) 2015-02-25
CN104376566B true CN104376566B (en) 2017-08-08

Family

ID=52555456

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410695659.3A Expired - Fee Related CN104376566B (en) 2014-11-26 2014-11-26 Steel strip surface defect image characteristic extracting method based on local feature space length

Country Status (1)

Country Link
CN (1) CN104376566B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110709688B (en) * 2017-04-13 2022-03-18 英卓美特公司 Method for predicting defects in an assembly unit
CN107784323B (en) * 2017-10-16 2019-11-08 武汉科技大学 Steel strip surface defect image classification method based on the uncorrelated constraint in part

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103514445A (en) * 2013-10-15 2014-01-15 武汉科技大学 Strip steel surface defect identification method based on multiple manifold learning
CN103530658A (en) * 2013-10-15 2014-01-22 武汉科技大学 Method for recognizing plant leaf data based on sparse representation
CN103940767A (en) * 2014-05-07 2014-07-23 重庆大学 Gas concentration inversion method based on multi-manifold learning

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006277415A (en) * 2005-03-29 2006-10-12 Sanyo Electric Co Ltd Registration method and device, and authentication method and device

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103514445A (en) * 2013-10-15 2014-01-15 武汉科技大学 Strip steel surface defect identification method based on multiple manifold learning
CN103530658A (en) * 2013-10-15 2014-01-22 武汉科技大学 Method for recognizing plant leaf data based on sparse representation
CN103940767A (en) * 2014-05-07 2014-07-23 重庆大学 Gas concentration inversion method based on multi-manifold learning

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
Bo Li等.Constrained Maximum Variance Mapping.《2008 International Joint Conference on Neural Networks(IJCNN 2008)》.2008,第534-537页. *
Bo Li等.Feature extraction using constrained maximum variance mapping.《Patten Recognition》.2008,第3287-3294页. *
LIU Yuchao等.Feature extraction using supervised constrained maximum variance mapping.《International Conference on Automatic Control and Artificial Intelligence (ACAI 2012)》.2012,第1049-1052页. *
吴飞等.基于流形学习的整体正交稀疏保留鉴别分析.《计算机技术与发展》.2014,第24卷(第6期),第63-66页. *
王娜等.基于特征子空间邻域的局部保持流形学习算法.《计算机应用研究》.2012,第29卷(第4期),第1318-1321页. *

Also Published As

Publication number Publication date
CN104376566A (en) 2015-02-25

Similar Documents

Publication Publication Date Title
Tang et al. A local binary pattern based texture descriptors for classification of tea leaves
CN105005565B (en) Live soles spoor decorative pattern image search method
CN100461204C (en) Method for recognizing facial expression based on 2D partial least square method
CN107377634B (en) A kind of hot-strip outlet Crown Prediction of Media method
CN103262118B (en) Attribute value estimation device and property value method of estimation
CN107423760A (en) Based on pre-segmentation and the deep learning object detection method returned
CN102298773B (en) Shape-adaptive non-local mean denoising method
Cao et al. Integrated prediction model of bauxite concentrate grade based on distributed machine vision
CN101958000B (en) Face image-picture generating method based on sparse representation
CN107743225B (en) A method of it is characterized using multilayer depth and carries out non-reference picture prediction of quality
CN107392237B (en) Cross-domain foundation cloud picture classification method based on migration visual information
CN106599833B (en) A kind of face identification method adapted to based on field and manifold distance is measured
CN112818850B (en) Cross-posture face recognition method and system based on progressive neural network and attention mechanism
CN101303728A (en) Method for identifying fingerprint facing image quality
Faraki et al. Approximate infinite-dimensional region covariance descriptors for image classification
CN105427313A (en) Deconvolutional network and adaptive inference network based SAR image segmentation method
CN112488211A (en) Fabric image flaw classification method
CN103440494A (en) Horrible image identification method and system based on visual significance analyses
CN106169083A (en) The film of view-based access control model feature recommends method and system
CN104376566B (en) Steel strip surface defect image characteristic extracting method based on local feature space length
CN105447488B (en) SAR image target detection method based on sketch line segment topological structure
CN105205816A (en) Method for extracting high-resolution SAR image building zone through multi-feature weighted fusion
CN110399814A (en) It is a kind of to indicate that field adapts to the face identification method of measurement based on local linear
CN105023239A (en) Hyperspectral data dimensionality reduction method based on ultra-pixel and maximum boundary distribution
CN107784323B (en) Steel strip surface defect image classification method based on the uncorrelated constraint in part

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20170808

Termination date: 20211126