CN105095864B - Aurora image detecting method based on deep learning two-dimensional principal component analysis network - Google Patents

Aurora image detecting method based on deep learning two-dimensional principal component analysis network Download PDF

Info

Publication number
CN105095864B
CN105095864B CN201510418703.0A CN201510418703A CN105095864B CN 105095864 B CN105095864 B CN 105095864B CN 201510418703 A CN201510418703 A CN 201510418703A CN 105095864 B CN105095864 B CN 105095864B
Authority
CN
China
Prior art keywords
order
image
matrix
single order
aurora
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
Application number
CN201510418703.0A
Other languages
Chinese (zh)
Other versions
CN105095864A (en
Inventor
韩冰
贾中华
高新波
李洁
宋亚婷
王平
王秀美
王颖
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.)
Xidian University
Original Assignee
Xidian University
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 Xidian University filed Critical Xidian University
Priority to CN201510418703.0A priority Critical patent/CN105095864B/en
Publication of CN105095864A publication Critical patent/CN105095864A/en
Application granted granted Critical
Publication of CN105095864B publication Critical patent/CN105095864B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/13—Satellite images
    • 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
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Astronomy & Astrophysics (AREA)
  • Remote Sensing (AREA)
  • Multimedia (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a kind of aurora image detecting method based on deep learning two-dimensional principal component analysis, and it is insufficient mainly to solve the problems, such as that the prior art extracts aurora image information.Implementation step is：(1) extract single order feature vector using two-dimensional principal component analysis method and produce firstorder filter matrix, obtain single order global characteristics；(2) single order global characteristics are extracted with its second order feature vector, while produces second order filter matrix, obtains second order global characteristics；(3) blocked histogram statistics is carried out to second order global characteristics, extracts blocked histogram feature；(4) classified with support vector machine classifier to blocked histogram feature, obtain classification results.The present invention realizes the computer auto-detection to existing three classes aurora image, and has the advantages that classification accuracy is high, feature extraction and computer picture detection available for aurora image.

Description

Aurora image detecting method based on deep learning two-dimensional principal component analysis network
Technical field
The invention belongs to technical field of image processing, is related to aurora image detecting method, is carried available for aurora characteristics of image Take, aurora image classification.
Background technology
Aurora be the high energy particle from magnetosphere be deposited to upper atmosphere and with a kind of air of neutral compound collision excitation Luminescence phenomenon, is one of important behaviour form of the high latitude activity of the earth caused by solar wind-Magnetosphere-ionosphere coupling.So people The bulk information of magnetosphere and solar-terrestrial physics electromagnetic activity can be obtained by the systematic observation of aurora form and its evolution, help Influence mode and degree in further investigation solar activity to the earth, the changing rule of right solution space synoptic process is with important Meaning.An important form of the aurora image as nature, its distribution and discloses in geomagnetic activity and magnetosphere Pulsation phenomenon.The sort research of aurora image has reached a level at present, it is difficult to breaks through, we need to find a kind of energy now The method for effectively representing aurora characteristics of image.The textural characteristics of aurora image in itself, it should carry out sufficiently extraction with more preferable Characterize aurora characteristics of image.
An important form of the aurora image as nature, to their morphological feature and genesis mechanism in contrast Through clearer.Wang Q et al. are in article " Spatial texture based automatic classification of dayside aurora in all-sky images.Journal of Atmospheric and Solar-Terrestrial Physics,2010,72(5):Show that the statistical distribution of four kinds of aurora types is advised by studying still image in 498-508. " Rule, and preliminary identification Hu et al. is in article " Synoptic distribution of dayside aurora: Multiple-wavelength all-sky observation at Yellow River Station in Ny- Svalbard.Journal of Atmospheric and Solar-Terrestrial Physics,2009,71(8):794- Day side pole light is divided into the reasonability of four classes in 804. " according to zone of action, also gives these four classification from morphology direction Explanation.Han et al. is in article " Aurora images classification via features salient Coding.Journal of xidian university, 2013,40 (6), the feature based of the middle propositions of 180-186. " are notable Property coding aurora image classification method aurora image is divided into four classes.Yang Xi et al. are in article " Wavelet hierarchical model for aurora images classification.Journal of xidian university, 2013,40(2):Aurora image classification algorithms under a kind of layering wavelet model proposed in 18-24. ", profit Aurora image is divided into arcuation and three kinds of crown shape types with support vector machine classifier.Han et al. is in article " Aurora image classification based on LDA combining with saliency information.Ruan Jian Xue Bao/Journal of Software,2013,24(11):2758-2766. is " middle using based on the potential master for merging notable information Inscribe aurora image classification method and carry out aurora image classification.However, the existing research for aurora image mostly be carry out arcuation and Two sort researches of non-arcuation, do not obtain preferable effect yet in three sort research of aurora image, and reason is it not to pole The feature of light image is fully extracted, so as to have impact on the Accurate classification of aurora image.
The content of the invention
It is an object of the invention to the deficiency for above-mentioned prior art, proposes that one kind is based on deep learning two dimension principal component The aurora image detecting method of network is analyzed, improves the accuracy rate to aurora image classification.
The technical scheme is that：The advantage of aurora characteristics of image can be more fully extracted using deep learning, to aurora Image is detected, and implementation step includes as follows：
(1) single order global characteristics U is extracted：
1a) utilize the L of two-dimensional principal component analysis method extraction aurora image1A single order feature vector
1b) corresponding each single order feature vectorGenerate a firstorder filter matrix
1c) by i-th aurora image MiWith d-th of firstorder filter matrixConvolution is carried out, generates the independent feature of single order Image
1d) successively by N aurora images and d-th of firstorder filter matrixConvolution is carried out, generates single order local feature ImageWherein, M1dFor aurora image and d-th of firstorder filter matrix convolution The single order local feature image generated afterwards；
1c 1e) is carried out to all firstorder filter matrixes) and 1d) operate, generate corresponding single order local feature figure Picture, by L1A single order local feature image is combined as single order global characteristics
(2) second order global characteristics R is extracted：
2a) take d-th of single order local feature image M in single order global characteristics U1d, carried using two-dimensional principal component analysis method Take its L2A second order feature vector
2b) corresponding each second order feature vectorGenerate a second order filter matrix
2c) by i-th independent characteristic image of single orderWith k-th of second order filter matrixCarry out convolution, generation two The independent characteristic image of rank
2d) successively by the independent characteristic image of N number of single order and k-th of second order filter matrixConvolution is carried out, generates second order Local feature image
2e) by k-th of second order filter matrixWith single order local feature image M1dIn the independent characteristic pattern of each single order As carrying out 2c) and 2d) operate, generate corresponding second order integration characteristics image
All second order filter matrixes and single order global characteristics U 2f) are subjected to 2a) -2e) operation, generate corresponding Second order integration characteristics image, by L2A second order integration characteristics image is combined as second order global characteristics
(3) the blocked histogram feature f of second order global characteristics R is extracted；
(4) classify to blocked histogram feature f, arcuation aurora image h is detected from h aurora image1, radiation Coronal aurorae image h2With valance Coronal aurorae image h3Three classes.
The invention has the advantages that：
Two-dimensional principal component analysis method is passed through multilayer convolutional/topological structure by the present invention as a result of the thought of deep learning Connection, multilayer extraction is carried out to the feature of aurora image, and the positional information of aurora image is fully used, and is overcome existing There is the shortcomings that technology only carries out aurora characteristics of image individual layer extraction, so as to improve classification accuracy, realize three classes aurora The computer automatic sorting of image.
Brief description of the drawings
Fig. 1 be the present invention realize flow chart.
Embodiment
Present disclosure and effect are described further below in conjunction with attached drawing.
Step 1：Extract the single order global characteristics U of aurora image.
The targeted research object of this example is aurora image, it is existed by the all-sky video camera of China's Arctic Yellow River Station The aurora image that Yellow River Station photographs.The step of extracting the single order global characteristics of the aurora image is as follows：
1a) extract the L of aurora image1A single order feature vector
Extracting aurora image single order feature vector can use a variety of existing methods to carry out, such as two-dimensional principal component analysis method, The methods of Principal Component Analysis, this example extract the L of aurora image using two-dimensional principal component analysis method1A single order feature vectorStep is as follows：
After 1a1) inputting N aurora images, the single order mean matrix of N aurora images of extraction
The single order local mean value matrix of single width aurora image 1a2) is calculated respectively
1a3) by N number of single order local mean value matrixIt is combined into single order overall situation Mean Matrix
1a4) to single order overall situation Mean Matrix X and its transposed matrix XTIt is product generation single order product matrix P=XXT；
The feature vector of single order product matrix P 1a5) is calculated, and takes preceding L1A feature vector is as single order feature vector
1b) corresponding each single order feature vectorGenerate a firstorder filter matrix
1c) by i-th aurora image MiWith d-th of firstorder filter matrix Wd 1 Convolution is carried out, generation single order is independent Characteristic image
1d) successively by N aurora images and d-th of firstorder filter matrixConvolution is carried out, generates single order local feature ImageWherein, M1dFor aurora image and d-th of firstorder filter matrix convolution The single order local feature image generated afterwards；
1c 1e) is carried out to all firstorder filter matrixes) and 1d) operate, generate corresponding single order local feature figure Picture, by L1A single order local feature image is combined as single order global characteristics
Step 2：Extract second order global characteristics R.
2a) extract single order local feature image M1dL2A second order feature vector
Extract single order local feature image second order feature vector can use a variety of existing methods carry out, such as it is two-dimentional it is main into The methods of dividing analytic approach, Principal Component Analysis, this example extract the L of aurora image using two-dimensional principal component analysis method2A second order Feature vectorStep is as follows：
2a1) to single order local feature image M1d, extract the second order mean matrix of its all independent characteristic image of single order
The second order local mean value matrix of the independent characteristic image of single order 2a2) is calculated respectively
2a3) by N number of second order local mean value matrixSynthesize second order overall situation Mean Matrix
2a4) to second order overall situation Mean Matrix Y and its transposed matrix YTIt is product generation second-order product moment battle array Q=YYT；
The feature vector of second-order product moment battle array Q 2a5) is calculated, and takes preceding L2A feature vector is as second order feature vector
2b) corresponding each second order feature vectorGenerate a second order filter matrix
2c) by i-th independent characteristic image of single orderWith k-th of second order filter matrixCarry out convolution, generation two The independent characteristic image of rank
2d) successively by the independent characteristic image of N number of single order and k-th of second order filter matrixConvolution is carried out, generates second order Local feature image
2e) by k-th of second order filter matrixWith single order local feature image M1dIn the independent feature of each single order Image carries out 2c) and 2d) operate, generate corresponding second order integration characteristics image
All second order filter matrixes and single order global characteristics U 2f) are subjected to 2a) -2e) operation, generation corresponding two Rank integration characteristics image, by L2A second order integration characteristics image is combined as second order global characteristics
Step 3：Extract the blocked histogram feature f of second order global characteristics R.
3a) in second order global characteristics R, the independent characteristic pattern of i-th of second order of d-th of second order local feature image is chosen Picture, is translated into independent binary matrixWherein, H is transformed matrix symbol；
The independent characteristic image of i-th of second order of all second order local feature images 3b) is converted into independent binary system square Battle array, is combined into integration binary matrix
3c) integration blocked histogram statistics is done to integrating binary matrix
3a 3d) is carried out to all elements of all second order local feature images) -3c) operation, generate corresponding integrate and divide Block histogram, is combined into eigenmatrix f={ f1,f2,…,fi,…,fN}。
Step 4：Classification processing is carried out to blocked histogram feature f.
Classification processing is carried out to blocked histogram feature f many ripe existing sorting techniques, such as support vector machines Method, nearest neighbor classification, random forest classification etc..This example utilizes support vector machines method, and implementation step is：
First, eigenmatrix f is input to support vector machines；
Secondly, the fitting parameter of support vector machines is trained, obtains trained support vector machine classifier；
Finally, by classification of the trained support vector machine classifier completion to aurora image, i.e., by aurora image point For arcuation aurora, radiation Coronal aurorae and valance Coronal aurorae three types.
The effect of the present invention is further illustrated by following experiment simulation：
1. simulated conditions and method：
Hardware platform is：Intel Core i3、2.93GHz、3.45GB RAM；
Software platform is：MATLAB R2011b under Windows7 operating systems；
The data of experiment：The aurora data having had built up are selected in the all-sky aurora data of China's Arctic Yellow River Station Storehouse, these data come from December, 2003 in January, 2007, wherein aurora image totally 11150 width image, the aurora figure in sequence Gray level image as being 512*512.
2. emulation content and result
With it is of the invention and it is existing based on the aurora image detecting method of principal component analysis network to above-mentioned experimental data at the same time Classification processing is carried out, recall ratio and precision ratio the result such as table 1 of its aurora image separated.
1 arcuation aurora sequence recall ratio of table and precision ratio
As seen from Table 1, the present invention has good effect in aurora image classification, is significantly increased on classification accuracy.

Claims (4)

1. a kind of aurora image detecting method based on deep learning two-dimensional principal component analysis network, includes the following steps：
(1) single order global characteristics U is extracted：
1a) utilize the L of two-dimensional principal component analysis method extraction aurora image1A single order feature vector
1b) corresponding each single order feature vectorGenerate a firstorder filter matrix
1c) by i-th aurora image MiWith d-th of firstorder filter matrix Wd 1Convolution is carried out, generates the independent characteristic image of single order
1d) successively by N aurora images and d-th of firstorder filter matrix Wd 1Convolution is carried out, generates single order local feature imageWherein, M1dFor life after aurora image and d-th of firstorder filter matrix convolution Into single order local feature image；
1c 1e) is carried out to all firstorder filter matrixes) and 1d) operate, corresponding single order local feature image is generated, will L1A single order local feature image is combined as single order global characteristics
(2) second order global characteristics R is extracted：
2a) take d-th of single order local feature image M in single order global characteristics U1d, extract its L2A second order feature vector
2b) corresponding each second order feature vectorGenerate a second order filter matrix
2c) by i-th independent characteristic image of single orderWith k-th of second order filter matrix Wk 2Convolution is carried out, generates second order list Only characteristic image
2d) successively by the independent characteristic image of N number of single order and k-th of second order filter matrix Wk 2Convolution is carried out, generation second order is local Characteristic image
2e) by k-th of second order filter matrix Wk 2With single order local feature image M1dIn the independent characteristic image of each single order it is equal Carry out 2c) and 2d) operate, generate corresponding second order integration characteristics image M2k={ M21k,M22k,…,M2dk,…,M2L1k}；
All second order filter matrixes and single order global characteristics U 2f) are subjected to 2a) -2e) operation, generate corresponding second order Integration characteristics image, by L2A second order integration characteristics image is combined as second order global characteristics
(3) the blocked histogram feature f of second order global characteristics R is extracted；
(4) classify to blocked histogram feature f, arcuation aurora image h is detected from h aurora image1, radiation crown shape Aurora image h2With valance Coronal aurorae image h3Three classes.
2. according to the method described in claim 1, wherein described step 1a) in utilize two-dimensional principal component analysis method extraction aurora figure The L of picture1A single order feature vectorCarry out as follows：
After 1a1) inputting N aurora images, the single order mean matrix of N aurora images of extraction
The single order local mean value matrix of single width aurora image 1a2) is calculated respectively
1a3) by N number of single order local mean value matrixIt is combined into single order overall situation Mean Matrix
1a4) to single order overall situation Mean Matrix X and its transposed matrix XTIt is product generation single order product matrix P=XXT；
The feature vector of single order product matrix P 1a5) is calculated, and takes preceding L1A feature vector is as single order feature vector
3. according to the method described in claim 1, wherein described step 2a) in extract its L2A second order feature vectorIt is to utilize Two-dimensional principal component analysis method carries out, and step is as follows：
2a1) to single order local feature image M1d, extract the second order mean matrix of its all independent characteristic image of single order
The second order local mean value matrix of the independent characteristic image of single order 2a2) is calculated respectively
2a3) by N number of second order local mean value matrixSynthesize second order overall situation Mean Matrix
2a4) to second order overall situation Mean Matrix Y and its transposed matrix YTIt is product generation second-order product moment battle array Q=YYT；
The feature vector of second-order product moment battle array Q 2a5) is calculated, and takes preceding L2A feature vector is as second order feature vectorK=1, 2,…,L2。
4. the according to the method described in claim 1, blocked histogram of the extraction second order global characteristics R wherein described in step (3) Feature f, carries out as follows：
3a) in second order global characteristics R, the independent characteristic image of i-th of second order of d-th of second order local feature image is chosen, will It is converted into independent binary matrixWherein, H is transformed matrix symbol；
The independent characteristic image of i-th of second order of all second order local feature images 3b) is converted into independent binary matrix, group Binary matrix is integrated in synthesis
3c) integration blocked histogram statistics is done to integrating binary matrix
3a 3d) is carried out to all elements of all second order local feature images), 3b) and 3c) operate, generate corresponding integration Blocked histogram, is combined into eigenmatrix f={ f1,f2,…,fi,…,fN}。
CN201510418703.0A 2015-07-16 2015-07-16 Aurora image detecting method based on deep learning two-dimensional principal component analysis network Active CN105095864B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510418703.0A CN105095864B (en) 2015-07-16 2015-07-16 Aurora image detecting method based on deep learning two-dimensional principal component analysis network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510418703.0A CN105095864B (en) 2015-07-16 2015-07-16 Aurora image detecting method based on deep learning two-dimensional principal component analysis network

Publications (2)

Publication Number Publication Date
CN105095864A CN105095864A (en) 2015-11-25
CN105095864B true CN105095864B (en) 2018-04-17

Family

ID=54576253

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510418703.0A Active CN105095864B (en) 2015-07-16 2015-07-16 Aurora image detecting method based on deep learning two-dimensional principal component analysis network

Country Status (1)

Country Link
CN (1) CN105095864B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107194375A (en) * 2017-06-20 2017-09-22 西安电子科技大学 Video sequence sorting technique based on three-dimensional principal component analysis network
CN107273938B (en) * 2017-07-13 2020-05-29 西安电子科技大学 Multi-source remote sensing image ground object classification method based on two-channel convolution ladder network
CN109492593B (en) * 2018-11-16 2021-09-10 西安电子科技大学 Hyperspectral image classification method based on principal component analysis network and space coordinates
CN110514924B (en) * 2019-08-12 2021-04-27 武汉大学 Power transformer winding fault positioning method based on deep convolutional neural network fusion visual identification

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1975759A (en) * 2006-12-15 2007-06-06 中山大学 Human face identifying method based on structural principal element analysis
KR101117549B1 (en) * 2010-03-31 2012-03-07 경북대학교 산학협력단 Face recognition system and method thereof
CN103020640A (en) * 2012-11-28 2013-04-03 金陵科技学院 Facial image dimensionality reduction classification method based on two-dimensional principal component analysis
CN103473790A (en) * 2013-08-29 2013-12-25 西北工业大学 Online target tracking method based on increment bilateral two-dimensional principal component analysis (Bi-2DPCA) learning and sparse representation
CN103955697A (en) * 2014-03-12 2014-07-30 西安电子科技大学 2DPCA-based polarized SAR image classification method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1975759A (en) * 2006-12-15 2007-06-06 中山大学 Human face identifying method based on structural principal element analysis
KR101117549B1 (en) * 2010-03-31 2012-03-07 경북대학교 산학협력단 Face recognition system and method thereof
CN103020640A (en) * 2012-11-28 2013-04-03 金陵科技学院 Facial image dimensionality reduction classification method based on two-dimensional principal component analysis
CN103473790A (en) * 2013-08-29 2013-12-25 西北工业大学 Online target tracking method based on increment bilateral two-dimensional principal component analysis (Bi-2DPCA) learning and sparse representation
CN103955697A (en) * 2014-03-12 2014-07-30 西安电子科技大学 2DPCA-based polarized SAR image classification method

Also Published As

Publication number Publication date
CN105095864A (en) 2015-11-25

Similar Documents

Publication Publication Date Title
CN105512289B (en) Image search method based on deep learning and Hash
Kataoka et al. Pre-training without natural images
Garcia-Garcia et al. A review on deep learning techniques applied to semantic segmentation
CN108427920B (en) Edge-sea defense target detection method based on deep learning
CN111079685B (en) 3D target detection method
CN107239751B (en) High-resolution SAR image classification method based on non-subsampled contourlet full convolution network
CN108108751B (en) Scene recognition method based on convolution multi-feature and deep random forest
CN104217214B (en) RGB D personage's Activity recognition methods based on configurable convolutional neural networks
CN105095864B (en) Aurora image detecting method based on deep learning two-dimensional principal component analysis network
CN109829399A (en) A kind of vehicle mounted road scene point cloud automatic classification method based on deep learning
CN109034210A (en) Object detection method based on super Fusion Features Yu multi-Scale Pyramid network
CN106682697A (en) End-to-end object detection method based on convolutional neural network
CN110322453A (en) 3D point cloud semantic segmentation method based on position attention and auxiliary network
CN107944483B (en) Multispectral image classification method based on dual-channel DCGAN and feature fusion
CN109919011A (en) A kind of action video recognition methods based on more duration informations
CN107657634B (en) Shale digital core three-dimensional reconstruction method based on deep learning and support vector machine
CN108205806A (en) A kind of automatic analytic method of pyramidal CT image three-dimensional cranio-orbital tumor
CN109766934B (en) Image target identification method based on depth Gabor network
Wang et al. SCNet: Subdivision coding network for object detection based on 3D point cloud
CN105550712B (en) Aurora image classification method based on optimization convolution autocoding network
CN112037228A (en) Laser radar point cloud target segmentation method based on double attention
CN105956610B (en) A kind of remote sensing images classification of landform method based on multi-layer coding structure
CN112508066A (en) Hyperspectral image classification method based on residual error full convolution segmentation network
JP5464739B2 (en) Image area dividing apparatus, image area dividing method, and image area dividing program
CN113344110B (en) Fuzzy image classification method based on super-resolution reconstruction

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