CN108460400A - A kind of hyperspectral image classification method of combination various features information - Google Patents

A kind of hyperspectral image classification method of combination various features information Download PDF

Info

Publication number
CN108460400A
CN108460400A CN201810002038.0A CN201810002038A CN108460400A CN 108460400 A CN108460400 A CN 108460400A CN 201810002038 A CN201810002038 A CN 201810002038A CN 108460400 A CN108460400 A CN 108460400A
Authority
CN
China
Prior art keywords
high spectrum
various features
dictionary
image
spectrum 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.)
Granted
Application number
CN201810002038.0A
Other languages
Chinese (zh)
Other versions
CN108460400B (en
Inventor
杨明
张会敏
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.)
Nanjing Ciku Network Information Technology Co ltd
Original Assignee
Nanjing Normal 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 Nanjing Normal University filed Critical Nanjing Normal University
Priority to CN201810002038.0A priority Critical patent/CN108460400B/en
Publication of CN108460400A publication Critical patent/CN108460400A/en
Application granted granted Critical
Publication of CN108460400B publication Critical patent/CN108460400B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • 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
    • 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
    • G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
    • Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture

Abstract

The present invention discloses a kind of hyperspectral image classification method of combination various features information, includes the following steps：Step 1, the spectrum of high spectrum image to be sorted, gradient, texture, shape various features data are extracted：Step 2, high spectrum image to be sorted is split using fractional spins, is divided into several spatial neighbors groups；Step 3, dictionary and sparse coding are obtained with MFKSADL model learnings；Step 4, SVM classifier is trained using code coefficient, predicts high spectrum image test set label.The problems such as such method can solve the different spectrum of jljl present in high spectrum image, same object different images, can effectively improve classification hyperspectral imagery precision.

Description

A kind of hyperspectral image classification method of combination various features information
Technical field
The invention belongs to Hyperspectral imagery processing field, more particularly to a kind of high spectrum image of combination various features information Sorting technique.
Background technology
High-spectrum remote sensing, each pixel indicate by hundreds of spectral values, corresponds to from visible spectrum to infrared The different narrow wavelength of line, these values can provide the fine SPECTRAL DIVERSITY between different atural objects, this be with higher precision detection and It distinguishes various atural objects and provides possibility.Therefore, classification hyperspectral imagery is widely used in various fields, including Environmental protection, land use monitoring, urban planning, deep woods fire detection, air monitoring, military combat etc..In high spectrum image Abundant spectral information has also contained lot of challenges and problem, such as higher-dimension small sample classification problem, and " the different spectrum of jljl, with spectrum Foreign matter " phenomenon etc..
By the sparse representation method that human visual system's sparse coding mechanism inspires, obtained in classification hyperspectral imagery field To application.Although the pixel of high spectrum image is high dimensional signal, similar pixel is usually located in same lower dimensional space, and It is indicated by identical dictionary atom.Therefore, each pixel can be used several atoms in entire training dictionary sparsely to indicate, sparse system Position of the number comprising selected atom and weight.Then, the class label of each pixel can be determined by corresponding sparse coefficient.Dilute It dredges in representation method, the discrimination of dictionary is considerable.It may be lacked using training set as dictionary there are two apparent Point.First, initial data generally includes noise, this will reduce classification performance.Secondly, seek sparse coding from extensive dictionary Calculating cost it is higher, this may limit its practical application.In order to solve these problems, in recent years, from training sample middle school Handwriting practicing allusion quotation has been proved to that additional performance improvement can be provided for classification task.Dictionary learning method passes through code coefficient and word Allusion quotation improves performance.But the method for general dictionary learning is based only on single spectral signature information, and there is no consider EO-1 hyperion In advanced features information.
Single spectral signature only describes high spectrum image from an angle, and different types of feature has different resolutions Ability can provide related and complementary information.The present invention proposes a kind of new dictionary learning calculation based on multicharacteristic information Method solves the classification problem of high spectrum image.High spectrum image is divided into several spatial groups first and extracts high spectrum image Various features information data (spectrum, gradient, unity and coherence in writing and shape).It is then based on different type characteristic in the same space group There is corresponding sparse coding the hypothesis of identical sparse mode to constrain, and study obtains the dictionary and sparse coding of discriminating power. In addition, if the data of high spectrum image be linearly inseparable in original feature space or feature coding be it is similar, Then kernel method can project to initial data in more higher dimensional space to improve separability.
Invention content
The purpose of the present invention is to provide a kind of hyperspectral image classification method of combination various features information, can solve The problems such as certainly different spectrum of jljl present in high spectrum image, same object different images, can effectively improve classification hyperspectral imagery precision.
In order to achieve the above objectives, solution of the invention is：
A kind of hyperspectral image classification method of combination various features information, includes the following steps：
Step 1, the spectrum of high spectrum image to be sorted, gradient, texture, shape various features data are extracted：
Step 2, high spectrum image to be sorted is split using fractional spins, it is close is divided into several spaces Adjacent group；
Step 3, dictionary and sparse coding are obtained with MFKSADL model learnings；
Step 4, SVM classifier is trained using code coefficient, predicts high spectrum image test set label.
The detailed content of above-mentioned steps 2 is：The gradient image for extracting high spectrum image first, then utilizes watershed segmentation Algorithm is split gradient image, obtains the segmentation figure of high spectrum image, is divided into several spatial groups.
In above-mentioned steps 3, MFKSADL models obtain as follows：
A, if the high spectrum image that s category feature data indicateIn nuclear space Be expressed asWherein, s=1,2 ..., S indicate s kind features；N is pixel Number；bsFor the dimension of s category features；Pixel in same feature space carries out linear expression, table by other pixels Representation model is：Φ(Xs)=Φ (Ds)As, Φ (Ds) it is expression of the s category feature data dictionaries in nuclear space；AsIt is corresponding Encoder matrix；
B, if high spectrum image segmentation is divided into several spatial groups { g1,…,gG, the space that s kind characteristics indicate Group giMiddle pixel collectionWherein | gi| representation space group giMiddle pixel number is right The expression coefficient matrix answered
Step a and step b are combined by c, obtain such model：
Wherein,For DsI-th of atom.
In above-mentioned steps 3, the particular content of dictionary and sparse coding is obtained with MFKSADL model learnings is：
Model is transformed into following equivalent form：
DefinitionWherein aiWith bjRespectively the i-th row and jth row of matrix A and B；It is then described Step 3 comprises the concrete steps that：
Step 31, initialisation image matrix X, dictionary { Ds}s=1,…,S：
Step 32, fixed dictionary { Ds}s=1,…,S, update sparse coding Γ, it is below to each spatial group that former problem, which is degenerated, Split cavity oscillator：
Step 33, fixed sparse coding Γ, update dictionary { Ds}s=1,…,S：
Step 34, step 32-33 is repeated, until meet stopping criterion for iteration, the dictionary { D that output study obtainss}s =1,…,SWith sparse coding Γ.
The detailed content of above-mentioned steps 4 is：It is propped up using corresponding trained by the sparse coding that step 3 obtains of training set Hold vector machine classifier SVM；The sparse coding of pixel to be sorted is input in trained SVM classifier, is obtained corresponding Class label；To all pixels to be sorted, method is classified in due order, obtains final classification hyperspectral imagery result.
After adopting the above scheme, the present invention is specific to the image classification method of Hyperspectral imagery processing proposition.With it is existing There is technology to compare, the present invention has following characteristics：
First, by dividing spatial group to high spectrum image, bloom is preferably utilized in combining space information subsidiary classification The spatial information of spectrogram picture improves the classifying quality of high spectrum image with this；
Then, by merging various features information, it is effectively utilized the correlation and complementarity of various features information, is promoted Classification accuracy rate, enhancing classification robustness；
Furthermore using the representation theory for combining various features information, study more has the dictionary of discriminating power, and then is promoted dilute The discriminating power for dredging coding, the class label of sample to be sorted is obtained using sparse coding, improves high spectrum image indirectly Nicety of grading.Meanwhile the linear separability of data is improved by kernel method, it is same present in effective solution high spectrum image The different spectrum of object, same object different images problem, therefore there is higher use value.
Description of the drawings
Fig. 1 is the overall flow figure of the present invention；
Fig. 2 is that the present invention learns to obtain the flow chart of dictionary and sparse coding with MFKSADL.
Specific implementation mode
Below with reference to attached drawing, technical scheme of the present invention and advantageous effect are described in detail.
As shown in Figure 1, the present invention provides a kind of hyperspectral image classification method of combination various features information, including it is as follows Step：
Step 1, the spectrum of high spectrum image to be sorted, gradient, texture, shape various features data are extracted：Using existing Technology extracts the various features information of EO-1 hyperion, obtains the sample data in different characteristic space, lays the groundwork for step 3.A variety of spies Reference breath has correlation and complementarity, provides more effective information for the correct classification of high spectrum image, is further promoted Nicety of grading.
Step 2, high spectrum image to be sorted is split using watershed segmentation methods, it is close is divided into several spaces Adjacent group：The gradient image of high spectrum image is extracted first with the prior art, then utilizes fractional spins to gradient map As being split, the segmentation figure of high spectrum image is obtained, several spatial groups are divided into.Spatial information has consistent in group Property, classification is helped larger.
Step 3, dictionary and sparse coding are obtained with MFKSADL model learnings：Iteration updates dictionary and sparse coding, directly To stopping criterion for iteration is met, output study obtains dictionary and the corresponding coding of sample set data with discriminating power.
Step 4, SVM classifier is trained using code coefficient, predicts high spectrum image test set label：Utilize training set pair Being trained by the sparse coding that step 3 obtains of answering obtains support vector machine classifier SVM.By the sparse coding of pixel to be sorted It is input in trained SVM classifier, obtains corresponding class label.To all pixels to be sorted, method carries out in due order Classification, obtains final classification hyperspectral imagery result.
It should be noted that core of the invention step is to obtain dictionary and sparse volume with MFKSADL model learnings The description of code, specific implementation mode primarily focuses on step 3, and prior art realization can be used in steps 1 and 2 and step 4.
In the step 3, MFKSADL models can be obtained by the following steps：
(1) high spectrum image of s category feature data expression is setIn nuclear space Be expressed asWherein, s=1,2 ..., S indicate s kind features；N is pixel Number；bsFor the dimension of s category features.Pixel in same feature space can have other pixels to carry out linear expression, Indicate that model is：Φ(Xs)=Φ (Ds)As, Φ (Ds) it is expression of the s category feature data dictionaries in nuclear space；AsFor correspondence Encoder matrix.In practical applications, the possibility of Perfect Reconstruction is smaller, it is intended that reconstructed errorIt is as small as possible.
(2) it sets high spectrum image segmentation and is divided into several spatial groups { g1,…,gG, the sky that s kind characteristics indicate Between organize giMiddle pixel collectionWherein | gi| representation space group giMiddle pixel number. Corresponding expression coefficient matrixIt is bigger to belong to of a sort possibility for pixel in the same space group, can be by similar word Allusion quotation atom linear expression.And each pixel is made of similar dictionary atom as far as possible.In order to utilize neighborhood space Information, it is believed that the corresponding expression coefficient of pixel has row sparsity in the same space group.And correspond to the difference of the same space group Characteristic has the similitude of information, therefore in order to efficiently use complementation and the relevant information of various features information, we recognize Expression coefficient for the different characteristic data of corresponding the same space group has row sparsity.
(3) thought in step (1) and step (2) is combined, such model can be obtained：
Wherein, For DsI-th of atom.By space Similitude is combined with various features information, while learning a dictionary for having stronger identification, has been effectively retained more details Information, obtained corresponding sparse coding classification capacity higher.
As shown in Fig. 2, obtaining dictionary with MFKSADL model learnings and sparse coding is as follows：
Model is transformed into following equivalent form：
Here, we defineWherein aiWith bjRespectively the i-th row and jth of matrix A and B Row.
Step 31, initialisation image matrix X, dictionary { Ds}s=1,…,S：
Step 32, fixed dictionary { Ds}s=1,…,S, update sparse coding Γ, it is below to each spatial group that former problem, which is degenerated, Split cavity oscillator：
Step 33, fixed sparse coding Γ, update dictionary { Ds}s=1,…,S：
Step 34, step 32-33 is repeated, until meet stopping criterion for iteration, the dictionary { D that output study obtainss}s =1,…,SWith sparse coding Γ.
In summary, a kind of hyperspectral image classification method of combination various features information of the present invention, using combination MFKSADL (Multifeature Kernel Spatial-Aware Dictionary Learning) dictionary learning model, fills Divide and is adopted in conjunction with high spectrum image spatial neighborhood similitude using the correlation and complementarity of high spectrum image various features information With dictionary learning method, a kind of new hyperspectral image classification method is proposed.No matter the present invention is on subjective vision or in visitor It sees in evaluation index, has all accomplished being obviously improved for nicety of grading.In addition, the linear separability of data is improved by kernel method, The different spectrum of jljl present in high spectrum image, same object different images problem are efficiently solved, therefore there is higher use value.
Above example is merely illustrative of the invention's technical idea, and protection scope of the present invention cannot be limited with this, every According to technological thought proposed by the present invention, any change done on the basis of technical solution each falls within the scope of the present invention Within.

Claims (5)

1. a kind of hyperspectral image classification method of combination various features information, it is characterised in that include the following steps：
Step 1, the spectrum of high spectrum image to be sorted, gradient, texture, shape various features data are extracted：
Step 2, high spectrum image to be sorted is split using fractional spins, is divided into several spatial neighbors Group；
Step 3, dictionary and sparse coding are obtained with MFKSADL model learnings；
Step 4, SVM classifier is trained using code coefficient, predicts high spectrum image test set label.
2. a kind of hyperspectral image classification method of combination various features information as described in claim 1, it is characterised in that：Institute Stating the detailed content of step 2 is：The gradient image for extracting high spectrum image first, then utilizes fractional spins to gradient Image is split, and obtains the segmentation figure of high spectrum image, is divided into several spatial groups.
3. a kind of hyperspectral image classification method of combination various features information as described in claim 1, it is characterised in that：Institute It states in step 3, MFKSADL models obtain as follows：
A, if the high spectrum image that s category feature data indicateTable in nuclear space It is shown asWherein, s=1,2 ..., S indicate s kind features；N is of pixel Number；bsFor the dimension of s category features；Pixel in same feature space carries out linear expression by other pixels, indicates Model is：Φ(Xs)=Φ (Ds)As, Φ (Ds) it is expression of the s category feature data dictionaries in nuclear space；AsFor corresponding volume Code matrix；
B, if high spectrum image segmentation is divided into several spatial groups { g1,…,gG, the spatial group g that s kind characteristics indicatei Middle pixel collectionWherein | gi| representation space group giMiddle pixel number, it is corresponding Indicate coefficient matrix
Step a and step b are combined by c, obtain such model：
Wherein, For DsI-th of atom.
4. a kind of hyperspectral image classification method of combination various features information as claimed in claim 3, it is characterised in that：Institute It states in step 3, the particular content of dictionary and sparse coding is obtained with MFKSADL model learnings is：
Model is transformed into following equivalent form：
DefinitionWherein aiWith bjRespectively the i-th row and jth row of matrix A and B；The then step 3 comprise the concrete steps that：
Step 31, initialisation image matrix X, dictionary { Ds}s=1,…,S：
Step 32, fixed dictionary { Ds}s=1,…,S, update sparse coding Γ, it is to be detached below to each spatial group that former problem, which is degenerated, It solves：
Step 33, fixed sparse coding Γ, update dictionary { Ds}s=1,…,S：
Step 34, step 32-33 is repeated, until meet stopping criterion for iteration, the dictionary { D that output study obtainss}s=1,…,SWith Sparse coding Γ.
5. a kind of hyperspectral image classification method of combination various features information as claimed in claim 3, it is characterised in that：Institute Stating the detailed content of step 4 is：Train to obtain support vector machines point by the sparse coding that step 3 obtains using training set is corresponding Class device SVM；The sparse coding of pixel to be sorted is input in trained SVM classifier, corresponding class label is obtained； To all pixels to be sorted, method is classified in due order, obtains final classification hyperspectral imagery result.
CN201810002038.0A 2018-01-02 2018-01-02 Hyperspectral image classification method combining various characteristic information Active CN108460400B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810002038.0A CN108460400B (en) 2018-01-02 2018-01-02 Hyperspectral image classification method combining various characteristic information

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810002038.0A CN108460400B (en) 2018-01-02 2018-01-02 Hyperspectral image classification method combining various characteristic information

Publications (2)

Publication Number Publication Date
CN108460400A true CN108460400A (en) 2018-08-28
CN108460400B CN108460400B (en) 2022-05-20

Family

ID=63220654

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810002038.0A Active CN108460400B (en) 2018-01-02 2018-01-02 Hyperspectral image classification method combining various characteristic information

Country Status (1)

Country Link
CN (1) CN108460400B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109410223A (en) * 2018-11-07 2019-03-01 电子科技大学 A kind of SAR image segmentation method based on watershed algorithm and dictionary learning
CN110688890A (en) * 2019-08-13 2020-01-14 南京航空航天大学 Hyperspectral image classification method based on self-adaptive kernel sparse representation and multiple features
CN111046844A (en) * 2019-12-27 2020-04-21 中国地质大学（北京） Hyperspectral image classification method based on novel neighborhood selection constraint
CN112633045A (en) * 2019-10-09 2021-04-09 华为技术有限公司 Obstacle detection method, device, equipment and medium
CN113065403A (en) * 2021-03-05 2021-07-02 浙江大学 Hyperspectral imaging-based machine learning cell classification method and device
CN115546790A (en) * 2022-11-29 2022-12-30 深圳智能思创科技有限公司 Document layout segmentation method, device, equipment and storage medium

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101930547A (en) * 2010-06-24 2010-12-29 北京师范大学 Method for automatically classifying remote sensing image based on object-oriented unsupervised classification
CN103208011A (en) * 2013-05-05 2013-07-17 西安电子科技大学 Hyperspectral image space-spectral domain classification method based on mean value drifting and group sparse coding
CN103593853A (en) * 2013-11-29 2014-02-19 武汉大学 Remote-sensing image multi-scale object-oriented classification method based on joint sparsity representation
CN103971123A (en) * 2014-05-04 2014-08-06 南京师范大学 Hyperspectral image classification method based on linear regression Fisher discrimination dictionary learning (LRFDDL)
CN104123555A (en) * 2014-02-24 2014-10-29 西安电子科技大学 Super-pixel polarimetric SAR land feature classification method based on sparse representation
CN104281855A (en) * 2014-09-30 2015-01-14 西安电子科技大学 Hyperspectral image classification method based on multi-task low rank
CN106022358A (en) * 2016-05-11 2016-10-12 湖南大学 Hyper-spectral image classification method and hyper-spectral image classification device
US20160307073A1 (en) * 2015-04-20 2016-10-20 Los Alamos National Security, Llc Change detection and change monitoring of natural and man-made features in multispectral and hyperspectral satellite imagery
CN106778768A (en) * 2016-11-22 2017-05-31 广西师范大学 Image scene classification method based on multi-feature fusion
CN107358249A (en) * 2017-06-07 2017-11-17 南京师范大学 The hyperspectral image classification method of dictionary learning is differentiated based on tag compliance and Fisher
CN107527023A (en) * 2017-08-07 2017-12-29 西安理工大学 Classification of Polarimetric SAR Image method based on super-pixel and topic model

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101930547A (en) * 2010-06-24 2010-12-29 北京师范大学 Method for automatically classifying remote sensing image based on object-oriented unsupervised classification
CN103208011A (en) * 2013-05-05 2013-07-17 西安电子科技大学 Hyperspectral image space-spectral domain classification method based on mean value drifting and group sparse coding
CN103593853A (en) * 2013-11-29 2014-02-19 武汉大学 Remote-sensing image multi-scale object-oriented classification method based on joint sparsity representation
CN104123555A (en) * 2014-02-24 2014-10-29 西安电子科技大学 Super-pixel polarimetric SAR land feature classification method based on sparse representation
CN103971123A (en) * 2014-05-04 2014-08-06 南京师范大学 Hyperspectral image classification method based on linear regression Fisher discrimination dictionary learning (LRFDDL)
CN104281855A (en) * 2014-09-30 2015-01-14 西安电子科技大学 Hyperspectral image classification method based on multi-task low rank
US20160307073A1 (en) * 2015-04-20 2016-10-20 Los Alamos National Security, Llc Change detection and change monitoring of natural and man-made features in multispectral and hyperspectral satellite imagery
CN106022358A (en) * 2016-05-11 2016-10-12 湖南大学 Hyper-spectral image classification method and hyper-spectral image classification device
CN106778768A (en) * 2016-11-22 2017-05-31 广西师范大学 Image scene classification method based on multi-feature fusion
CN107358249A (en) * 2017-06-07 2017-11-17 南京师范大学 The hyperspectral image classification method of dictionary learning is differentiated based on tag compliance and Fisher
CN107527023A (en) * 2017-08-07 2017-12-29 西安理工大学 Classification of Polarimetric SAR Image method based on super-pixel and topic model

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
ALI SOLTANI-FARANI 等: "Spatial-Aware Dictionary Learning for Hyperspectral Image Classification", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》 *
ERLEI ZHANG 等: "Fast Multifeature Joint Sparse Representation for Hyperspectral Image Classification", 《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》 *
ERLEI ZHANG 等: "Weighted multifeature hyperspectral image classification via kernel joint sparse representation", 《NEUROCOMPUTING》 *
王凯 等: "一种多特征转换的高光谱影像自适应分类方法", 《武汉大学学报 信息科学版》 *
舒速 等: "基于分水岭分割和稀疏表示的高光谱图像分类方法", 《计算机科学》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109410223A (en) * 2018-11-07 2019-03-01 电子科技大学 A kind of SAR image segmentation method based on watershed algorithm and dictionary learning
CN110688890A (en) * 2019-08-13 2020-01-14 南京航空航天大学 Hyperspectral image classification method based on self-adaptive kernel sparse representation and multiple features
CN112633045A (en) * 2019-10-09 2021-04-09 华为技术有限公司 Obstacle detection method, device, equipment and medium
CN111046844A (en) * 2019-12-27 2020-04-21 中国地质大学（北京） Hyperspectral image classification method based on novel neighborhood selection constraint
CN113065403A (en) * 2021-03-05 2021-07-02 浙江大学 Hyperspectral imaging-based machine learning cell classification method and device
CN115546790A (en) * 2022-11-29 2022-12-30 深圳智能思创科技有限公司 Document layout segmentation method, device, equipment and storage medium

Also Published As

Publication number Publication date
CN108460400B (en) 2022-05-20

Similar Documents

Publication Publication Date Title
CN110414377B (en) Remote sensing image scene classification method based on scale attention network
CN109344736B (en) Static image crowd counting method based on joint learning
CN108038476B (en) A kind of facial expression recognition feature extracting method based on edge detection and SIFT
CN108460400A (en) A kind of hyperspectral image classification method of combination various features information
Fang et al. Spectral–spatial classification of hyperspectral images with a superpixel-based discriminative sparse model
Narihira et al. Learning lightness from human judgement on relative reflectance
CN105184298B (en) A kind of image classification method of quick local restriction low-rank coding
CN108108751B (en) Scene recognition method based on convolution multi-feature and deep random forest
CN110348399B (en) Hyperspectral intelligent classification method based on prototype learning mechanism and multidimensional residual error network
CN106845341B (en) Unlicensed vehicle identification method based on virtual number plate
CN107085716A (en) Across the visual angle gait recognition method of confrontation network is generated based on multitask
CN109766858A (en) Three-dimensional convolution neural network hyperspectral image classification method combined with bilateral filtering
CN111046732B (en) Pedestrian re-recognition method based on multi-granularity semantic analysis and storage medium
CN105989336B (en) Scene recognition method based on deconvolution deep network learning with weight
CN111160249A (en) Multi-class target detection method of optical remote sensing image based on cross-scale feature fusion
Jiang et al. A deep evaluator for image retargeting quality by geometrical and contextual interaction
CN108389189B (en) Three-dimensional image quality evaluation method based on dictionary learning
CN108509925B (en) Pedestrian re-identification method based on visual bag-of-words model
CN110210282A (en) A kind of moving target detecting method decomposed based on non-convex low-rank sparse
CN108182449A (en) A kind of hyperspectral image classification method
CN108932455B (en) Remote sensing image scene recognition method and device
CN105608478A (en) Combined method and system for extracting and classifying features of images
CN107767416A (en) The recognition methods of pedestrian's direction in a kind of low-resolution image
Duarte-Carvajalino et al. Multiscale representation and segmentation of hyperspectral imagery using geometric partial differential equations and algebraic multigrid methods
CN103955709A (en) Weighted synthetic kernel and triple markov field (TMF) based polarimetric synthetic aperture radar (SAR) image classification method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right

Effective date of registration: 20240105

Address after: Room B1015, 10th Floor, North Campus Teaching Complex Building, Nanjing University Gulou Campus, No. 22 Hankou Road, Gulou District, Nanjing City, Jiangsu Province, 210093

Patentee after: NANJING CIKU NETWORK INFORMATION TECHNOLOGY CO.,LTD.

Address before: No. 122, Ninghai Road, Gulou District, Nanjing, Jiangsu 210023

Patentee before: NANJING NORMAL University

TR01 Transfer of patent right