CN108021874A - A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum - Google Patents
A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum Download PDFInfo
- Publication number
- CN108021874A CN108021874A CN201711178807.4A CN201711178807A CN108021874A CN 108021874 A CN108021874 A CN 108021874A CN 201711178807 A CN201711178807 A CN 201711178807A CN 108021874 A CN108021874 A CN 108021874A
- Authority
- CN
- China
- Prior art keywords
- mrow
- pixel
- spectrum image
- high spectrum
- cluster
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F18/23—Clustering techniques
-
- 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
-
- 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
- G06V10/30—Noise filtering
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of EO-1 hyperion Endmember extraction preprocess method combined based on empty spectrum, including step to have:(1) multidimensional gaussian filtering is carried out to original high spectrum image, obtains a series of new high spectrum image;(2) spatially uniform calculating is carried out to new high spectrum image, obtains the corresponding space uniform sex index of each pixel in new high spectrum image;(3) original high spectrum image is handled with unsupervised clustering, and presorted to image, be a cluster per class;(4) each cluster is determined to the subset of pixel according to the evenness index of pixel, and determines most extreme cluster;(5) cluster is selected, Endmember extraction is carried out to the pixel in its subset, each cluster extracts an end member pixel, obtains one group of end member abundance figure;(6) identification and classification to image are completed according to end member abundance figure.The problem of present invention solves high spectrum image and has very in the case of strong noise, and high optical spectrum image end member extraction accuracy is not high.
Description
Technical field
The present invention relates to image processing field, more particularly to a kind of EO-1 hyperion Endmember extraction combined based on sky-spectrum to locate in advance
Reason method.
Background technology
The rise of high-spectrum remote-sensing is one of the maximum achievement of remote sensing technology the 1980s.Hyperspectral imager is obtained
The ground object anti-(hair) taken is penetrated spectral signal and is recorded in units of pixel.Due to sensor spatial resolution limitation with
And the complexity of background so that high spectrum image generally existing mixed pixel.If using mixed pixel as pure pixel into
The application studies such as row classification, target detection and identification, as a result have very big error.This just causes EO-1 hyperion Decomposition of Mixed Pixels
Problem becomes an important and key link of hyperspectral data processing.The work that we need to carry out be exactly inside pixel,
Decomposition of Mixed Pixels for Pure pixel (end member) and is obtained into shared ratio, graphical analysis is entered sub-pixed mapping rank.Solution is mixed
First step Endmember extraction have the function that important, it is the basis of follow-up study, and the precision of Endmember extraction will be to subsequent treatment
Precision, which produces, directly to be influenced.
Each pixel in high spectrum image is described jointly by spectral information and spatial information.Spectral information
It is to handle pixel by the way of independent, and spatial information then considers relations of they and surrounding neighbor.Traditional end
First extraction algorithm only considers merely spectral information when handling data, data is considered as a unordered list of spectral measurement
Handled not as piece image, have ignored existing spatial coherence between pixel so that exist in Endmember extraction result
Certain error.How to be combined using the spatial information between pixel with spectral information and carry out Endmember extraction, improve spectrum extraction essence
Degree, becomes the emphasis for Recent study.
The content of the invention
Carried in view of the above-mentioned drawbacks of the prior art, the present invention provides a kind of EO-1 hyperion end member combined based on sky-spectrum
Preprocess method is taken, in the case of solving high spectrum image with very strong noise, high optical spectrum image end member extraction accuracy is not high
The problem of.
A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum provided by the invention, its improvements are existed
In described method includes following steps:
(1) multidimensional gaussian filtering is carried out to original high spectrum image, obtains a series of new high spectrum image;
(2) spatially uniform calculating is carried out to the new high spectrum image, obtains each picture in the new high spectrum image
The corresponding space uniform sex index of element;
(3) the original high spectrum image is handled with unsupervised clustering, and presorted to image, be one per class
A cluster;
(4) each cluster is determined to the subset of pixel according to the evenness index of pixel, and determines most extreme cluster;
(5) cluster is selected, Endmember extraction is carried out to the pixel in its subset, each cluster extracts an end member pixel, obtains
To one group of end member abundance figure;
(6) identification and classification to image are completed according to the end member abundance figure.
Preferably, the formula of step (1) progress multidimensional gaussian filtering includes:
In formula, N (x | μ, σ) represents the probability density function of x;X represents to need to do the pixel of gaussian filtering;μ represents it is expected;
σ represents covariance matrix;D represents the dimension of x, is positive integer.
More preferably, the control parameter of gaussian filtering, different covariance matrix values are used as by the use of covariance matrix σ parameters
Correspondence obtains the new high spectrum image of different degree of filtration.
More preferably, step (2), which carries out spatially uniform calculating, includes:
Root-mean-square error, its formula are carried out to a series of new high spectrum image and the original high spectrum image
For:
In formula, RMSE represents root-mean-square error;N represents pixel number;X1,iRepresent picture in a series of new high spectrum image
Element;X2,iRepresent respective pixel in original high spectrum image;i∈[1,…,n];
Make i=1 ..., n, according to different covariance matrix σ values, calculates every in a series of new high spectrum image
The root-mean-square error coefficient of a pixel;
To the different error coefficients of the same pixel under same covariance matrix σ values, it is averaged, its formula is:
In formula,Represent the average value of the different error coefficient sums of same pixel;M represents the value of covariance matrix σ
Sum;NiRepresent new high spectrum image and the difference of the respective pixel of original high spectrum image;σjRepresent different σ values, j=
1,…,m;
By the average value of the different error coefficient sums of same pixelRefer to as the corresponding spatially uniform of each pixel
Number.
More preferably, step (3) includes:
Minimal noise separation conversion is carried out to the original high spectrum image, obtains the image of denoising, and new peak
The quantity of the atural object classification of spectrum picture;
ISODATA methods are used to the image of the denoising, definitely the other species of species, each classification is one
A cluster, number of clusters p.
More preferably, each cluster is determined the subset of pixel by step (4) according to the evenness index of pixel, and is determined most
The step of extreme cluster, includes:
1) according to the corresponding space uniform sex index of each pixel of step (2), ranking is carried out to pixel;
2) threshold value set according to the difference of the pixel and user, adheres to the pixel separately cluster;
3) according to the space uniform sex index, most extreme cluster is gone out using Orthogonal subspace projection method choice, is moved
Except the cluster with mixed pixel.
In technical scheme, before the method that tradition is used only that spectral information carries out Endmember extraction, add
A kind of preprocess method, carries out spatially uniform calculating using multiple dimensioned Gaussian filter algorithm, introduces spatial information, utilize light
Spectral clustering is merged spatial information and spectral information, then carries out Endmember extraction operation again.And this method is to making an uproar
Sound has robustness, and in the case of solving high spectrum image with very strong noise, high optical spectrum image end member extraction accuracy is not high
The problem of.It is also relatively more reasonable in principle at the same time, there is suitable practical value.
Brief description of the drawings
Fig. 1 is the flow chart of the embodiment of the present invention.
Embodiment
For the objects, technical solutions and advantages of the present invention are more clearly understood, referring to the drawings and preferred reality is enumerated
Example is applied, the present invention is described in more detail.However, it is necessary to illustrate, many details listed in specification are only to be
Reader is set to have a thorough explanation to the one or more aspects of the present invention, even without these specific details can also
Realize the aspects of the invention.
A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum provided by the invention, its flow chart such as Fig. 1
It is shown, specifically comprise the following steps:
(1) multidimensional gaussian filtering is carried out to original high spectrum image, obtains a series of new high spectrum image;
Gaussian filtering is a kind of linear smoothing filtering, suitable for eliminating Gaussian noise, is widely used in subtracting for image procossing
Make an uproar process.The concrete operations of gaussian filtering are:With each pixel in a template (or convolution, mask) scan image,
The weighted average gray value of pixel goes the value of alternate template central pixel point in the neighborhood determined with template.Formula is:
In formula, N (x μ, σ) represents the probability density function of x;X represents to need to do the pixel of gaussian filtering;μ represents it is expected;σ
Represent covariance matrix;D represents the dimension of x, is positive integer.Different σ values correspondences obtain the high-spectrum of different degree of filtration
Picture.Parameter σ value values are higher, and high-spectrum image space is more smooth, so the present embodiment is used as height by the use of covariance matrix σ parameters
The control parameter of this filtering, different σ values correspondence obtain the new high spectrum image of different degree of filtration.
(2) spatially uniform calculating is carried out to the new high spectrum image, obtains each picture in the new high spectrum image
The corresponding space uniform sex index of element, specifically includes:
Root-mean-square error, its formula are carried out to a series of new high spectrum image and the original high spectrum image
For:
In formula, RMSE represents root-mean-square error;N represents pixel number;X1,iRepresent picture in a series of new high spectrum image
Element;X2,iRepresent respective pixel in original high spectrum image;i∈[1,…,n];
I is made since 1, i=1 ..., n, according to different covariance matrix σ values, calculate a series of new EO-1 hyperion
The root-mean-square error coefficient of each pixel in image;
To the different error coefficients of the same pixel under same covariance matrix σ values, it is averaged, its formula is:
In formula,Represent the average value of the different error coefficient sums of same pixel;NiRepresent new high spectrum image and original
The difference of the respective pixel of beginning high spectrum image;σjRepresent different σ values, j=1 ..., m, m is σ values sum;
By the average value of the different error coefficient sums of same pixelRefer to as the corresponding spatially uniform of each pixel
Number.
(3) the original high spectrum image is handled with unsupervised clustering, and presorted to image, be one per class
A cluster, specifically includes:
Minimal noise separation conversion is carried out to the original high spectrum image, obtains the image of denoising, and new peak
The quantity of the atural object classification of spectrum picture;
ISODATA (dynamic clustering or iteration self-organizing data analysis) method is used to the image of the denoising,
The definitely other species of species, each classification are a cluster, number of clusters p.
(4) each cluster is determined to the subset of pixel according to the evenness index of pixel, and determines most extreme cluster, specifically
Include the following steps:
1) according to the corresponding space uniform sex index of each pixel of step (2), ranking is carried out to pixel;
2) threshold value set according to the difference of the pixel and user, adheres to the pixel separately cluster;The present embodiment uses
Pixel difference is scope less than 0.001, by pixel classifications.
3) according to the space uniform sex index, most extreme cluster is gone out using Orthogonal subspace projection method choice, is moved
Except the cluster with mixed pixel.
(5) cluster is selected, Endmember extraction is carried out to the pixel in its subset, each cluster extracts an end member pixel, obtains
To one group of end member abundance figure;
(6) each end member corresponds to a width abundance figure, and it is shared in entire image that this abundance figure represents this kind of end member
Determine ratio, the identification and classification to image are completed according to the end member abundance figure.
The above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, without departing from the principle of the present invention, some improvements and modifications can also be made, these improvements and modifications also should
It is considered as protection scope of the present invention.
Claims (6)
1. a kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum, it is characterised in that the described method includes as follows
Step:
(1) multidimensional gaussian filtering is carried out to original high spectrum image, obtains a series of new high spectrum image;
(2) spatially uniform calculating is carried out to the new high spectrum image, obtains each pixel pair in the new high spectrum image
The space uniform sex index answered;
(3) the original high spectrum image is handled with unsupervised clustering, and presorted to image, be a collection per class
Group;
(4) each cluster is determined to the subset of pixel according to the evenness index of pixel, and determines most extreme cluster;
(5) cluster is selected, Endmember extraction is carried out to the pixel in its subset, each cluster extracts an end member pixel, obtains one
Group end member abundance figure;
(6) identification and classification to image are completed according to the end member abundance figure.
2. EO-1 hyperion Endmember extraction preprocess method as claimed in claim 1, it is characterised in that it is high that step (1) carries out multidimensional
The formula of this filtering includes:
<mrow>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>|</mo>
<mi>&mu;</mi>
<mo>,</mo>
<mi>&sigma;</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msup>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mi>&pi;</mi>
<mo>)</mo>
</mrow>
<mrow>
<mi>D</mi>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msup>
</mfrac>
<mfrac>
<mn>1</mn>
<mrow>
<mo>|</mo>
<mi>&sigma;</mi>
<msup>
<mo>|</mo>
<mrow>
<mn>1</mn>
<mo>/</mo>
<mn>2</mn>
</mrow>
</msup>
</mrow>
</mfrac>
<mi>exp</mi>
<mo>{</mo>
<mo>-</mo>
<mfrac>
<mn>1</mn>
<mn>2</mn>
</mfrac>
<msup>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mo>)</mo>
</mrow>
<mi>T</mi>
</msup>
<msup>
<mi>&sigma;</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msup>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>-</mo>
<mi>&mu;</mi>
<mo>)</mo>
</mrow>
<mo>}</mo>
<mo>;</mo>
</mrow>
In formula, N (x | μ, σ) represents the probability density function of x;X represents to need to do the pixel of gaussian filtering;μ represents it is expected;σ tables
Show covariance matrix;D represents the dimension of x, is positive integer.
3. EO-1 hyperion Endmember extraction preprocess method as claimed in claim 2, it is characterised in that joined using covariance matrix σ
Control parameter of the number as gaussian filtering, different covariance matrix values correspond to obtain the new high-spectrum of different degree of filtration
Picture.
4. EO-1 hyperion Endmember extraction preprocess method as claimed in claim 1, it is characterised in that it is equal that step (2) carries out space
Even property, which calculates, to be included:
Root-mean-square error is carried out to a series of new high spectrum image and the original high spectrum image, its formula is:
<mrow>
<mi>R</mi>
<mi>M</mi>
<mi>S</mi>
<mi>E</mi>
<mo>=</mo>
<msqrt>
<mfrac>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>X</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>-</mo>
<msub>
<mi>X</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
<mi>n</mi>
</mfrac>
</msqrt>
<mo>;</mo>
</mrow>
In formula, RMSE represents root-mean-square error;N represents pixel number;X1,iRepresent pixel in a series of new high spectrum image;X2,i
Represent respective pixel in original high spectrum image;i∈[1,…,n];
Make i=1 ..., n, according to different covariance matrix σ values, calculates each picture in a series of new high spectrum image
The root-mean-square error coefficient of element;
To the different error coefficients of the same pixel under same covariance matrix σ values, it is averaged, its formula is:
<mrow>
<msub>
<mover>
<mi>N</mi>
<mo>&OverBar;</mo>
</mover>
<mi>i</mi>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>m</mi>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</munderover>
<msub>
<mi>N</mi>
<mi>i</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>&sigma;</mi>
<mi>j</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
In formula,Represent the average value of the different error coefficient sums of same pixel;M represents the value sum of covariance matrix σ;
NiRepresent new high spectrum image and the difference of the respective pixel of original high spectrum image;σjRepresent different σ values, j=1 ..., m;
By the average value of the different error coefficient sums of same pixelAs the corresponding space uniform sex index of each pixel.
5. EO-1 hyperion Endmember extraction preprocess method as claimed in claim 1, it is characterised in that step (3) includes:
Minimal noise separation conversion is carried out to the original high spectrum image, obtains the image of denoising, and new EO-1 hyperion
The quantity of the atural object classification of image;
ISODATA methods are used to the image of the denoising, definitely the other species of species, each classification is a collection
Group, number of clusters p.
6. EO-1 hyperion Endmember extraction preprocess method as claimed in claim 1, it is characterised in that step (4) is by each cluster
Evenness index according to pixel determines the subset of pixel, and the step of determining most extreme cluster includes:
1) according to the corresponding space uniform sex index of each pixel of step (2), ranking is carried out to pixel;
2) threshold value set according to the difference of the pixel and user, adheres to the pixel separately cluster;
3) according to the space uniform sex index, most extreme cluster is gone out using Orthogonal subspace projection method choice, removes band
There is the cluster of mixed pixel.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711178807.4A CN108021874A (en) | 2017-11-23 | 2017-11-23 | A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711178807.4A CN108021874A (en) | 2017-11-23 | 2017-11-23 | A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108021874A true CN108021874A (en) | 2018-05-11 |
Family
ID=62080090
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711178807.4A Pending CN108021874A (en) | 2017-11-23 | 2017-11-23 | A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108021874A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109948556A (en) * | 2019-03-21 | 2019-06-28 | 中国农业科学院农业资源与农业区划研究所 | Crops Classification recognition methods and system |
WO2020062843A1 (en) * | 2018-09-28 | 2020-04-02 | 河北华讯方舟太赫兹技术有限公司 | Pill coating image processing method and apparatus, and computer device and storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101030299A (en) * | 2007-03-29 | 2007-09-05 | 复旦大学 | Method for decomposing remote-sensing-mixed image element based on data space orthogonality |
CN101221243A (en) * | 2007-11-01 | 2008-07-16 | 复旦大学 | Remote sensing image mixed pixels decomposition method based on nonnegative matrix factorization |
CN101504315A (en) * | 2009-02-23 | 2009-08-12 | 北京航空航天大学 | Expansion morphology and orthogonal subspace projection combined end member automatic extraction method |
CN102346853A (en) * | 2010-07-29 | 2012-02-08 | 张灵芝 | Mixed pixel decomposition and end member extraction method |
CN103530875A (en) * | 2013-10-09 | 2014-01-22 | 哈尔滨工程大学 | End member extraction data preprocessing method |
-
2017
- 2017-11-23 CN CN201711178807.4A patent/CN108021874A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101030299A (en) * | 2007-03-29 | 2007-09-05 | 复旦大学 | Method for decomposing remote-sensing-mixed image element based on data space orthogonality |
CN101221243A (en) * | 2007-11-01 | 2008-07-16 | 复旦大学 | Remote sensing image mixed pixels decomposition method based on nonnegative matrix factorization |
CN101504315A (en) * | 2009-02-23 | 2009-08-12 | 北京航空航天大学 | Expansion morphology and orthogonal subspace projection combined end member automatic extraction method |
CN102346853A (en) * | 2010-07-29 | 2012-02-08 | 张灵芝 | Mixed pixel decomposition and end member extraction method |
CN103530875A (en) * | 2013-10-09 | 2014-01-22 | 哈尔滨工程大学 | End member extraction data preprocessing method |
Non-Patent Citations (2)
Title |
---|
GABRIEL MARTÍN ETAL: ""Spatial-Spectral Preprocessing Prior to Endmember Identification and Unmixing of Remotely Sensed Hyperspectral Data"", 《IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING》 * |
潘仲明: "《随机信号与系统》", 31 August 2013, 北京:国防工业出版社 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020062843A1 (en) * | 2018-09-28 | 2020-04-02 | 河北华讯方舟太赫兹技术有限公司 | Pill coating image processing method and apparatus, and computer device and storage medium |
CN109948556A (en) * | 2019-03-21 | 2019-06-28 | 中国农业科学院农业资源与农业区划研究所 | Crops Classification recognition methods and system |
CN109948556B (en) * | 2019-03-21 | 2020-10-27 | 中国农业科学院农业资源与农业区划研究所 | Crop classification identification method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Huang et al. | Spatial and spectral image fusion using sparse matrix factorization | |
Chen et al. | A comprehensive approach to mode clustering | |
CN104966085B (en) | A kind of remote sensing images region of interest area detecting method based on the fusion of more notable features | |
CN104732200B (en) | A kind of recognition methods of skin type and skin problem | |
CN112580590A (en) | Finger vein identification method based on multi-semantic feature fusion network | |
CN115410050B (en) | Tumor cell detection equipment based on machine vision and method thereof | |
CN108960404B (en) | Image-based crowd counting method and device | |
CN105976310A (en) | Partition-based VCA end-member extraction method | |
CN104281835B (en) | Face recognition method based on local sensitive kernel sparse representation | |
CN105184302B (en) | A kind of high optical spectrum image end member extraction method | |
CN107301643B (en) | Well-marked target detection method based on robust rarefaction representation Yu Laplce's regular terms | |
CN111652273B (en) | Deep learning-based RGB-D image classification method | |
CN108875623B (en) | Face recognition method based on image feature fusion contrast technology | |
CN106960182A (en) | A kind of pedestrian integrated based on multiple features recognition methods again | |
CN112906550B (en) | Static gesture recognition method based on watershed transformation | |
CN112308873B (en) | Edge detection method for multi-scale Gabor wavelet PCA fusion image | |
CN105046272A (en) | Image classification method based on concise unsupervised convolutional network | |
CN103886334A (en) | Multi-index fused hyperspectral remote sensing image dimensionality reduction method | |
Ma et al. | Multiscale 2-D singular spectrum analysis and principal component analysis for spatial–spectral noise-robust feature extraction and classification of hyperspectral images | |
CN108021874A (en) | A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum | |
CN107133579A (en) | Based on CSGF (2D)2The face identification method of PCANet convolutional networks | |
CN111666999A (en) | Remote sensing image classification method | |
CN112966781A (en) | Hyperspectral image classification method based on triple loss and convolutional neural network | |
Hong et al. | Saliency-based feature learning for no-reference image quality assessment | |
CN112149582A (en) | Hyperspectral image material identification method and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180511 |