CN105138966A - Quick density peak value clustering based polarimetric SAR image classification method - Google Patents
Quick density peak value clustering based polarimetric SAR image classification method Download PDFInfo
- Publication number
- CN105138966A CN105138966A CN201510467832.9A CN201510467832A CN105138966A CN 105138966 A CN105138966 A CN 105138966A CN 201510467832 A CN201510467832 A CN 201510467832A CN 105138966 A CN105138966 A CN 105138966A
- Authority
- CN
- China
- Prior art keywords
- class
- pixel
- classification
- polarimetric sar
- sar 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
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
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Probability & Statistics with Applications (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Multimedia (AREA)
- Radar Systems Or Details Thereof (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses a quick density peak value clustering based polarimetric SAR image classification method, mainly to solve the problem of low classification precision of an existing unsupervised polarimetric SAR classification method. The method comprises: 1. filtering an to-be-classified polarimetric SAR image; 2. performing Yamaguchi decomposition on the filtered polarimetric SAR image, and calculating four kinds of scattering power of each pixel point; 3. extracting main scattering power of each pixel point, and initially dividing the whole image into four classifications; 4. according to the main scattering power of all pixel points in an initial classification, dividing the whole polarimetric SAR image into M classifications; 5. using center points of the M classifications as new pixel points for clustering, and converting a clustering result into a pre-classification result of the whole image; and 6. performing iterative classification on the pre-classification result to obtain a final classification result. The experiment shows that the method provided by the prevent invention has a better classification effect, and can be used to perform unsupervised classification on various polarimetric SAR images.
Description
Technical field
The invention belongs to technical field of image processing, relate to Classification of Polarimetric SAR Image method, can be used for target identification.
Background technology
Along with remote sensing technology is in the development in the fields such as satellite, manned space flight, moon exploration program, polarization SAR has become the development trend of SAR, polarization SAR can obtain abundanter target information, there is investigation and application widely in agricultural, forestry, military affairs, geology, hydrology and ocean etc. be worth, as different field such as topographic mapping, resource exploration, disaster monitoring and astronomical researches.The object of polarization Images Classification utilizes polarization measurement data that the are airborne or acquisition of borne polarization sensor, determines the classification belonging to each pixel.In recent years, the international remote sensing fields that is sorted in utilizing polarization SAR measurement data to carry out is paid much attention to, and has become the main direction of studying of SAR image classification.
According to the difference of disposal route, full polarimetric SAR sorting technique can be divided into supervised classification and unsupervised classification.Supervised classification method mainly comprises the sorting technique of Corpus--based Method knowledge, neural network, wavelet analysis and fuzzy logic.Maximum in not supervised classification is the decomposition method of based target scattering mechanism, as H/ α goal decomposition method and Freeman decomposition method.For supervised classification method, the full-polarization SAR not supervised classification based on scattering mechanism does not need the training data of known class, has stronger adaptability.
Classical Classification of Polarimetric SAR Image method comprises:
The people such as 1.Cloude propose the non-supervisory Classification of Polarimetric SAR Image method based on H/ α goal decomposition, see CloudeSR, PottierE.Anentropybasedclassificationschemeforlandapplic ationsofpolarimetricSAR [J] .IEEETrans.Geosci.RemoteSensing.1997, 35 (1): 549-557. the method mainly decompose the feature obtaining H and α two sign polarization data by Cloude, then according to H and α composition H/ α plane artificial be divided into 9 regions, remove a region that can not exist in theory, image is divided into 8 classes the most at last.The defect that H/ alpha taxonomy exists is that the division in region is too dogmatic, when of a sort Data distribution8 is on the border of two classes or a few class, classifier performance will be deteriorated, another weak point be when coexist in same region several different atural object time, can not effectively distinguish.
The people such as 2.Lee propose the non-supervisory polarization SAR sorting technique based on H/ α goal decomposition and Wishart sorter, see LeeJS, GrunesMR, AinsworthTL, eta1.Unsupervisedclassificationusingpolarimetricdecompos itionandthecomplexWishartclassifier [J] .IEEETrans.Geosci.RemoteSensing.1999, 37 (5): 2249-2258. the method mainly utilize Wishart sorter to repartition each pixel to 8 classes after H/ α division, thus effectively raise the precision of classification, but there is the deficiency of the polarization scattering characteristics that well can not keep all kinds of.
Summary of the invention
The object of the invention is to the deficiency for prior art, propose a kind of Classification of Polarimetric SAR Image method based on fast density peak value cluster, to improve classification accuracy rate.
For achieving the above object, the present invention includes following steps:
(1) filtering is carried out to Polarimetric SAR Image to be sorted, remove speckle noise, obtain filtered Polarimetric SAR Image;
(2) Yamaguchi decomposition is carried out to the coherence matrix T of pixel each in filtered Polarimetric SAR Image, obtain the volume scattering power P of each pixel
v, dihedral angle scattering power P
d, surface scattering power P
swith conveyor screw scattering component P
h;
(3) according to each pixel four scattering power P
s, P
d, P
v, P
hmaximal value be four classes by Polarimetric SAR Image initial division:
If max is (P
s, P
d, P
v, P
h)=P
s, then the pixel of its correspondence is divided into a class, wherein P
sfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
d, then the pixel of its correspondence is divided into a class, wherein P
dfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
v, then the pixel of its correspondence is divided into a class, wherein P
vfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
h, then the pixel of its correspondence is divided into a class, wherein P
hfor such main scattering power;
Wherein, max () represents maximal value;
(4) press ascending sequence to the main scattering power of all pixels of every class initial category, and 300 pixels every in every class are further subdivided into a class, the most whole Polarimetric SAR Image is divided into M class;
(5) in an acquired M classification, using the central point of each class as new pixel, M new pixel A is obtained
i, i=1 ..., M, and with the new pixel A of every class
irepresent all pixels in this classification;
(6) fast density peak value cluster is carried out to above-mentioned M new pixel, M new pixel is gathered for k class;
(7) in the cluster result of M new pixel, will by new pixel A
iall pixels of representative are labeled as and new pixel A
iidentical classification, completes presorting to entire image;
(8) multiple Wishart iteration is carried out to the result of presorting of whole Polarimetric SAR Image, obtain classification results more accurately.
The present invention has the following advantages compared with prior art:
1. the present invention takes full advantage of Yamaguchi and decomposes the four kinds of scattering power P obtained
v, P
d, P
s, P
hvalidity in classification, makes the symmetric complex region of some discontented foot reflexs can more enough good classification.
2. a lot of classical taxonomy method in the classification of existing polarization SAR that the present invention is directed to all is confined to the defect of specific class categories number, adopt fast density peak value cluster, self-adaption cluster can be carried out according to the concrete condition of different Polarimetric SAR Image, adaptive selection class number, and classification results region consistency divides better, the edge after zones of different divides is also more clear.
Accompanying drawing explanation
Fig. 1 is general flow chart of the present invention;
Fig. 2 is the sub-process figure of fast density peak value cluster in the present invention;
Fig. 3 is original San Francisco polarization SAR data image that the present invention emulates use;
Fig. 4 is to the classification results figure of Fig. 3 by existing H/ α method, H/ α-Wishart method and the inventive method;
Fig. 5 is the original Fu Laifulan farmland polarization SAR data image that the present invention emulates use;
Fig. 6 is with the classification results figure of the present invention to Fig. 5.
Embodiment
With reference to Fig. 1, performing step of the present invention is as follows:
Step 1, carries out filtering to Polarimetric SAR Image to be sorted, removes speckle noise, obtains filtered Polarimetric SAR Image.
Usually be all adopt existing exquisiteness polarization LEE filter method to the filtering of Polarimetric SAR Image, the size of filter window is 7 × 7.
Step 2, carries out Yamaguchi decomposition to the coherence matrix T of pixel each in filtered Polarimetric SAR Image, obtains the volume scattering power P of each pixel
v, dihedral angle scattering power P
d, surface scattering power P
swith spiral scattering component P
h.
It is a kind of Polarization target decomposition method that Yamaguchi decomposes, different ground object targets has different scattering propertiess, the method is according to the scattering properties of different ground object target, the coherence matrix of each pixel is resolved into the linear combination of multiple scattering component, the advantage of the method other goal decomposition method relative is to which increase conveyor screw scattering component, make the symmetric complex region of some discontented foot reflexs can more enough good classification, see YamaguchiY, MaoriyamaT, IshidoM, eta1.Four-componentscatteringmodelforpolarimetricSARimag edecomposition [J] .IEEETrans.Geosci.RemoteSensing.2005, 43 (8): 1699-1706.The concrete steps that Yamaguchi decomposes are as follows:
(2a) read in each pixel of filtered image, these pixels are the coherence matrix T of 3 × 3, obtain covariance matrix C according to coherence matrix T;
Wherein, U is intermediate variable,
H represents horizontal polarization, and V represents vertical polarization, S
hHexpression level to launch and level to reception echo data, S
vVrepresent the echo data that Vertical dimension is launched and Vertical dimension receives, S
hVthe echo data that expression level receives to transmitting Vertical dimension, ()
*represent the conjugation of these data, < > represents average by looking number;
(2b) covariance matrix C is resolved into following expression:
Wherein, f
sfor the coefficient of dissociation of in-plane scatter component, f
dfor the coefficient of dissociation of dihedral angle scattering component, f
vfor the coefficient of dissociation of volume scattering component, f
hfor the coefficient of dissociation of spiral scattering component, β is the ratio that horizontal emission level receives back scattering reflection coefficient and Vertical Launch vertical reception back scattering emission ratio, and α is defined as α=I
hi
vH/ I
vi
vV, I
hand I
vrepresent level and the vertical reflection coefficient on earth's surface respectively, I
vHand I
vVrepresent level and the vertical reflection coefficient of vertical body of wall respectively, j represents imaginary number;
(2c) result of being decomposed by covariance matrix C, obtains one and has six unknown number f
s, f
v, f
d, f
h, the system of equations of α, β and five equations is as follows:
(2d) calculate in pixel covariance matrix C
value, if
then make α=-1, if
then make β=1, after the value of given α or β, remaining 4 unknown numbers are then according to formula 3) solve, wherein real part is got in Re () expression;
(2e) according to the f solved
s, f
v, f
d, f
h, α, β, solve volume scattering power P
v, dihedral angle scattering power P
d, surface scattering power P
swith spiral scattering power P
h:
P
s=f
s(1+|β|
2)
P
d=f
d(1+|a|
2)
P
v=f
v
P
h=f
h。
Step 3, according to each pixel four scattering power P
s, P
d, P
v, P
hmaximal value be four classes by Polarimetric SAR Image initial division:
If max is (P
s, P
d, P
v, P
h)=P
s, then the pixel of its correspondence is divided into a class, wherein P
sfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
d, then the pixel of its correspondence is divided into a class, wherein P
dfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
v, then the pixel of its correspondence is divided into a class, wherein P
vfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
h, then the pixel of its correspondence is divided into a class, wherein P
hfor such main scattering power;
Wherein, max () represents maximal value.
Step 4, press ascending sequence to the main scattering power of all pixels of every class initial category, and 300 pixels every in every class are further subdivided into a class, the most whole Polarimetric SAR Image is divided into M class.
Step 5, in an acquired M classification, using the central point of each class as new pixel, obtains M new pixel A
i, i=1 ..., M, and with the new pixel A of every class
irepresent all pixels in this classification.
Step 6, carries out fast density peak value cluster to above-mentioned M new pixel, gathers M new pixel for k class.
Conventional clustering method mainly comprises K-means cluster, spectral clustering, fuzzy clustering and fast density peak value cluster.This example adopts fast density peak value cluster, this clustering method is a kind of new clustering method in machine learning field, the method thinks that the data that cluster centre has low local density's point by some are surrounded, and these low-density data points are larger apart from other high density data point distance, and take this as a foundation adaptively selected cluster centre and cluster classification, see RodriguezA, andAlessandroL.Clusteringbyfastsearchandfindofdensitypea ks [J] .Science27.2014,344 (6191): 1492-1496.
With reference to Fig. 2, being implemented as follows of this step:
6a) calculate any two new pixel A
iand A
jbetween phase mutual edge distance d
ij:
d
ij=Tr((T
i)
-1+(T
j)
-1T
i)-q,
Wherein, T
i, T
jrepresent new pixel A respectively
iand A
jcoherence matrix, (T
i)
-1(T
j)
-1represent respectively matrix T
iand T
jinvert, q is constant, and value is q=3, Tr () is matrix trace;
6b) calculate new pixel A
ilocal density ρ
i:
Wherein, d is worked as
ij< d
ctime, χ (d
ij-d
c)=1, otherwise χ (d
ij-d
c)=0; M represents new pixel A
inumber; d
cfor constant, its value for by phase mutual edge distance d a little
ijascending arrangement, will be positioned at the value of this arrangement 2% position as d
cvalue;
6c) calculate new pixel A
idistance δ
i:
6d) select local density ρ
iwith distance δ
ithe maximum k of product new pixel is as cluster centre;
After 6e) cluster centre is determined, a class is represented with each cluster centre, k cluster centre represents k class altogether, relatively remain the distance of new pixel and each class cluster centre, if this new pixel and m class cluster centre is nearest, then this new pixel is divided into m class, m=1,, k.
Step 7, in the cluster result of M new pixel, will by new pixel A
iall pixels of representative are labeled as and new pixel A
iidentical classification, completes presorting to entire image;
Step 8, the entire image obtained presorting, with reflecting that the Wishart sorter of polarization SAR distribution character carries out Iterative classification, obtains classification results more accurately.
Wishart sorter is a kind of sorter that can reflect polarization SAR Data distribution8 characteristic, by the distance of compared pixels point and each cluster centre, judge the generic of pixel, see LeeJS, GrunesMR, PottierE, eta1.Unsupervisedterrainclassificationpreservingpolarime tricscatteringcharacteristic [J] .IEEETrans.Geosci.RemoteSensing.2004,42 (4): 722-731. its steps are as follows:
(8a) whole Polarimetric SAR Image is presorted the k class division result obtained, ask the cluster centre B of each class according to following formula
c:
Wherein T
ρthe coherence matrix of each pixel in c class, n
cit is the number of the pixel belonging to c class;
(8b) according to the cluster centre B of each class
c, calculate the distance d of each pixel i to c class cluster centre
ic:
Wherein T is the coherence matrix of pixel, and < > represents average by looking number, the determinant of [] representing matrix, the mark of Tr () representing matrix,
represent cluster centre B
cinvert;
(8c) distance of more each pixel and each cluster centre, if this new pixel and m class cluster centre is nearest, is then divided into m class by this pixel, m=1,, k, completes repartitioning view picture Polarimetric SAR Image classification after presorting;
(8d) repeat step (8a)-(8c), until iterations equals given iterations μ=4, obtain classification results.
Effect of the present invention can by following experimental verification:
1, experiment condition and method
Hardware platform is: Intel (R) Pentium (R) 1CPU2.4GHz;
Software platform is: WindowXPProfessional, MATLAB7.0.4;
Experimental technique: be respectively the method for existing H/ α and H/ α-Wishart method and the present invention, wherein these two kinds of methods existing are all quote more classical way in polarization SAR Data classification.
2. experiment content and result
Experiment one, to be that the SanFranciscoBay Polarimetric SAR Image of four is as test pattern depending on number shown in Fig. 3, carry out classification by the inventive method and existing H/ α method and H/ α-Wishart method to Fig. 3 to emulate, classification results is shown in Fig. 4, wherein, Fig. 4 (a) is the result of H/ α classification, and Fig. 4 (b) is the result of H/ α-Wishart classification, the classification results that Fig. 4 (c) is the inventive method.
From Fig. 4 (a), the waters part in image obtains reasonable division, but city and greenery patches etc. are obscured seriously.Therefore, the method classifying rules is too dogmatic, causes classifying quality not good.
From Fig. 4 (b), in conjunction with the H/ α-Wishart sorting technique of H/ α and Wishart sorter, it is more careful that image-region divides, but also have more Region dividing unclear, even occur that mistake divides, the mistake that the sea area as Fig. 4 (b) upper right corner occurs divides.
From Fig. 4 (c), classification results of the present invention visually sees better effects if, wherein the region consistency in the region such as golf course, racecourse, parking lot is significantly better than first two method, and between zones of different, sorted edge is also more level and smooth.
Experiment two will be that the Flevoland Polarimetric SAR Image of four is as test pattern depending on number shown in Fig. 5.Carry out classification emulation with the present invention to Fu Laifulan province farmland Polarimetric SAR Image, classification results is shown in Fig. 6.
As seen from Figure 6, the Region dividing of the present invention to farmland is more careful, and edge keeps better, and nicety of grading is high.
In sum, the Classification of Polarimetric SAR Image method based on fast density peak value cluster that the present invention proposes can obtain better classification results to Classification of Polarimetric SAR Image, and can be used for classifying to various Polarimetric SAR Image.
Claims (4)
1., based on a Classification of Polarimetric SAR Image method for fast density peak value cluster, comprise the steps:
(1) filtering is carried out to Polarimetric SAR Image to be sorted, remove speckle noise, obtain filtered Polarimetric SAR Image;
(2) Yamaguchi decomposition is carried out to the coherence matrix T of pixel each in filtered Polarimetric SAR Image, obtain the volume scattering power P of each pixel
v, dihedral angle scattering power P
d, surface scattering power P
swith conveyor screw scattering component P
h;
(3) according to each pixel four scattering power P
s, P
d, P
v, P
hmaximal value be four classes by Polarimetric SAR Image initial division:
If max is (P
s, P
d, P
v, P
h)=P
s, then the pixel of its correspondence is divided into a class, wherein P
sfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
d, then the pixel of its correspondence is divided into a class, wherein P
dfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
v, then the pixel of its correspondence is divided into a class, wherein P
vfor such main scattering power;
If max is (P
s, P
d, P
v, P
h)=P
h, then the pixel of its correspondence is divided into a class, wherein P
hfor such main scattering power;
Wherein, max () represents maximal value;
(4) press ascending sequence to the main scattering power of all pixels of every class initial category, and 300 pixels every in every class are further subdivided into a class, the most whole Polarimetric SAR Image is divided into M class;
(5) in an acquired M classification, using the central point of each class as new pixel, M new pixel A is obtained
i, i=1 ..., M, and with the new pixel A of every class
irepresent all pixels in this classification;
(6) fast density peak value cluster is carried out to above-mentioned M new pixel, M new pixel is gathered for k class;
(7) in the cluster result of M new pixel, will by new pixel A
iall pixels of representative are labeled as and new pixel A
iidentical classification, completes presorting to entire image;
(8) multiple Wishart iteration is carried out to the result of presorting of whole Polarimetric SAR Image, obtain classification results more accurately.
2. the Classification of Polarimetric SAR Image method based on fast density peak value cluster according to claims 1, carries out Yamaguchi decomposition to the coherence matrix T of each pixel wherein described in step (2), carries out as follows:
(2a) read in each pixel of filtered image, these pixels are the coherence matrix T of 3 × 3, obtain covariance matrix C according to coherence matrix T;
Wherein, U is intermediate variable,
H represents horizontal polarization, and V represents vertical polarization, S
hHexpression level to launch and level to reception echo data, S
vVrepresent the echo data that Vertical dimension is launched and Vertical dimension receives, S
hVthe echo data that expression level receives to transmitting Vertical dimension, ()
*represent the conjugation of these data, <> represents average by looking number;
(2b) covariance matrix C is resolved into following expression:
Wherein, f
sfor the coefficient of dissociation of in-plane scatter component, f
dfor the coefficient of dissociation of dihedral angle scattering component, f
vfor the coefficient of dissociation of volume scattering component, f
hfor the coefficient of dissociation of spiral scattering component, β is the ratio that horizontal emission level receives back scattering reflection coefficient and Vertical Launch vertical reception back scattering emission ratio, and α is defined as α=I
hi
vH/ I
vi
vV, I
hand I
vrepresent level and the vertical reflection coefficient on earth's surface respectively, I
vHand I
vVrepresent level and the vertical reflection coefficient of vertical body of wall respectively, j represents imaginary number;
(2c) result of being decomposed by covariance matrix C, obtains one and has six unknowm coefficient f
s, f
v, f
d, f
h, the system of equations of α, β and five equations is as follows:
(2d) calculate in pixel covariance matrix C
value, if
then make α=-1, if
then make β=1, after the value of given α or β, remaining 4 unknown numbers are then according to solving equations in (2c), and wherein real part is got in Re () expression;
(2e) according to the f solved
s, f
v, f
d, f
h, α, β, solve volume scattering power P
v, dihedral angle scattering power P
d, surface scattering power P
swith spiral scattering power P
h:
P
s=f
s(1+|β|
2)
P
d=f
d(1+|a|
2)
P
v=f
v
P
h=f
h。
3. the Classification of Polarimetric SAR Image method based on fast density peak value cluster according to claims 1, carries out fast density peak value cluster to the new pixel of M wherein described in step (6), carries out as follows:
6a) calculate new pixel A
ilocal density ρ
i:
Wherein, d is worked as
ij< d
ctime, χ (d
ij-d
c)=1, otherwise χ (d
ij-d
c)=0; M represents new pixel A
inumber; d
ijrepresent the phase mutual edge distance between any two points:
d
ij=Tr((T
i)
-1+(T
j)
-1T
i)-q
Wherein, T
i, T
jrepresent new pixel A respectively
iand A
jcoherence matrix, (T
i)
-1(T
j)
-1represent respectively matrix T
iand T
jinvert, q is constant, and value is q=3, Tr () is matrix trace; d
cfor constant, its value for by phase mutual edge distance d a little
ijascending arrangement, will be positioned at the value of this arrangement 2% position as d
cvalue.
6b) calculate new pixel A
idistance δ
i:
6c) select local density ρ
iwith distance δ
ithe maximum k of product new pixel is as cluster centre;
After 6d) cluster centre is determined, each cluster centre represents a class, k cluster centre represents k class altogether, relatively remain the distance of new pixel and each class cluster centre, if this new pixel and m class cluster centre is nearest, then this new pixel is divided into m class, m=1,, k.
4. the Classification of Polarimetric SAR Image method based on fast density peak value cluster according to claims 1, carries out multiple Wishart iteration to the result of presorting of whole Polarimetric SAR Image wherein described in step (8), carries out as follows:
(8a) whole Polarimetric SAR Image is presorted the k class division result obtained, ask the cluster centre B of each class according to following formula
c:
Wherein T
ρthe coherence matrix of each pixel in c class, n
cit is the number of the pixel belonging to c class;
(8b) according to the cluster centre B of each class
c, calculate the distance d of each pixel i to c class cluster centre
ic:
Wherein T is the coherence matrix of pixel, and <> represents average by looking number, the determinant of [] representing matrix, the mark of Tr () representing matrix,
represent cluster centre B
cinvert;
(8c) distance of more each pixel and each cluster centre, if this new pixel and m class cluster centre is nearest, is then divided into m class by this pixel, m=1,, k, completes repartitioning view picture Polarimetric SAR Image classification after presorting;
(8d) repeat step (8a)-(8c), until iterations equals given iterations μ=4, obtain classification results.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510467832.9A CN105138966B (en) | 2015-08-03 | 2015-08-03 | Classification of Polarimetric SAR Image method based on fast density peak value cluster |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510467832.9A CN105138966B (en) | 2015-08-03 | 2015-08-03 | Classification of Polarimetric SAR Image method based on fast density peak value cluster |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105138966A true CN105138966A (en) | 2015-12-09 |
CN105138966B CN105138966B (en) | 2018-11-16 |
Family
ID=54724310
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510467832.9A Active CN105138966B (en) | 2015-08-03 | 2015-08-03 | Classification of Polarimetric SAR Image method based on fast density peak value cluster |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105138966B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106339416A (en) * | 2016-08-15 | 2017-01-18 | 常熟理工学院 | Grid-based data clustering method for fast researching density peaks |
CN106408939A (en) * | 2016-10-29 | 2017-02-15 | 浙江大学 | Traffic flow sequence classification method based on density peak value clustering |
CN107169492A (en) * | 2017-05-26 | 2017-09-15 | 西安电子科技大学 | Polarization SAR object detection method based on FCN CRF master-slave networks |
CN107239757A (en) * | 2017-05-23 | 2017-10-10 | 西安电子科技大学 | A kind of polarization SAR silhouette target detection method based on depth ladder net |
CN107729802A (en) * | 2017-08-18 | 2018-02-23 | 浙江大学宁波理工学院 | Face picture clustering method based on coring density peaks |
CN111665504A (en) * | 2020-05-29 | 2020-09-15 | 武汉大学 | Fully-polarized SAR target decomposition method for angle optimization processing |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102968640A (en) * | 2012-10-25 | 2013-03-13 | 西安电子科技大学 | Polarized synthetic aperture radar (SAR) image classification method based on Freeman decomposition and data distribution characteristics |
CN103186794A (en) * | 2013-03-27 | 2013-07-03 | 西安电子科技大学 | Polarized SAT (synthetic aperture radar) image classification method based on improved affinity propagation clustering |
CN104331707A (en) * | 2014-11-02 | 2015-02-04 | 西安电子科技大学 | Polarized SAR (synthetic aperture radar) image classification method based on depth PCA (principal component analysis) network and SVM (support vector machine) |
-
2015
- 2015-08-03 CN CN201510467832.9A patent/CN105138966B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102968640A (en) * | 2012-10-25 | 2013-03-13 | 西安电子科技大学 | Polarized synthetic aperture radar (SAR) image classification method based on Freeman decomposition and data distribution characteristics |
CN103186794A (en) * | 2013-03-27 | 2013-07-03 | 西安电子科技大学 | Polarized SAT (synthetic aperture radar) image classification method based on improved affinity propagation clustering |
CN104331707A (en) * | 2014-11-02 | 2015-02-04 | 西安电子科技大学 | Polarized SAR (synthetic aperture radar) image classification method based on depth PCA (principal component analysis) network and SVM (support vector machine) |
Non-Patent Citations (1)
Title |
---|
SHANE ROBERT CLOUDE: "An Entropy Based Classification Scheme for Land Applications of Polarimetric SAR", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106339416A (en) * | 2016-08-15 | 2017-01-18 | 常熟理工学院 | Grid-based data clustering method for fast researching density peaks |
CN106339416B (en) * | 2016-08-15 | 2019-11-08 | 常熟理工学院 | Educational data clustering method based on grid fast searching density peaks |
CN106408939A (en) * | 2016-10-29 | 2017-02-15 | 浙江大学 | Traffic flow sequence classification method based on density peak value clustering |
CN106408939B (en) * | 2016-10-29 | 2019-05-31 | 浙江大学 | Magnitude of traffic flow sequence division methods based on density peaks cluster |
CN107239757A (en) * | 2017-05-23 | 2017-10-10 | 西安电子科技大学 | A kind of polarization SAR silhouette target detection method based on depth ladder net |
CN107239757B (en) * | 2017-05-23 | 2020-12-08 | 西安电子科技大学 | Polarized SAR image target detection method based on deep ladder network |
CN107169492A (en) * | 2017-05-26 | 2017-09-15 | 西安电子科技大学 | Polarization SAR object detection method based on FCN CRF master-slave networks |
CN107169492B (en) * | 2017-05-26 | 2020-07-07 | 西安电子科技大学 | Polarized SAR target detection method based on FCN-CRF master-slave network |
CN107729802A (en) * | 2017-08-18 | 2018-02-23 | 浙江大学宁波理工学院 | Face picture clustering method based on coring density peaks |
CN111665504A (en) * | 2020-05-29 | 2020-09-15 | 武汉大学 | Fully-polarized SAR target decomposition method for angle optimization processing |
CN111665504B (en) * | 2020-05-29 | 2022-02-01 | 武汉大学 | Fully-polarized SAR target decomposition method for angle optimization processing |
Also Published As
Publication number | Publication date |
---|---|
CN105138966B (en) | 2018-11-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107563428B (en) | Based on the Classification of Polarimetric SAR Image method for generating confrontation network | |
CN102208031B (en) | Freeman decomposition and homo-polarization rate-based polarized synthetic aperture radar (SAR) image classification method | |
CN105138966A (en) | Quick density peak value clustering based polarimetric SAR image classification method | |
CN102982338B (en) | Classification of Polarimetric SAR Image method based on spectral clustering | |
CN102999762B (en) | Decompose and the Classification of Polarimetric SAR Image method of spectral clustering based on Freeman | |
CN104123555B (en) | Super-pixel polarimetric SAR land feature classification method based on sparse representation | |
CN103186794B (en) | Based on the Classification of Polarimetric SAR Image method of the neighbour's propagation clustering improved | |
CN102968640B (en) | Decompose and the Classification of Polarimetric SAR Image method of data distribution characteristics based on Freeman | |
CN105138970A (en) | Spatial information-based polarization SAR image classification method | |
CN105930772A (en) | City impervious surface extraction method based on fusion of SAR image and optical remote sensing image | |
CN108734171A (en) | A kind of SAR remote sensing image ocean floating raft recognition methods of depth collaboration sparse coding network | |
CN103886336A (en) | Polarized SAR image classifying method based on sparse automatic encoder | |
CN105868793A (en) | Polarization SAR image classification method based on multi-scale depth filter | |
CN105005789B (en) | A kind of remote sensing images terrain classification method of view-based access control model vocabulary | |
CN108256436A (en) | A kind of radar HRRP target identification methods based on joint classification | |
CN107292336A (en) | A kind of Classification of Polarimetric SAR Image method based on DCGAN | |
CN107358203A (en) | A kind of High Resolution SAR image classification method based on depth convolution ladder network | |
CN103839073A (en) | Polarization SAR image classification method based on polarization features and affinity propagation clustering | |
CN104463193A (en) | Polarization SAR image classifying method based on depth sparsity ICA | |
CN104732244A (en) | Wavelet transform, multi-strategy PSO (particle swarm optimization) and SVM (support vector machine) integrated based remote sensing image classification method | |
CN105760900A (en) | Hyperspectral image classification method based on affinity propagation clustering and sparse multiple kernel learning | |
CN107895139A (en) | A kind of SAR image target recognition method based on multi-feature fusion | |
CN104318246A (en) | Depth self-adaption ridgelet network based polarimetric SAR (Synthetic Aperture Radar) image classification | |
CN107330457B (en) | A kind of Classification of Polarimetric SAR Image method based on multi-feature fusion | |
CN105913081A (en) | Improved PCAnet-based SAR image classification method |
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 |