CN107203791A - Based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy - Google Patents
Based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy Download PDFInfo
- Publication number
- CN107203791A CN107203791A CN201710522622.4A CN201710522622A CN107203791A CN 107203791 A CN107203791 A CN 107203791A CN 201710522622 A CN201710522622 A CN 201710522622A CN 107203791 A CN107203791 A CN 107203791A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- mtd
- pixel
- msup
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Medicines Containing Antibodies Or Antigens For Use As Internal Diagnostic Agents (AREA)
- Image Analysis (AREA)
Abstract
The present invention proposes a kind of based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, the technical problem low for solving nicety of grading present in existing unsupervised Classification of Polarimetric SAR Image method.Realize that step is:Remove the coherent speckle noise in Polarimetric SAR Image to be sorted;Freeman decomposition is carried out to Polarimetric SAR Image, three kinds of scattered powers of image are obtained;The scattered power entropy of image is calculated according to three kinds of scattered powers;Polarimetric SAR Image initial division it is 7 classes using three kinds of scattered powers and scattered power entropy;Calculate the ratio between the co-polarization and cross polar component of each pixel in Polarimetric SAR Image:Heteropolar ratio;Using the ratio to the subdivision of each class progress in proportion in 7 class polarization SAR data;Classification results are merged based on merging criterion between specific class;Multiple Wishart iteration is carried out to the result after merging and is painted, final color classification figure is obtained.
Description
Technical field
The invention belongs to image data processing technology field, it is related to a kind of Classification of Polarimetric SAR Image method, and in particular to one
Plant based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, the terrain classification available for Polarimetric SAR Image.
Background technology
Polarimetric synthetic aperture radar is a kind of multi-parameter, the radar imaging system of multichannel, due to its stronger target letter
Acquisition capability is ceased, military in agricultural, the field such as ocean is widely studied and applied, and has become synthetic aperture radar
One of important directions of development, Classification of Polarimetric SAR Image is as one of important application of polarimetric synthetic aperture radar, by more next
More concerns.Classification of Polarimetric SAR Image is the polarization measurement data provided using airborne or borne polarization SAR system, to figure
Each pixel as in carries out category division, so as to realize the classification of Polarimetric SAR Image.Existing Classification of Polarimetric SAR Image side
Method mainly includes unsupervised segmentation method and supervised classification method, but supervised classification method needs substantial amounts of training sample,
And nicety of grading is easily influenceed by training data.When carrying out terrain classification to image, in many cases, we are not
Know the True Data of earth's surface, then artificial selection training sample is highly difficult for Polarimetric SAR Image.Therefore, mesh
Preceding terrain classification the characteristics of focusing on unsupervised segmentation, unsupervised segmentation does not need training sample exactly, it is not necessary to know
The True Data distribution of atural object, makes full use of the data message included in image to classify image in image.Compare through
The method of the Polarimetric SAR Image unsupervised segmentation of allusion quotation has:Freeman propose three-component scattering model and Lee propose based on
Sorting technique of H/ α goal decompositions and Wishart graders etc., on the basis of these classical sorting algorithms, first classifies, then
The classification policy of merging is suggested, including:
2004, J.S.Lee et al. decomposed the polarization SAR for proposing a kind of holding target scattering characteristics based on Freeman
Image unsupervised segmentation algorithm, is shown in J.S.Lee, M.R.Grunes, E.Pottier and L.Ferro-Famil, "
Unsupervised terrain classification preserving polarimetric scattering
characteristics,"in IEEE Transactions on Geoscience and Remote Sensing,
Vol.42, no.4, pp.722-731, April2004. this method on the basis of Freeman scattering models, by picture breakdown into
Surface scattering class, dihedral angle scattering class and volume scattering class three major types, then according to the scattering work(of each pixel in each class data
Each class is subdivided into 30 classes or more by the size of rate, carries out categories combination and Wishart iteration optimizations to data afterwards.Should
Method combination Freeman scattering models and multiple Wishart iteration, with the holding pure property of the main scattering mechanism of multilevel configuration
Characteristic, but this method do not account in Polarimetric SAR Image mix scattering mechanism presence, therefore, nicety of grading still has
Wait to improve.
2007, Cao et al. proposed the polarization of the adaptive classification based on SPAN/H/ α/A and multiple Wishart algorithms
The sorting technique of SAR image, is shown in F.Cao, W.Hong, Y.Wu, and E.Pottier. " An Unsupervised
segmentation with an adaptive number of clusters using the SPAN/H/α/A space
and the complex Wishart clustering for fully polarimetric SAR data analysis.”
IEEE Trans.Geosci.Remote Sensing, vol.45, no.8, pp.3454-3467, Nov.2007. this method are combined
Polarimetric SAR Image, is divided into 48 classes, afterwards using Wishart by the back scattering power SPAN of data and H/ α/A information of data
The method of test statisticses realizes the Agglomerative hierarchical clustering of data, is assessed data come automatic using the likelihood probability of data and is drawn
The suitable class number divided.This method can carry out the division of adaptive classification to polarization SAR data, and the robustness of method is obtained
Ensure, but this method is still without the presence in view of mixing scattering mechanism in Polarimetric SAR Image, moreover, being used in this method
Agglomerative hierarchical clustering and every time cluster are assessed the likelihood that data are carried out, and cause the computation complexity of whole method higher.
2013, Wang et al. proposed the sorting technique based on scattered power entropy and same polarization ratio, sees S.Wang, K.Liu,
J.Pei,M.Gong and Y.Liu,"Unsupervised Classification of Fully Polarimetric SAR
Images Based on Scattering Power Entropy and Copolarized Ratio,"in IEEE
This method of Geoscience and Remote Sensing Letters, vol.10, no.3, pp.622-626, May 2013.
On the basis of scattered power entropy, picture breakdown is scattered into class, volume scattering class, surface-dihedral angle into surface scattering class, dihedral angle
Class, surface-volume scattering scattering class, dihedral angle-volume scattering class, and the mixing major class of class seven are scattered, the same polarization of data is then utilized
Each major class is divided into three groups than feature, finally classification results optimized using multiple Wishart alternative manners.This method
The presence that scattering mechanism is mixed in pixel is take into account, but the same polarization used is than having related only in Polarimetric SAR Image
Co-polarization have ignored the cross polar component of Polarimetric SAR Image than component, cause partial target mistake point occur, or classify not smart
Thin the problem of.
The content of the invention
It is an object of the invention to overcome the shortcomings of above-mentioned prior art, it is proposed that one kind is based on heteropolar ratio and scattering work(
The Classification of Polarimetric SAR Image method of rate entropy, to realize the more accurate terrain classification of Polarimetric SAR Image.
The present invention technical thought be:First on the basis of Freeman decomposition, using scattered power entropy to polarization SAR
Image carries out preliminary classification;Then according between the co-polarization component and cross polar component of each pixel in Polarimetric SAR Image
Ratio, each class to image initialization point makees further divide;Merge afterwards using between the class based on Wishart distances
Criterion is merged to above-mentioned classification results, and carries out Wishart iteration to amalgamation result;Finally image is painted, coloured silk is obtained
The classification results figure of color, implementing step is:
(1) coherent speckle noise in Polarimetric SAR Image to be sorted is removed, filtered Polarimetric SAR Image is obtained;
(2) three kinds of scattered powers of each pixel in Polarimetric SAR Image are obtained:Filtered Polarimetric SAR Image is carried out
Freeman is decomposed, and obtains the surface scattering power P of each pixels, dihedral angle scattered power PdWith volume scattering power Pv;
(3) using three kinds of scattered powers of each pixel, the scattered power entropy H of each pixel is calculated respectivelyp, wherein
0≤Hp≤1;
(4) preliminary classification is carried out to Polarimetric SAR Image:According to three kinds of scattering work(of each pixel in Polarimetric SAR Image
Rate and scattered power entropy Hp, preliminary classification is carried out to Polarimetric SAR Image, 7 class Polarimetric SAR Images are obtained:
(4a) is according to the scattered power entropy H of Polarimetric SAR ImagepSelect two classification thresholds x1And x2;
(4b) is according to scattered power entropy HpWith two classification thresholds x1And x2Relation, Polarimetric SAR Image is initially divided
Class, be specially:
As 0≤Hp≤x1, P will be mets=max (Ps,Pd,Pv) pixel be divided into surface scattering class, meet Pd=max
(Ps,Pd,Pv) pixel be divided into dihedral angle scattering class, meet Pv=max (Ps,Pd,Pv) pixel be divided into volume scattering
Class;
Work as x1< Hp≤x2, P will be mets=min (Ps,Pd,Pv) pixel be divided into dihedral angle-volume scattering mixing class,
Meet Pd=min (Ps,Pd,Pv) pixel be divided into surface-volume scattering mixing class, meet Pv=min (Ps,Pd,Pv) pixel
Point is divided into surface-dihedral angle scattering mixing class;
Work as x2< Hp≤ 1, then such pixel is divided into scattering mixing class;
(5) definition is heteropolar than formula, and utilizes the co-polarization component of each pixel in formula calculating Polarimetric SAR Image
With the ratio of cross polar component, the heteropolar ratio of each pixel is obtained, the heteropolar score of Polarimetric SAR Image is then counted
Cloth;
(6) subseries again is carried out to Polarimetric SAR Image:According to the heteropolar ratio of each pixel in Polarimetric SAR Image and pole
Change the heteropolar than distribution of SAR image, in proportion thin is carried out to each class in 7 class Polarimetric SAR Images in step (4)
Point, 7n class Polarimetric SAR Images are obtained, wherein, n represents the classification number of each class data subdividing, and n >=2;
(7) categories combination is carried out to the 7n classes Polarimetric SAR Image of acquisition, obtains m class Polarimetric SAR Images, wherein, m value
It is the ground species number that includes in Polarimetric SAR Image to determine;
(8) classification belonging to each pixel in m class Polarimetric SAR Images is updated, classification results are obtained;
(9) step (8) is given according to the principle of three primary colours as three primary colours with red R, green G and tri- color components of blueness B
In obtain classification results colouring, obtain final color classification result figure.
The present invention compared with prior art, with advantages below:
The present invention carries out preliminary classification using scattered power entropy on the basis of Freeman decomposition to Polarimetric SAR Image,
On the basis of preliminary classification, the polarization information included in Polarimetric SAR Image is made full use of:Co-polarization component and cross polarization point
Amount, according to the log-of-ratio of two kinds of polarization components according to more meticulously being divided.Because scattered power entropy information can be effectively
The scattering type of each pixel is determined, co-polarization component can preferably characterize different atural objects from the ratio of cross polar component
Between difference, be conducive to distinguishing different types of ground objects, therefore the combination of the ratio and scattered power entropy of two kinds of polarization components
The precision of Classification of Polarimetric SAR Image can be effectively improved.
Brief description of the drawings
Fig. 1 is the implementation process block diagram of the present invention;
Fig. 2 is the Polarimetric SAR Image San Francisco images that present invention emulation is used;
Fig. 3 is that the heteropolar of major surface features compares distribution map in present invention emulation Polarimetric SAR Image used;
Fig. 4 is of the invention and the existing sorting technique based on scattered power entropy and same polarization ratio is to San Francisco
The classification simulation result figure of image.
Embodiment
Below in conjunction with the drawings and specific embodiments, the present invention is described in further detail.
Reference picture 1, the Classification of Polarimetric SAR Image method based on heteropolar ratio and scattered power entropy, comprise the following steps:
Step 1) coherent speckle noise in San Francisco images is removed using exquisiteness Lee filtering methods, filtered
San Francisco images afterwards;
Step 2) Freeman decomposition is carried out to filtered San Francisco images, the surface for obtaining each pixel dissipates
Penetrate power Ps, dihedral angle scattered power PdWith volume scattering power Pv, realize that step is:
The covariance matrix C of each pixel in (2a) input San Francisco images:
Wherein, H represents horizontal polarization, and V represents vertical polarization, SHHRepresent the echo that horizontal emission level is received, SVV
Represent the echo of Vertical Launch vertical reception, SHVRepresent the echo of horizontal emission vertical reception, SHHAnd SVVIt is referred to as
Co-polarization component, SHVIt is referred to as cross polar component, * represents conjugation, | | Modulus of access is represented,<·>Represent to press and regard number averagely;
Covariance matrix C is resolved into following form by (2b):
Wherein, fsFor the decomposition coefficient of surface scattering component, fdFor the decomposition coefficient of dihedral angle scattering component, fvDissipated for body
The decomposition coefficient of component is penetrated, β represents that horizontal emission level receives back scattering reflectance factor and Vertical Launch vertical reception is backward
The ratio of scattered reflection coefficient, α is defined as α=RghRvh/RgvRvv, Rgh, RvhThe horizontal and vertical reflection system of earth's surface is represented respectively
Number, Rgv, RvvThe horizontal and vertical reflectance factor of vertical wall is represented respectively;
(2c) is according to formula 1) and formula 2) corresponding element is equal, draw containing five unknown number fs, fd, fv, the four of α and β
Individual equation:
(2d) solves equation group 3), obtain fs, fd, fv, α and β value:
Calculate in each pixel covariance matrixValue and judge positive and negative, ifMake α
=-1, otherwise makes β=1, and equation group 3 is solved according to α and β value), and then obtain fs, fdAnd fvValue, wherein Re () represent
Take real part;
(2e) is according to fs, fd, fv, α and β value obtains three kinds of scattered power P of each pixel using following formulas, PdAnd Pv:
Wherein, Ps, PdAnd PvSurface scattering power, dihedral angle scattered power and volume scattering power are represented respectively.
Step 3) using three kinds of scattered powers of each pixel, the scattered power entropy H of each pixel is calculated respectivelyp,
Wherein 0≤Hp≤1;
(3a) calculates the surface scattering power P of each pixels, dihedral angle scattered power PdAnd volume scattering power PvTotal
Proportion in scattered power, calculation formula is:
Wherein, P1, P2And P3Surface scattering power is represented respectively, and dihedral angle scattered power and volume scattering power dissipate always
Penetrate proportion in power;
(3b) utilizes formula 6) calculate the scattered power entropy H of each pixel in San Francisco imagesp
Step 4) preliminary classification is carried out to San Francisco images:According to each pixel in San Francisco images
The three kinds of scattered powers and scattered power entropy H of pointp, San Francisco image initials are divided into 7 classes;
Scattered power entropy is the description to scattering mechanism randomness, and its span is:0≤Hp≤ 1, work as HpWhen=0, number
A kind of only unique scattering mechanism, works as H inpWhen=1, data are the extreme case of three kinds of scattering mechanism mixing, as 0 < Hp<
When 1, with the increase of scattered power entropy, the randomness of scattering process also gradually increases, when scattered power entropy is smaller, in data
Only a kind of main scattering mechanism, when scattered power entropy is larger, has two kinds or three kinds of main scattering mechanisms in data, therefore,
Scattered power entropy may be used to determine the scattering mechanism of pixel, and image is classified according to the scattering mechanism of pixel,
Realize that step is:
(4a) is according to the scattered power entropy H of Polarimetric SAR ImagepSelect two classification thresholds x1=0.48 and x2=0.85;
(4b) is according to scattered power entropy HpWith two classification thresholds x1And x2Relation, Polarimetric SAR Image is initially divided
Class, be specially:
As 0 < Hp≤x1, P will be mets=max (Ps,Pd,Pv) pixel be divided into surface scattering class, meet Pd=max
(Ps,Pd,Pv) pixel be divided into dihedral angle scattering class, meet Pv=max (Ps,Pd,Pv) pixel be divided into volume scattering
Class;
Work as x1< Hp≤x2, P will be mets=min (Ps,Pd,Pv) pixel be divided into dihedral angle-volume scattering mixing class,
Meet Pd=min (Ps,Pd,Pv) pixel be divided into surface-volume scattering mixing class, meet Pv=min (Ps,Pd,Pv) pixel
Point is divided into surface-dihedral angle scattering mixing class;
Work as x2< Hp≤ 1, then such pixel is divided into scattering mixing class;
Step 5) definition is heteropolar than formula, and calculate each pixel in San Francisco images using the formula
The ratio of co-polarization component and cross polar component, obtains the heteropolar ratio of each pixel;
Ratio between heteropolar co-polarization component and cross polar component than representing Polarimetric SAR Image, can be altogether
The ratio of the ratio or cross polar component of polarization components and cross polar component polarization components together.Either which
A kind of ratio is all without the heteropolar than being distributed of influence data.Due to not only including the copolar of Polarimetric SAR Image in heteropolar ratio
Change component, the also cross polar component comprising image, it is more abundant to the Information Pull in Polarimetric SAR Image, therefore, use heteropole
The classifying quality of data can more be improved to be classified to image by changing ratio.Here with co-polarization component and the ratio of cross polar component
Exemplified by value, that is, the calculation formula for defining heteropolar ratio is:
Wherein SHHRepresent the echo that horizontal emission level is received, SVVRepresent the reflectogram of Vertical Launch vertical reception
Picture, SHHAnd SVVIt is referred to as co-polarization component, SHVRepresent the echo of horizontal emission vertical reception, SVHRepresent Vertical Launch water
Flush the echo of receipts, SHVAnd SVHIt is referred to as cross polar component, | | represent to take the modulus value of the number.
Step 6) subseries again is carried out to San Francisco images:According to each pixel in San Francisco images
It is heteropolar than and San Francisco images it is heteropolar than distribution, to step 4) in 7 class San Francisco images
In each class carry out subdivision in proportion.San Francisco images are mainly comprising 3 kinds of different atural objects:Ocean, building and plant
Quilt, therefore we are by step 4) in 7 class San Francisco images in each class data be divided into 3 classes again, obtain 21
The San Francisco images of class, realize that step is:
(6a) selects two classification thresholds y according to the heteropolar than being distributed of San Francisco images1=6, y2=13;
(6b) compares R according to the heteropolar of each pixel in San Francisco imagesdWith two classification thresholds y1And y2Pass
System, step 4) in 7 class San Francisco images in each class be subdivided into 3 classes, be specially:R will be metd≤y1Picture
Vegetarian refreshments is divided into a class, meets y1< Rd≤y2Pixel be divided into a class, meet Rd≤y2Pixel be divided into a class;
Step 7) the San Francisco images of 21 classes are merged using merging criterion between the class based on Wishart distances
Into the San Francisco images of 8 classes, realize that step is:
(7a) calculates the cluster centre per class data, and calculation formula is:
Wherein, i=1,2 ... 21NiRepresent the number of pixel in the i-th class data, ClRepresent in the i-th class data l-th
The covariance matrix of pixel;
(7b) calculates the Wishart distances between every two classes data, and calculation formula is:
Dij=(Ni+Nj)ln|V|-Ni ln|Vi|-Njln|Vj| 9)
Wherein, i, j=1,2 ... 21, i ≠ j, V represent the cluster centre after two classes merging;
Two minimum class data of (7c) combined distance;
(7d) repeat step (7a)-(7c), until total classification number of Polarimetric SAR Image is 8.
Step 8) 8 class data in San Francisco images are carried out with multiple Wishart iteration, update each pixel
Affiliated classification, obtains more accurate classification results:
(8a) sets primary iteration number of times t=1, and maximum iteration is 4;
(8b) utilize formula 8) calculate merge after each class data cluster centre;
(8c) calculates each pixel to the Wishart distances at the i-th class data clusters center, and calculation formula is:
d(C,Vi)=ln | Vi|+tr(Vi -1C) 10)
Wherein, i=1,2 ... 8, C be the covariance matrix of pixel, the mark of tr representing matrixs, Vi -1Represent ViMatrix
Inverse matrix;
(8d) is reclassified according to the Wishart distances of each pixel to each cluster centre to pixel:If
Pixel meets d (C, Vi)≤d(C,Vj), the pixel is then divided into the i-th class by i, j=1,2 ..., 8, i ≠ j;
(8e) iterations increases by 1, i.e. t=t+1;
(8f) repeat step (8b)-(8e), is 4 to iterations is arrived.
Step 9) red R, green G and blueness tri- color components of B are used as three primary colours, according to the principle of three primary colours to step
8) the classification results colouring obtained in, obtains final color classification result figure.Colouring principle is generally:Waters part blueness
Adjust, culture's red color tone, vegetation part green hue.
Below in conjunction with emulation experiment, the technique effect of the present invention is described further.
1. simulated conditions and method:
Simulated environment:Intel(R)Core(TM)i5-3210M CPU@2.50GHz 2.50GHz Windows10;
Software platform:Matlab2015a;
Emulation mode:The present invention and the sorting technique based on scattered power entropy and same polarization ratio are completed to Polarimetric SAR Image
Classification emulation experiment.
2. emulation content and interpretation of result:
Experiment content:The present invention carries out emulation experiment using the San Francisco images shown in Fig. 2, and the image is regarded
Number is 4, and size is 900 × 1024 pixels;
Different atural objects is heteropolar than distribution, San Francisco figures in experiment one, analysis San Francisco images
It is main as in include the region 1 marked in three kinds of types of ground objects, such as Fig. 2, region 2 and region 3, ocean is represented respectively, is built, and is planted
Fig. 3 is shown in quilt, their heteropolar ratio distribution.
As can be seen from Figure 3 there is larger difference in heteropolar between three kinds of different atural objects than being distributed, therefore we can
It is different heteropolar than classification thresholds to set, to distinguish different types of ground objects.
Experiment two, is schemed with the present invention and the sorting technique based on scattered power entropy and same polarization ratio to San Francisco
As carrying out classification emulation experiment, classification results are shown in that Fig. 4, wherein Fig. 4 (a) are the classification side based on scattered power entropy and same polarization ratio
The classification results of method, Fig. 4 (b) is classification results of the invention.
From Fig. 4 (a) as can be seen that the sorting technique based on scattered power entropy and same polarization ratio, although can substantially distinguish figure
The different type of ground objects as in, but the division to sea area is still not careful enough, there is wrong point of phenomenon in some areas.
From Fig. 4 (b) as can be seen that the classification results of the present invention are compared with Fig. 4 (a), classification results are more accurate, and atural object is thin
It is more fine that section embodies, and the classifying edge of different zones is also smoother, for example, sea area division is more careful, builds
The classification results for building area are more nearly true atural object, and golf course, the uniformity in the region such as racecourse and parking lot is bright
The aobvious classification results for being better than Fig. 4 (a).
In summary, the sorting technique proposed by the present invention for Polarimetric SAR Image, by using Polarimetric SAR Image
Three kinds of scattered powers and scattered power entropy carry out preliminary classification to image, utilize co-polarization component and intersection in Polarimetric SAR Image
The ratio of polarization components carries out subseries again to image, and afterwards classification results are carried out with categories combination and multiple Wishart iteration, is entered
One step improves the classifying quality of data, different types of ground objects can be more accurately distinguished, with more preferable classification performance.
Claims (7)
1. it is a kind of based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, comprise the following steps:
(1) coherent speckle noise in Polarimetric SAR Image to be sorted is removed, filtered Polarimetric SAR Image is obtained;
(2) three kinds of scattered powers of each pixel in Polarimetric SAR Image are obtained:Filtered Polarimetric SAR Image is carried out
Freeman is decomposed, and obtains the surface scattering power P of each pixels, dihedral angle scattered power PdWith volume scattering power Pv;
(3) using three kinds of scattered powers of each pixel, the scattered power entropy H of each pixel is calculated respectivelyp, wherein 0≤Hp
≤1;
(4) preliminary classification is carried out to Polarimetric SAR Image:According to three kinds of scattered powers of each pixel in Polarimetric SAR Image and
Scattered power entropy Hp, preliminary classification is carried out to Polarimetric SAR Image, 7 class Polarimetric SAR Images are obtained:
(4a) is according to the scattered power entropy H of Polarimetric SAR ImagepSelect two classification thresholds x1And x2;
(4b) is according to scattered power entropy HpWith two classification thresholds x1And x2Relation, to Polarimetric SAR Image carry out preliminary classification,
Specially:
As 0≤Hp≤x1, P will be mets=max (Ps,Pd,Pv) pixel be divided into surface scattering class, meet Pd=max (Ps,
Pd,Pv) pixel be divided into dihedral angle scattering class, meet Pv=max (Ps,Pd,Pv) pixel be divided into volume scattering class;
Work as x1< Hp≤x2, P will be mets=min (Ps,Pd,Pv) pixel be divided into dihedral angle-volume scattering mixing class, meet Pd
=min (Ps,Pd,Pv) pixel be divided into surface-volume scattering mixing class, meet Pv=min (Ps,Pd,Pv) pixel dot-dash
It is divided into surface-dihedral angle scattering mixing class;
Work as x2< Hp≤ 1, then such pixel is divided into scattering mixing class;
(5) definition is heteropolar than formula, and calculates the co-polarization component of each pixel and friendship in Polarimetric SAR Image using the formula
The ratio of polarization components is pitched, the heteropolar ratio of each pixel is obtained, the heteropolar than distribution of Polarimetric SAR Image is then counted;
(6) subseries again is carried out to Polarimetric SAR Image:According to the heteropolar ratio and polarization SAR of each pixel in Polarimetric SAR Image
The heteropolar ratio distribution of image, subdivision in proportion is carried out to each class in 7 class Polarimetric SAR Images in step (4), is obtained
7n class Polarimetric SAR Images, wherein, n represents the classification number of each class data subdividing, and n >=2;
(7) categories combination is carried out to the 7n classes Polarimetric SAR Image of acquisition, obtains m class Polarimetric SAR Images, wherein, m value be by
The ground species number that is included in Polarimetric SAR Image is determined;
(8) classification belonging to each pixel in m class Polarimetric SAR Images is updated, classification results are obtained;
(9) obtained with red R, green G and tri- color components of blueness B as three primary colours according to the principle of three primary colours in step (8)
The classification results colouring arrived, obtains final color classification result figure.
2. it is according to claim 1 based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, its feature
It is that the coherent speckle noise in the Polarimetric SAR Image to be sorted of the removal described in step (1) uses exquisite Lee filtering
Method.
3. it is according to claim 1 based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, its feature
It is, Freeman decomposition is carried out to filtered Polarimetric SAR Image in step (2), realizes that step is:
(2a) inputs the covariance matrix C of each pixel in Polarimetric SAR Image to be sorted:
<mrow>
<mi>C</mi>
<mo>=</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mrow>
<mo><</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>></mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msqrt>
<mn>2</mn>
</msqrt>
<mo><</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
</msub>
<msubsup>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
<mo>*</mo>
</msubsup>
<mo>></mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mo><</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
</msub>
<msubsup>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
<mo>*</mo>
</msubsup>
<mo>></mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msqrt>
<mn>2</mn>
</msqrt>
<mo><</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
</msub>
<msubsup>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
<mo>*</mo>
</msubsup>
<mo>></mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mn>2</mn>
<mo><</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>></mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msqrt>
<mn>2</mn>
</msqrt>
<mo><</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
</msub>
<msubsup>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
<mo>*</mo>
</msubsup>
<mo>></mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo><</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
</msub>
<msubsup>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
<mo>*</mo>
</msubsup>
<mo>></mo>
</mrow>
</mtd>
<mtd>
<mrow>
<msqrt>
<mn>2</mn>
</msqrt>
<mo><</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
</msub>
<msubsup>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
<mo>*</mo>
</msubsup>
<mo>></mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mo><</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>></mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
Wherein, H represents horizontal polarization, and V represents vertical polarization, SHHRepresent the echo that horizontal emission level is received, SVVRepresent
The echo of Vertical Launch vertical reception, SHVRepresent the echo of horizontal emission vertical reception, SHHAnd SVVIt is referred to as copolar
Change component, SHVIt is referred to as cross polar component, * represents conjugation, | | Modulus of access is represented,<·>Represent to press and regard number averagely;
Covariance matrix C is resolved into following form by (2b):
<mrow>
<mi>C</mi>
<mo>=</mo>
<msub>
<mi>f</mi>
<mi>s</mi>
</msub>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mrow>
<mo>|</mo>
<mi>&beta;</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mi>&beta;</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
</mtr>
<mtr>
<mtd>
<msup>
<mi>&beta;</mi>
<mo>*</mo>
</msup>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>d</mi>
</msub>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mrow>
<mo>|</mo>
<mi>&alpha;</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mi>&alpha;</mi>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
</mtr>
<mtr>
<mtd>
<mi>&alpha;</mi>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>v</mi>
</msub>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mrow>
<mn>1</mn>
<mo>/</mo>
<mn>3</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mrow>
<mn>2</mn>
<mo>/</mo>
<mn>3</mn>
</mrow>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mn>1</mn>
<mo>/</mo>
<mn>3</mn>
</mrow>
</mtd>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mn>1</mn>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
Wherein, fsFor the decomposition coefficient of surface scattering component, fdFor the decomposition coefficient of dihedral angle scattering component, fvFor volume scattering point
The decomposition coefficient of amount, β represents that horizontal emission level receives back scattering reflectance factor and Vertical Launch vertical reception back scattering
The ratio of reflectance factor, α is defined as α=RghRvh/RgvRvv, Rgh, RvhThe horizontal and vertical reflectance factor of earth's surface is represented respectively,
Rgv, RvvThe horizontal and vertical reflectance factor of vertical wall is represented respectively;
(2c) is according to formula 1) and formula 2) corresponding element is equal, draw containing five unknown number fs, fd, fv, α and β four sides
Journey:
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mo><</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>></mo>
<mo>=</mo>
<msub>
<mi>f</mi>
<mi>s</mi>
</msub>
<mo>|</mo>
<mi>&beta;</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>d</mi>
</msub>
<mo>|</mo>
<mi>&alpha;</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>v</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo><</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>></mo>
<mo>=</mo>
<msub>
<mi>f</mi>
<mi>s</mi>
</msub>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>d</mi>
</msub>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>v</mi>
</msub>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo><</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
</msub>
<msubsup>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
<mo>*</mo>
</msubsup>
<mo>></mo>
<mo>=</mo>
<msub>
<mi>f</mi>
<mi>s</mi>
</msub>
<mi>&beta;</mi>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>d</mi>
</msub>
<mi>&alpha;</mi>
<mo>+</mo>
<msub>
<mi>f</mi>
<mi>v</mi>
</msub>
<mo>/</mo>
<mn>3</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo><</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>></mo>
<mo>=</mo>
<msub>
<mi>f</mi>
<mi>v</mi>
</msub>
<mo>/</mo>
<mn>3</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
(2d) solves equation group 3), obtain fs, fd, fv, α and β value:
Calculate in each pixel covariance matrixValue and judge positive and negative, ifα=- 1 is made,
Otherwise β=1 is made, equation group 3 is solved according to α and β value), and then obtain fs, fdAnd fvValue, wherein Re () represent take reality
Portion;
(2e) is according to fs, fd, fv, α and β value obtains three kinds of scattered power P of each pixel using following formulas, PdAnd Pv:
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>P</mi>
<mi>s</mi>
</msub>
<mo>=</mo>
<msub>
<mi>f</mi>
<mi>s</mi>
</msub>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>+</mo>
<mo>|</mo>
<mi>&beta;</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>P</mi>
<mi>d</mi>
</msub>
<mo>=</mo>
<msub>
<mi>f</mi>
<mi>d</mi>
</msub>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>+</mo>
<mo>|</mo>
<mi>&alpha;</mi>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msub>
<mi>P</mi>
<mi>v</mi>
</msub>
<mo>=</mo>
<mn>8</mn>
<msub>
<mi>f</mi>
<mi>v</mi>
</msub>
<mo>/</mo>
<mn>3</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
Wherein, Ps, PdAnd PvSurface scattering power, dihedral angle scattered power and volume scattering power are represented respectively.
4. it is according to claim 1 based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, its feature
It is, using three kinds of scattered powers of each pixel in step (3), the scattered power entropy H of each pixel is calculated respectivelyp,
Realize that step is:
(3a) calculates the surface scattering power P of each pixels, dihedral angle scattered power PdAnd volume scattering power PvIn total scattering
Proportion in power, calculation formula is:
<mrow>
<msub>
<mi>P</mi>
<mn>1</mn>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>P</mi>
<mi>s</mi>
</msub>
<mrow>
<msub>
<mi>P</mi>
<mi>s</mi>
</msub>
<mo>+</mo>
<msub>
<mi>P</mi>
<mi>d</mi>
</msub>
<mo>+</mo>
<msub>
<mi>P</mi>
<mi>v</mi>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
<msub>
<mi>P</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>P</mi>
<mi>d</mi>
</msub>
<mrow>
<msub>
<mi>P</mi>
<mi>s</mi>
</msub>
<mo>+</mo>
<msub>
<mi>P</mi>
<mi>d</mi>
</msub>
<mo>+</mo>
<msub>
<mi>P</mi>
<mi>v</mi>
</msub>
</mrow>
</mfrac>
<mo>,</mo>
<msub>
<mi>P</mi>
<mn>3</mn>
</msub>
<mo>=</mo>
<mfrac>
<msub>
<mi>P</mi>
<mi>v</mi>
</msub>
<mrow>
<msub>
<mi>P</mi>
<mi>s</mi>
</msub>
<mo>+</mo>
<msub>
<mi>P</mi>
<mi>d</mi>
</msub>
<mo>+</mo>
<msub>
<mi>P</mi>
<mi>v</mi>
</msub>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
Wherein, P1, P2And P3Surface scattering power is represented respectively, and dihedral angle scattered power and volume scattering power are in total scattering work(
Proportion in rate;
(3b) utilizes formula 6) calculate the scattered power entropy H of each pixel in Polarimetric SAR Imagep
<mrow>
<msub>
<mi>H</mi>
<mi>p</mi>
</msub>
<mo>=</mo>
<mo>-</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mn>3</mn>
</munderover>
<msub>
<mi>P</mi>
<mi>i</mi>
</msub>
<msub>
<mi>log</mi>
<mn>3</mn>
</msub>
<msub>
<mi>P</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mn>0</mn>
<mo>&le;</mo>
<msub>
<mi>H</mi>
<mi>p</mi>
</msub>
<mo>&le;</mo>
<mn>1</mn>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
5. it is according to claim 1 based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, its feature
It is, the co-polarization component and the ratio of cross polar component of each pixel in Polarimetric SAR Image is calculated in step (5), calculates
Formula is:
<mrow>
<msub>
<mi>R</mi>
<mi>d</mi>
</msub>
<mo>=</mo>
<mn>10</mn>
<mo>&times;</mo>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mo>(</mo>
<mfrac>
<mrow>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>H</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
<mo>)</mo>
<mo>=</mo>
<mn>10</mn>
<mo>&times;</mo>
<mi>l</mi>
<mi>o</mi>
<mi>g</mi>
<mo>(</mo>
<mfrac>
<mrow>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>H</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
<mo>+</mo>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>V</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
<mrow>
<mn>2</mn>
<mo>|</mo>
<msub>
<mi>S</mi>
<mrow>
<mi>H</mi>
<mi>V</mi>
</mrow>
</msub>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
<mo>)</mo>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
Wherein SHHRepresent the echo that horizontal emission level is received, SVVRepresent the echo of Vertical Launch vertical reception, SHH
And SVVIt is referred to as co-polarization component, SHVRepresent the echo of horizontal emission vertical reception, SVHRepresent that Vertical Launch level is received
Echo, SHVAnd SVHIt is referred to as cross polar component, | | represent to take the modulus value of the number.
6. it is according to claim 1 based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, its feature
It is, categories combination is carried out to the 7n classes Polarimetric SAR Image of acquisition in step (7), realizes that step is:
(7a) calculates the cluster centre per class data, and calculation formula is:
<mrow>
<msub>
<mi>V</mi>
<mi>I</mi>
</msub>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<msub>
<mi>N</mi>
<mi>i</mi>
</msub>
</mfrac>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>l</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<msub>
<mi>N</mi>
<mi>i</mi>
</msub>
</munderover>
<msub>
<mi>C</mi>
<mi>l</mi>
</msub>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>8</mn>
<mo>)</mo>
</mrow>
Wherein, i=1,2 ... 7n NiRepresent the number of pixel in the i-th class data, ClRepresent l-th of pixel in the i-th class data
The covariance matrix of point;
(7b) calculates the Wishart distances between every two classes data, and calculation formula is:
Dij=(Ni+Nj)ln|V|-Niln|Vi|-Njln|Vj| 9)
Wherein, i, j=1,2 ... 7n, i ≠ j, V represent the cluster centre after two classes merging;
Two minimum class data of (7c) combined distance;
(7d) repeat step (7a)-(7c), until total classification number of Polarimetric SAR Image is m.
7. it is according to claim 1 based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy, its feature
It is, the classification belonging to each pixel in m class Polarimetric SAR Images is updated in step (8), realizes that step is:
(8a) sets primary iteration number of times t=1, and maximum iteration is T, wherein, T > 1;
(8b) utilize formula 8) calculate merge after each class data cluster centre;
(8c) calculates each pixel to the Wishart distances at the i-th class data clusters center, and calculation formula is:
<mrow>
<mi>d</mi>
<mrow>
<mo>(</mo>
<mi>C</mi>
<mo>,</mo>
<msub>
<mi>V</mi>
<mi>i</mi>
</msub>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mi>l</mi>
<mi>n</mi>
<mo>|</mo>
<msub>
<mi>V</mi>
<mi>i</mi>
</msub>
<mo>|</mo>
<mo>+</mo>
<mi>t</mi>
<mi>r</mi>
<mrow>
<mo>(</mo>
<msubsup>
<mi>V</mi>
<mi>i</mi>
<mrow>
<mo>-</mo>
<mn>1</mn>
</mrow>
</msubsup>
<mi>C</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mn>10</mn>
<mo>)</mo>
</mrow>
Wherein, i=1,2 ... m, C are the covariance matrixes of pixel, the mark of tr representing matrixs,Represent ViInverse of a matrix square
Battle array;
(8d) is reclassified according to the Wishart distances of each pixel to each cluster centre to pixel:If pixel
Meet d (C, Vi)≤d(C,Vj), the pixel is then divided into the i-th class by i, j=1,2 ..., m, i ≠ j;
(8e) iterations increases by 1, i.e. t=t+1;
(8f) repeat step (8b)-(8e), to maximum iteration is reached, that is, meets t=T, wherein T > 1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710522622.4A CN107203791B (en) | 2017-06-30 | 2017-06-30 | Polarimetric SAR image classification method based on heteropolarity ratio and scattering power entropy |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710522622.4A CN107203791B (en) | 2017-06-30 | 2017-06-30 | Polarimetric SAR image classification method based on heteropolarity ratio and scattering power entropy |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107203791A true CN107203791A (en) | 2017-09-26 |
CN107203791B CN107203791B (en) | 2019-12-31 |
Family
ID=59910448
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710522622.4A Active CN107203791B (en) | 2017-06-30 | 2017-06-30 | Polarimetric SAR image classification method based on heteropolarity ratio and scattering power entropy |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107203791B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108872987A (en) * | 2018-06-29 | 2018-11-23 | 中国科学院电子学研究所 | The target extraction method with cylindrical structure based on polarization circular track SAR data |
CN110646795A (en) * | 2019-09-16 | 2020-01-03 | 武汉大学 | Ocean oil spill detection method and system of simple polarization SAR |
CN112529817A (en) * | 2020-12-23 | 2021-03-19 | 中国科学院空天信息创新研究院 | Sea spike suppression method for uncalibrated airborne SAR image based on optimal polarization ratio |
CN115063687A (en) * | 2022-08-19 | 2022-09-16 | 航天宏图信息技术股份有限公司 | Polarized SAR image classification method and device |
CN116797845A (en) * | 2023-07-05 | 2023-09-22 | 中国科学院空天信息创新研究院 | Unsupervised reduced polarization classification method based on scattering mechanism |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110222781A1 (en) * | 2010-03-15 | 2011-09-15 | U.S. Government As Represented By The Secretary Of The Army | Method and system for image registration and change detection |
CN103413146A (en) * | 2013-08-23 | 2013-11-27 | 西安电子科技大学 | Method for finely classifying polarized SAR images based on Freeman entropy and self-learning |
CN104504393A (en) * | 2014-12-04 | 2015-04-08 | 西安电子科技大学 | SAR (Synthetic Aperture Radar) image semi-supervised classification method based on integrated learning |
CN105160353A (en) * | 2015-08-18 | 2015-12-16 | 西安电子科技大学 | Polarimetric SAR data ground object classification method based on multiple feature sets |
US20170069062A1 (en) * | 2015-09-08 | 2017-03-09 | The Johns Hopkins University | Small maritime target detector |
-
2017
- 2017-06-30 CN CN201710522622.4A patent/CN107203791B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110222781A1 (en) * | 2010-03-15 | 2011-09-15 | U.S. Government As Represented By The Secretary Of The Army | Method and system for image registration and change detection |
CN103413146A (en) * | 2013-08-23 | 2013-11-27 | 西安电子科技大学 | Method for finely classifying polarized SAR images based on Freeman entropy and self-learning |
CN104504393A (en) * | 2014-12-04 | 2015-04-08 | 西安电子科技大学 | SAR (Synthetic Aperture Radar) image semi-supervised classification method based on integrated learning |
CN105160353A (en) * | 2015-08-18 | 2015-12-16 | 西安电子科技大学 | Polarimetric SAR data ground object classification method based on multiple feature sets |
US20170069062A1 (en) * | 2015-09-08 | 2017-03-09 | The Johns Hopkins University | Small maritime target detector |
Non-Patent Citations (2)
Title |
---|
WANG SHUANG,ET AL.: "《Unsupervised classification of fully polarimetric SAR images based on scattering power entropy and copolarized ratio》", 《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》 * |
邹同元,ET AL.: "《一种新的极化SAR图像非监督分类算法研究》", 《武汉大学学报 信息科技版》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108872987A (en) * | 2018-06-29 | 2018-11-23 | 中国科学院电子学研究所 | The target extraction method with cylindrical structure based on polarization circular track SAR data |
CN110646795A (en) * | 2019-09-16 | 2020-01-03 | 武汉大学 | Ocean oil spill detection method and system of simple polarization SAR |
CN112529817A (en) * | 2020-12-23 | 2021-03-19 | 中国科学院空天信息创新研究院 | Sea spike suppression method for uncalibrated airborne SAR image based on optimal polarization ratio |
CN115063687A (en) * | 2022-08-19 | 2022-09-16 | 航天宏图信息技术股份有限公司 | Polarized SAR image classification method and device |
CN116797845A (en) * | 2023-07-05 | 2023-09-22 | 中国科学院空天信息创新研究院 | Unsupervised reduced polarization classification method based on scattering mechanism |
CN116797845B (en) * | 2023-07-05 | 2024-01-26 | 中国科学院空天信息创新研究院 | Unsupervised reduced polarization classification method based on scattering mechanism |
Also Published As
Publication number | Publication date |
---|---|
CN107203791B (en) | 2019-12-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102208031B (en) | Freeman decomposition and homo-polarization rate-based polarized synthetic aperture radar (SAR) image classification method | |
CN107203791A (en) | Based on the heteropolar Classification of Polarimetric SAR Image method than with scattered power entropy | |
CN108564006B (en) | Polarized SAR terrain classification method based on self-learning convolutional neural network | |
CN102982338B (en) | Classification of Polarimetric SAR Image method based on spectral clustering | |
CN108846426A (en) | Polarization SAR classification method based on the twin network of the two-way LSTM of depth | |
CN102999762B (en) | Decompose and the Classification of Polarimetric SAR Image method of spectral clustering based on Freeman | |
CN108388927A (en) | Small sample polarization SAR terrain classification method based on the twin network of depth convolution | |
CN102968640B (en) | Decompose and the Classification of Polarimetric SAR Image method of data distribution characteristics based on Freeman | |
CN103839073B (en) | Polarization SAR image classification method based on polarization features and affinity propagation clustering | |
CN103413146B (en) | Polarimetric SAR Image sophisticated category method based on Freeman entropys and self study | |
CN105913076A (en) | Polarimetric SAR image classification method based on depth direction wave network | |
CN103186794B (en) | Based on the Classification of Polarimetric SAR Image method of the neighbour's propagation clustering improved | |
CN105138966B (en) | Classification of Polarimetric SAR Image method based on fast density peak value cluster | |
CN104166859A (en) | Polarization SAR image classification based on SSAE and FSALS-SVM | |
CN102999761B (en) | Based on the Classification of Polarimetric SAR Image method that Cloude decomposes and K-wishart distributes | |
CN105160353B (en) | Polarization SAR data terrain classification method based on multiple features collection | |
CN106778884A (en) | The Classification of Polarimetric SAR Image method of plane is decomposed based on scattering entropy and three-component | |
CN109446894A (en) | The multispectral image change detecting method clustered based on probabilistic segmentation and Gaussian Mixture | |
CN105930846A (en) | Neighborhood information and SVGDL (support vector guide dictionary learning)-based polarimetric SAR image classification method | |
CN104361346B (en) | Classification of Polarimetric SAR Image method based on K SVD and rarefaction representation | |
CN105913081A (en) | Improved PCAnet-based SAR image classification method | |
CN107403434A (en) | SAR image semantic segmentation method based on two-phase analyzing method | |
CN106611422A (en) | Stochastic gradient Bayesian SAR image segmentation method based on sketch structure | |
CN103365985B (en) | The adaptive polarization SAR sorting technique of one kind | |
CN109635789A (en) | Based on intensity than the High Resolution SAR image classification method that is extracted with spatial structure characteristic |
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 |