CN117111063A - Polarized SAR image feature extraction method - Google Patents
Polarized SAR image feature extraction method Download PDFInfo
- Publication number
- CN117111063A CN117111063A CN202311059470.0A CN202311059470A CN117111063A CN 117111063 A CN117111063 A CN 117111063A CN 202311059470 A CN202311059470 A CN 202311059470A CN 117111063 A CN117111063 A CN 117111063A
- Authority
- CN
- China
- Prior art keywords
- scattering
- component
- asymmetric
- matrix
- 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.)
- Pending
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 19
- 239000011159 matrix material Substances 0.000 claims abstract description 66
- 230000001427 coherent effect Effects 0.000 claims abstract description 24
- 230000010287 polarization Effects 0.000 claims abstract description 24
- 230000003044 adaptive effect Effects 0.000 claims abstract description 22
- 239000002131 composite material Substances 0.000 claims abstract description 11
- 238000000034 method Methods 0.000 claims description 27
- 230000014509 gene expression Effects 0.000 claims description 10
- 238000012935 Averaging Methods 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 2
- 238000005388 cross polarization Methods 0.000 claims description 2
- 238000000354 decomposition reaction Methods 0.000 description 18
- 238000010276 construction Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000015572 biosynthetic process Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000003786 synthesis reaction Methods 0.000 description 3
- 241000282414 Homo sapiens Species 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000013535 sea water Substances 0.000 description 2
- 241001272720 Medialuna californiensis Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000004035 construction material Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
- G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
- G01S13/9021—SAR image post-processing techniques
- G01S13/9027—Pattern recognition for feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Electromagnetism (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention relates to a polarized SAR image feature extraction method, which comprises the following steps: collecting original polarized SAR image data to obtain a coherent matrix of a polarized SAR image; performing polarization direction angle compensation and ellipticity angle compensation on the coherent matrix; calculating the proportion of the power corresponding to the scattering asymmetric component to the total power; decomposing the coherence matrix after ellipticity angle compensation into a scattering symmetrical part and a scattering asymmetrical part, and obtaining the coherence matrix and scattering power of each part; determining an adaptive volume scattering model; decomposing the obtained scattering symmetry part into a surface scattering component, an even scattering component and an adaptive volume scattering component; the scattering asymmetric portion is decomposed into a helical scattering component, a director scattering component and a composite asymmetric scattering component. The invention can describe the ground objects in different states more precisely, can better extract the scattering information of different ground objects, and plays a great role in analyzing the scattering characteristics of different ground objects.
Description
Technical Field
The invention relates to a polarized SAR image feature extraction method based on a two-stage target decomposition model, and belongs to the technical field of polarized SAR image feature extraction.
Background
The polarized Synthetic Aperture Radar (SAR) is used as an advanced microwave remote sensing means, has the advantages of all weather, all-day time, high resolution and large area coverage, and can more systematically and comprehensively reflect the backscattering characteristic of an object by observing echo information under different transceiving polarization combinations and distinguishing parameters such as the detailed structure, the target orientation, the material composition and the like of the object, thereby obtaining more comprehensive and rich ground object information and having wide application prospect in the remote sensing field.
The most representative characteristic of polarized SAR images is their rich polarized electromagnetic scattering information. The mode of receiving and transmitting electromagnetic waves in various polarization modes in a combined way can reflect information such as structures, sizes, dielectric constants and the like of different ground objects, electromagnetic scattering characteristics different from optical images distinguished by human eyes can be obtained, and the acquisition and recognition capabilities of human beings on target information are expanded. On the other hand, from the aspect of space dimension, the polarized SAR image also has an image expression form, and can reflect the space distribution information among different types of ground features, so that more and richer application means are provided for interpretation of the polarized SAR image. Therefore, the full-polarization scattering characteristics of the targets in the polarized SAR images are researched, the relation between the full-polarization characteristics of the targets and the materialized characteristics of the targets is analyzed, valuable target characteristics are extracted, and further effective target classification or identification is realized, which is the key point and the difficulty of polarized SAR image interpretation.
In a plurality of ground object information extraction theories and methods, the polarized target decomposition method becomes a research hotspot which is concerned in the field of polarized SAR because of the advantages of simple algorithm, easy interpretation of results, clear physical meaning and the like. Therefore, the research on an effective polarization target decomposition method has important theoretical research value and wide application prospect. There are still some problems with the currently existing target decomposition methods. First, the assumption of a scattering model needs to be considered, and when the scattering model does not match with actual data, negative energy may occur in some scattering components, so that physical significance is lost. The second is the scattering symmetry assumption problem, which is due to the scattering symmetry phenomenon in polarized object decomposition, thus the scattering symmetry case needs to be distinguished from the scattering asymmetry case. Finally, polarization orientation angle compensation is needed to be considered for the even scattering of rotation so as to ensure the accuracy of decomposition. Therefore, how to improve the above problems, designing an effective target decomposition method is a critical problem to be solved by polarized SAR image interpretation.
Disclosure of Invention
The invention is used for solving the problem that the incoherent target decomposition method cannot fully consider the scattering symmetry assumption, and provides a polarized SAR image feature extraction method based on a two-stage target decomposition model.
As shown in fig. 1, the polarized SAR image feature extraction method includes the following steps:
firstly, acquiring original polarized SAR image data, and preprocessing to obtain a coherent matrix of a polarized SAR image;
step two, compensating the polarization direction angle of the coherent matrix;
thirdly, carrying out ellipticity angle compensation on the coherent matrix subjected to the polarization pointing angle compensation;
calculating the proportion of the power corresponding to the scattering asymmetric component to the total power;
fifthly, decomposing the coherence matrix subjected to ellipticity angle compensation in the third step into a scattering symmetrical part and a scattering asymmetrical part, and obtaining the coherence matrix and scattering power of each part;
step six, determining an adaptive volume scattering model;
step seven, decomposing the scattering symmetrical part obtained in the step five into a surface scattering component, an even scattering component and an adaptive volume scattering component;
and step eight, decomposing the scattering asymmetric part obtained in the step five into a spiral scattering component, a directional dipole scattering component and a composite asymmetric scattering component.
Further, the method for compensating the polarization direction angle of the coherent matrix in the second step includes the following steps:
the method comprises the following steps of determining a coherent matrix T of a polarized SAR image to be subjected to feature extraction:
each element S in pq (p, q=h, v) represents the target backscattering coefficient at the time of transmission in q-polarized mode, reception in p-polarized mode,<>representing spatial domain averaging;
calculating a rotation angle theta:
wherein T is 23 、T 22 And the like denote elements in the coherence matrix T, and the rotation matrix U (θ) is calculated:
rotating the coherence matrix by T (θ) = [ U (θ) ]]T[U(θ)] *T 。
Further, performing ellipticity angle compensation on the coherence matrix after the polarization pointing angle compensation in the third step includes:
calculating ellipticity angle
Computing a rotation matrix
Wherein j is an imaginary number, and performing ellipticity angle rotation on the coherent matrix after polarization pointing angle compensation
Further, in the fourth step, the ratio Cor of the power corresponding to the scattering asymmetric component to the total power is calculated as:
further, the step five of obtaining the coherence matrix and the scattered power of each part includes the following:
for scattering symmetric targets, element T in the coherence matrix 13 ＝T 23 =0, then the coherence matrix T of the scattering symmetry part sym The elements of (a) satisfy the following formula:
wherein A, B, E, F denotes T sym Unknown elements, adopting a normalized scattering model to model scattering asymmetric components of a target, wherein the expression is as follows:
where m represents the ratio of the backscattering coefficients of horizontal and vertical polarizations, n represents the ratio of the backscattering coefficients of cross-polarization and vertical polarizations, then the coherence matrix T of the asymmetric portion of the scattering asym Expressed as:
to compensate the coherence matrixIs decomposed intoA scattering symmetric part and a scattering asymmetric part, the expressions of which are as follows;
wherein,P sym and P asym Respectively representing scattering power corresponding to scattering symmetric component and scattering asymmetric component, < >>Representing the total power, in the equation:
P asym ＝CorP t
P sym ＝(1-Cor)P t ’
the coherence matrix of the scattering symmetric part and the scattering asymmetric part is:
further, the sixth step includes the steps of:
the adaptive volume scattering model is expressed as:
wherein,
further, in the seventh step, the specific process of decomposing the scattering symmetry part obtained in the fifth step into a surface scattering component, an even scattering component and an adaptive volume scattering component is as follows:
t in the formula s 、T d 、T v Respectively representing a coherent matrix corresponding to the surface scattering model, the even scattering model and the adaptive volume scattering model, wherein alpha and beta are parameters in the model;
make the left and right sides of the formula equal sign correspond to each other, set
S, D, C is a parameter in the solving process, the solving is available,
to ensure uniqueness of solution, useTo determine whether surface scattering or even scattering is the main scattering contribution, when +.>In this case, it is considered that the surface scattering component mainly contributes, and when α=0 is set
When (when)In this case, the even scattering component is considered to be the main contribution, and if β=0 is set
Wherein P is s ，P d And P v The energy of surface scattering, even scattering and adaptive bulk scattering are represented, respectively.
Further, the specific processes of the scattering asymmetric part spiral scattering component, the directional dipole scattering component and the composite asymmetric scattering component obtained in the step five are as follows:
firstly, the scattering asymmetric part spiral scattering component and the directional dipole scattering component obtained in the step five are mixed
Wherein T is h 、T od Respectively representing a coherent matrix corresponding to the spiral scattering model and the directional dipole scattering model, and j represents an imaginary number to obtain
Wherein,and->Respectively represent T asym Middle T 23 And T 13 An element;
P ma is compound asymmetric scattering, and the expression is expressed as P ma ＝P asym -P h -P od 。
The invention has the following beneficial effects:
the method can be used for describing the ground objects in different states more carefully, can better extract the scattering information of the different ground objects, simultaneously increases more representative scattering components, decomposes the scattering asymmetric part in more detail, plays a great role in analyzing the scattering characteristics of the different ground objects, and effectively distinguishes various types of ground objects.
Drawings
FIG. 1 is a flow chart of a polarized SAR image feature extraction method based on a two-stage target decomposition model according to the present invention;
FIG. 2 is a pseudo-color composite image of a polarized SAR image Pauli decomposition to be feature extracted;
FIG. 3 is a first stage resolved scattering symmetry power P sym A result diagram;
FIG. 4 is a first stage resolved scattered asymmetric power P asym A result diagram;
FIG. 5 is a schematic diagram of pseudo-color synthesis of the second stage decomposition result, wherein red represents P d Green represents P v Blue represents P s ；
FIG. 6 is a schematic diagram of pseudo-color synthesis of the second stage decomposition result, wherein red represents P od Green represents P v Blue represents P h 。
Detailed Description
The method for extracting the characteristics of the polarized SAR image provided by the invention is further described in detail below with reference to the accompanying drawings and the detailed description.
Firstly, acquiring original polarized SAR image data, and preprocessing to obtain a polarized SAR image coherence matrix to be subjected to feature extraction;
step two, compensating a polarization direction angle of the coherent matrix;
thirdly, carrying out ellipticity angle compensation on the coherent matrix subjected to the polarization pointing angle compensation;
calculating the proportion of the power corresponding to the scattering asymmetric component to the total power;
fifthly, decomposing the coherence matrix subjected to ellipticity angle compensation in the third step into a scattering symmetrical part and a scattering asymmetrical part, and obtaining the coherence matrix and scattering power of each part;
step six, determining an adaptive volume scattering model;
step seven, decomposing the scattering symmetrical part obtained in the step five into a surface scattering component, an even scattering component and an adaptive volume scattering component;
and step eight, decomposing the scattering asymmetric part obtained in the step five into a spiral scattering component, a directional dipole scattering component and a composite asymmetric scattering component.
Further, the method for compensating the polarization direction angle of the coherent matrix in the second step includes: the coherence matrix < T > of the polarized SAR image to be feature extracted is determined as follows:
each element S in pq (p, q=h, v) represents the target backscattering coefficient at the time of transmission in q-polarized mode, reception in p-polarized mode,<>representing spatial domain averaging;
calculating a rotation angle:
wherein T is 23 、T 22 And the like denote elements in the coherence matrix T, and the rotation matrix U (θ) is calculated:
rotating the coherence matrix:
T(θ)＝[U(θ)]T[U(θ)] *T 。
further, performing ellipticity angle compensation on the coherence matrix after the polarization pointing angle compensation in the third step includes: calculating ellipticity angle
Computing a rotation matrix
Wherein j is an imaginary number, and performing ellipticity angle rotation on the coherent matrix after polarization pointing angle compensation
And step four, calculating the proportion Cor of the power corresponding to the scattering asymmetric component to the total power as follows:
the step five of obtaining the coherence matrix and the scattered power of each part comprises the following contents:
for scattering symmetric targets, element T in the coherence matrix 13 ＝T 23 =0, then the coherence matrix T of the scattering symmetry part sym The elements of (a) satisfy the following formula:
wherein A, B, E, F denotes T sym Unknown elements, a general form of normalized scattering model is adopted to model the scattering asymmetric components of the target, and the expression is that
Where m represents the ratio of the backscattering coefficients of horizontal and vertical polarizations, and n represents the backscattering of cross and vertical polarizationsThe ratio of the coefficients; then scatter the coherence matrix T of the asymmetric part asym Expressed as:
to compensate the coherence matrixIs decomposed into a scattering symmetrical part and a scattering asymmetrical part, and the expression is that
Wherein,P sym and P asym Respectively representing scattering power corresponding to scattering symmetric component and scattering asymmetric component, < >>Representing the total power of the formula
P asym ＝CorP t
P sym ＝(1-Cor)P t
The coherent matrix of the scattering symmetrical part and the scattering asymmetrical part is that
The sixth step comprises the following steps:
the adaptive volume scattering model is expressed as:
wherein,
in the seventh step, the specific process of decomposing the scattering symmetry part obtained in the fifth step into a surface scattering component, an even scattering component and an adaptive volume scattering component is as follows:
t in the formula s 、T d 、T v Respectively representing a coherent matrix corresponding to the surface scattering model, the even scattering model and the adaptive volume scattering model, wherein alpha and beta are parameters in the model;
make the left and right sides of the formula equal sign correspond to each other, set
S, D, C is a parameter in the solving process, the solving is available,
to ensure uniqueness of solution, useTo determine whether surface scattering or even scattering is the main scattering contribution, when +.>In this case, it is considered that the surface scattering component mainly contributes, and when α=0 is set
When (when)In this case, the even scattering component is considered to be the main contribution, and if β=0 is set
Wherein P is s ，P d And P v The energy of surface scattering, even scattering and adaptive bulk scattering are represented, respectively. The unknown number P can be obtained by the above solution s ，P d And P v I.e. the contribution of surface scattering, even scattering and adaptive bulk scattering components.
The specific processes of the scattering asymmetric part spiral scattering component, the directional dipole scattering component and the composite asymmetric scattering component obtained in the step five are as follows:
firstly, the scattering asymmetric part spiral scattering component and the directional dipole scattering component obtained in the step five are mixed
Wherein T is h 、T od Respectively representing a coherent matrix corresponding to the spiral scattering model and the directional dipole scattering model, and j represents an imaginary number to obtain
Wherein,and->Respectively represent T asym Middle T 23 And T 13 An element; meanwhile, in order to ensure the full utilization of energy, a new scattering component is provided, and the asymmetric scattering P is compounded ma To characterize the energy of this part, its expression can be expressed as
P ma ＝P asym -P h -P od
The effects of the present invention can be further illustrated by the following experiments.
1. Experimental data
The data used in the experiment of the invention is a polarized SAR image full-polarization complex coherence format of the U.S. half-moon bay area acquired by an on-board UAVSAR system of the U.S. and the acquisition time is 2014, 11, 24 days. The distance and azimuth resolutions were 1.6m and 0.6m, respectively. The image is 5501 pixels long and 3377 pixels wide, and the pseudo-color composite image of Pauli decomposition is shown in FIG. 2. From the optical map taken on google map, the main ground types of the region include ocean, forest, crops, ships and towns. Because the optical image and the experimental image are acquired at different times, slight differences exist between the types of features.
2. Experimental content and analysis
And for the polarized SAR image, the result of the polarized SAR image characteristic extraction method based on the two-stage target decomposition model is shown in figures 3-6. The first stage target decomposition results obtained using UAVSAR experimental data are shown in FIGS. 3-4, wherein FIG. 3 is the scattering symmetric power P sym FIG. 4 shows the scattered asymmetric power P asym . As can be seen from the figure, firstly, since the scattering symmetry phenomenon is found in natural regions, the scattering symmetry component in natural regions such as forests and farmlands is stronger than the scattering asymmetry component; second, the building area is alsoAlso, higher surface scattering and bulk scattering components are present, so that the scattering symmetric power thereof is also higher; finally, because the artificial target areas such as buildings and the like have more complex space structures, the scattering models of the artificial target areas do not meet more symmetrical conditions, and therefore, the scattering asymmetric power of the artificial target areas is higher than that of natural areas such as forests, farmlands and the like.
The second stage further decomposes the scattering symmetric part and the scattering asymmetric part, wherein the scattering symmetric part P sym Is further decomposed into surface scattering P s Even scattering P d And adaptive volume scattering P v Scattering asymmetric part P asym Is continuously decomposed into spiral scattering P h Directional dipole scattering P od And composite asymmetric scattering P ca . The false color synthesis results of the target decomposition are shown in fig. 5-6.
It can be found that, for a forest region, the forest canopy has strong backscattering, a high gray value and a speckled texture, and the main scattering mechanism is a bulk scattering component, so that the forest canopy is mostly green in target decomposition. The farmland area is generally represented as a dark area with smaller gray values on the image, because the farmland area is flat, the scattering echo received by the radar is relatively less, and the scattering mechanism is mainly based on surface scattering components. The construction area is complicated in structure due to the artificial construction process, and thus the performance of the construction area is affected by various factors including the properties of the construction itself such as the construction material and the size structure of the construction, the incident angle of electromagnetic waves, the resolution, etc. In general, the structure characteristics of a large number of dihedral shapes exist, the distribution fluctuation of the spatial textures is large, and a certain rule does not exist. The polarized SAR image is generally reflected as even scattering, and more volume scattering components exist due to the complex spatial arrangement of the building area. And thus is typically shown as white or purple in the pseudo-color composite map. For the ocean area, the backscattering of the smooth and calm sea water surface is weak, and the echo of the water body is small, so that the main scattering component of the sea water is represented as surface scattering, and the gray value is small. However, when the image is acquired, the sea surface is not calm, and as can be seen from the texture of the ocean area in the figure, the image has a large number of waves, and the waves have obvious structures and textures, so that the scattering characteristics of the waves are changed, and more scattering asymmetric components exist.
The results of the invention show that besides the marine area being able to find a relatively high scattering asymmetric component, the complex texture structure of the ocean wave also causes more scattering asymmetric components to exist, which also illustrates that the proposed method can be more detailed for different states of ground objects. Experimental results show that the invention can better extract the scattering information of different ground objects, meanwhile, more representative scattering components are added, the scattering asymmetric parts are decomposed in more detail, the invention plays a great role in analyzing the scattering characteristics of different ground objects, and various types of ground objects can be effectively distinguished.
While the present invention has been described in detail through the foregoing description of the preferred embodiment, it should be understood that the foregoing description is not to be considered as limiting the invention. Many modifications and substitutions of the present invention will become apparent to those of ordinary skill in the art upon reading the foregoing. Accordingly, the scope of the invention should be limited only by the attached claims.
Claims (8)
1. The polarized SAR image feature extraction method is characterized by comprising the following steps of:
firstly, acquiring original polarized SAR image data, and preprocessing to obtain a coherent matrix of a polarized SAR image;
step two, compensating the polarization direction angle of the coherent matrix;
thirdly, carrying out ellipticity angle compensation on the coherent matrix subjected to the polarization pointing angle compensation;
calculating the proportion of the power corresponding to the scattering asymmetric component to the total power;
fifthly, decomposing the coherence matrix subjected to ellipticity angle compensation in the third step into a scattering symmetrical part and a scattering asymmetrical part, and obtaining the coherence matrix and scattering power of each part;
step six, determining an adaptive volume scattering model;
step seven, decomposing the scattering symmetrical part obtained in the step five into a surface scattering component, an even scattering component and an adaptive volume scattering component;
and step eight, decomposing the scattering asymmetric part obtained in the step five into a spiral scattering component, a directional dipole scattering component and a composite asymmetric scattering component.
2. The method for extracting features of polarized SAR image according to claim 1, wherein said method for compensating the polarization orientation angle of the coherence matrix in step two comprises the following steps:
the method comprises the following steps of determining a coherent matrix T of a polarized SAR image to be subjected to feature extraction:
each element S in pq (p, q=h, v) represents the target backscattering coefficient at the time of transmission in q-polarized mode, reception in p-polarized mode,<>representing spatial domain averaging;
calculating a rotation angle theta:
wherein T is 23 、T 22 And the like denote elements in the coherence matrix T, and the rotation matrix U (θ) is calculated:
rotating the coherence matrix by T (θ) = [ U (θ) ]]T[U(θ)] *T 。
3. The polarized SAR image feature extraction method of claim 2, wherein performing ellipticity angle compensation on the coherence matrix after the polarized pointing angle compensation in step three comprises:
calculating ellipticity angle
Computing a rotation matrix
Wherein j is an imaginary number, and performing ellipticity angle rotation on the coherent matrix after polarization pointing angle compensation
4. The polarized SAR image feature extraction method as set forth in claim 3, wherein,
and step four, calculating the proportion Cor of the power corresponding to the scattering asymmetric component to the total power as follows:
5. the method for extracting features of polarized SAR image according to claim 4, wherein said obtaining the coherence matrix and scattering power of each part in step five comprises:
for scattering symmetric targets, element T in the coherence matrix 13 ＝T 23 =0, then the coherence moment of the scattering symmetry partArray T sym The elements of (a) satisfy the following formula:
wherein A, B, E, F denotes T sym Unknown elements, adopting a normalized scattering model to model scattering asymmetric components of a target, wherein the expression is as follows:
where m represents the ratio of the backscattering coefficients of horizontal and vertical polarizations, n represents the ratio of the backscattering coefficients of cross-polarization and vertical polarizations, then the coherence matrix T of the asymmetric portion of the scattering asym Expressed as:
to compensate the coherence matrixThe method is divided into a scattering symmetrical part and a scattering asymmetrical part, and the expression is as follows;
the scattering power corresponding to the scattering asymmetric component,representing the total power, in the equation:
the coherence matrix of the scattering symmetric part and the scattering asymmetric part is:
6. the polarized SAR image feature extraction method of claim 5, wherein said step six comprises the steps of:
the adaptive volume scattering model is expressed as:
wherein,
7. the polarized SAR image feature extraction method according to claim 6, wherein the specific process of decomposing the scattering symmetry part obtained in step five into the surface scattering component, the even scattering component and the adaptive volume scattering component in step seven is:
t in the formula s 、T d 、T v Respectively representing a coherent matrix corresponding to the surface scattering model, the even scattering model and the adaptive volume scattering model, wherein alpha and beta are parameters in the model;
make the left and right sides of the formula equal sign correspond to each other, set
S, D, C is a parameter in the solving process, the solving is available,
to ensure uniqueness of solution, useTo determine whether surface scattering or even scattering is the main scattering contribution, when +.>In this case, it is considered that the surface scattering component mainly contributes, and when α=0 is set
When (when)In this case, the even scattering component is considered to be the main contribution, and if β=0 is set
Wherein P is s ，P d And P v The energy of surface scattering, even scattering and adaptive bulk scattering are represented, respectively.
8. The polarized SAR image feature extraction method according to claim 7, wherein the specific procedures of scattering asymmetric partial spiral scattering component, directional dipole scattering component and composite asymmetric scattering component obtained in step five are:
firstly, the scattering asymmetric part spiral scattering component and the directional dipole scattering component obtained in the step five are mixed
Wherein T is h 、T od Respectively representing a coherent matrix corresponding to the spiral scattering model and the directional dipole scattering model, and j represents an imaginary number to obtain
Wherein,and->Respectively represent T asym Middle T 23 And T 13 An element;
P ma is compound asymmetric scattering, and the expression is expressed as P ma ＝P asym -P h -P od 。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311059470.0A CN117111063A (en) | 2023-08-22 | 2023-08-22 | Polarized SAR image feature extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311059470.0A CN117111063A (en) | 2023-08-22 | 2023-08-22 | Polarized SAR image feature extraction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117111063A true CN117111063A (en) | 2023-11-24 |
Family
ID=88806918
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311059470.0A Pending CN117111063A (en) | 2023-08-22 | 2023-08-22 | Polarized SAR image feature extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117111063A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117390355A (en) * | 2023-12-12 | 2024-01-12 | 江西师范大学 | Polarization target decomposition method based on LPC compensation |
-
2023
- 2023-08-22 CN CN202311059470.0A patent/CN117111063A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117390355A (en) * | 2023-12-12 | 2024-01-12 | 江西师范大学 | Polarization target decomposition method based on LPC compensation |
| CN117390355B (en) * | 2023-12-12 | 2024-03-15 | 江西师范大学 | Polarization target decomposition method based on LPC compensation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xiang et al. | Unsupervised polarimetric SAR urban area classification based on model-based decomposition with cross scattering | |
| Chen et al. | General polarimetric model-based decomposition for coherency matrix | |
| Ma et al. | What catch your attention in SAR images: Saliency detection based on soft-superpixel lacunarity cue | |
| Chen et al. | Spectral classification of ecological spatial polarization SAR image based on target decomposition algorithm and machine learning | |
| CN101498789B (en) | Ground object target classification method and apparatus based on polarimetric synthetic aperture radar | |
| CN107527023B (en) | Polarized SAR image classification method based on superpixels and topic models | |
| CN117111063A (en) | Polarized SAR image feature extraction method | |
| CN107203791B (en) | Polarimetric SAR image classification method based on heteropolarity ratio and scattering power entropy | |
| CN112884029B (en) | Collaborative classification method integrating fully-polarized SAR and hyperspectral remote sensing | |
| Gao et al. | Scattering characteristic-aware fully polarized SAR ship detection network based on a four-component decomposition model | |
| CN112630741B (en) | Full-polarization synthetic aperture radar image target compensation PEOC method | |
| Chen et al. | Polarimetric coherence pattern: A visualization tool for PolSAR data investigation | |
| Fan et al. | Marine floating raft aquaculture dynamic monitoring based on multi-source GF imagery | |
| Hu et al. | Model-based polarimetric target decomposition with power redistribution for urban areas | |
| Souissi et al. | Analysis of orientation angle shifts on the polarimetric data using Radarsat2 images | |
| CN107507251B (en) | Pseudo-color synthesis method and device of dual-polarization SAR image | |
| CN107144841B (en) | Polarized SAR image target decomposition method based on minimum residual power | |
| Mishra et al. | Critical analysis of model-based incoherent polarimetric decomposition methods and investigation of deorientation effect | |
| Arai et al. | Polarimetric SAR image classification with the maximum curvature of the trajectory in the eigen space converted from the polarization signature | |
| Karachristos et al. | A Review on PolSAR Decompositions for Feature Extraction | |
| Xue et al. | An efficient classification method of fully polarimetric SAR image based on polarimetric features and spatial features | |
| CN107578064B (en) | Sea surface oil spill detection method based on superpixel and utilizing polarization similarity parameters | |
| Chen et al. | Equivalent complex valued deep semantic segmentation network for sar images | |
| Gou et al. | Spectral clustering with eigenvalue similarity metric method for POL-SAR image segmentation of land cover | |
| CN111125622A (en) | Improved hybrid Freeman/Eigenvalue decomposition method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |