CN110703220B - Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors - Google Patents

Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors Download PDF

Info

Publication number
CN110703220B
CN110703220B CN201910968899.9A CN201910968899A CN110703220B CN 110703220 B CN110703220 B CN 110703220B CN 201910968899 A CN201910968899 A CN 201910968899A CN 110703220 B CN110703220 B CN 110703220B
Authority
CN
China
Prior art keywords
vegetation
baseline
rvog
time
polarization
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.)
Active
Application number
CN201910968899.9A
Other languages
Chinese (zh)
Other versions
CN110703220A (en
Inventor
付海强
张兵
朱建军
李志伟
汪长城
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central South University
Original Assignee
Central South University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Central South University filed Critical Central South University
Priority to CN201910968899.9A priority Critical patent/CN110703220B/en
Publication of CN110703220A publication Critical patent/CN110703220A/en
Application granted granted Critical
Publication of CN110703220B publication Critical patent/CN110703220B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
    • 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
    • G01S13/9023—SAR image post-processing techniques combined with interferometric techniques
    • 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
    • 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
    • G01S13/904—SAR modes
    • G01S13/9076—Polarimetric features in SAR
    • 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
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/024—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using polarisation effects

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Processing (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention discloses a multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors, which comprises the following steps: step 0, introducing a time decoherence factor into the RVoG model to obtain the RVoG extended model considering the time decoherence factor; step 1, preprocessing and polarization interference processing are carried out on a multi-baseline polarization interference SAR image of vegetation to obtain a multi-baseline multi-polarization complex phase coherence coefficient observation value; step 2, setting and calculating an inversion initial value of the vegetation parameter of the RVoG extended model; step 3, selecting one of the N baselines as a reference baseline, and considering that the influence of a time decoherence factor does not exist; step 4, matching the observed value with the RVoG extended model to construct an observation equation; and finally, performing vegetation parameter inversion by adopting a nonlinear iterative algorithm according to the obtained initial parameter value. The invention can improve the inversion precision of the vegetation parameters and is easy to realize.

Description

Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors
Technical Field
The invention relates to the field of vegetation parameter inversion of a polarimetric interferometric synthetic aperture radar (PolInSAR), in particular to a multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors.
Background
The polarimetric interference synthetic aperture radar (PolInSAR) technology is a novel microwave remote sensing technology, the radar polarimetric concept is expanded to an interference space, the PolInSAR technology can be used for separating the positions and the vertical distribution conditions of different scattering mechanisms in the same resolution unit, the positions of vegetation canopies and scattering centers of earth surfaces can be resolved in vegetation areas, and then the vegetation height is extracted from the vegetation areas. At present, the PolInSAR technology is one of the most potential means for regional high-precision and even global scale vegetation height inversion.
In the field of vegetation height inversion by adopting the PolInSAR technology, the model which is most widely applied at present is a Random Volume over group (RVoG) model, and a functional relation between an InSAR observation value and a vegetation parameter is established. However, in the modeling process of the RVoG model, other non-body decorrelation factors such as time decorrelation and the like are ignored; in addition, in the resolving process of the RVoG model, the problems of insufficient observation information and unknown parameter resolving rank caused by excessive parameterization exist.
In fact, in the heavy-rail interference mode, a certain time interval exists between two radar images; during this time interval, the shape, position, dielectric properties, etc. of the scatterers may change due to the influence of wind, rainfall, temperature, human factors, etc. Therefore, in the heavy-orbit interference mode, the PolInSAR technology is adopted to carry out vegetation height inversion, the influence of a time decorrelation factor cannot be ignored, and the vegetation height inversion accuracy is relatively low otherwise.
Therefore, a multi-baseline PolInSAR vegetation parameter inversion method is needed to be designed, and the influence of a time decoherence factor in a heavy-orbit interference mode on vegetation parameter inversion can be compensated; meanwhile, the problem of rank deficiency of unknown parameter calculation caused by insufficient observation information in the calculation process based on the model can be solved.
Disclosure of Invention
The invention aims to solve the technical problem of providing a multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors and improving vegetation parameters obtained by vegetation parameter inversion.
In order to achieve the technical purpose, the invention adopts the following technical scheme:
a multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors comprises the following steps:
step 0, introducing a time decorrelation factor into the RVoG model to obtain an RVoG extended model considering the time decorrelation factor:
Figure GDA0003024000320000021
vegetation coverage scenarios on microwave-opaque terrain are represented using the RVoG extended model, where ω represents the polarization state relative to baseline and polarization mode,
Figure GDA0003024000320000022
representing the phase of the earth surface, gammaTVRepresenting the temporal decoherence factor, gamma, of the vegetation bodyTGRepresenting the surface time decorrelation factor, mu(ω)Representing the earth body amplitude ratio corresponding to the polarization state omega; gamma rayvRepresenting a pure decoherence coefficient as a function of vegetation height and extinction coefficient;
step 0.5, parameter normalization: earth amplitude ratio mu for interfering N base lines(ω)Normalizing in a complex plane to obtain the same normalized parameter x of the earth volume amplitude ratiorAnd normalizing the surface elevation parameters; wherein, the body amplitude ratio normalization parameter xrThe distance from the polarized complex coherence coefficient of the volume scattering dominant channel of the reference baseline r to the pure body decorrelation coefficient is referred to;
step 1, carrying out registration, flat ground removal, multi-vision and polarization interference processing on a multi-baseline polarization interference SAR image of vegetation to generate an interference pattern, and acquiring a multi-baseline multi-polarization complex phase correlation coefficient observed value gamma of the vegetation(ω)(ii) a Setting and generating N base lines, wherein each base line adopts M polarization modes;
step 2, setting the initial value of the body amplitude ratio normalization parameter as 0, and setting the initial values of the time decorrelation factors of the vegetation body and the earth surface as 1; then calculating an initial value of vegetation height, an initial value of extinction coefficient and an initial value of surface elevation by using the RVoG model and a three-stage vegetation height inversion algorithm;
step 3, selecting one of the N baselines as a reference baseline r, wherein the reference baseline has no influence of time decoherence, and the reference baselineSurface time decoherence factor gamma TG1 and vegetation body time decoherence factor gamma TV1, and is fixed; in the RVoG extended model, the earth surface time decorrelation factors and vegetation body time decorrelation factors of other baselines except the reference baseline refer to corresponding time decorrelation factors when interfering relative to the reference baseline;
step 4, multi-baseline PolInSAR vegetation parameter inversion considering time decoherence factors;
for each base line, matching the M observation values obtained in the step 1 with the RVoG expansion model constructed in the step 0 to obtain an observation equation, and splitting a real part and an imaginary part to obtain 2MN split observation equations;
solving 2(N-1) +4 unknown parameters in 2MN observation equations by adopting a nonlinear iterative algorithm according to the initial values obtained in the step 2 and the time decoherence factors of the earth surface and the vegetation body of the reference base line obtained in the step 3: vegetation height, extinction coefficient, surface elevation, surface-to-body amplitude ratio normalization parameter, surface time decorrelation factor of other N-1 baselines except reference base line and vegetation body time decorrelation factor.
Further, the reference baseline r is selected by the following method: acquiring complex phase coherence coefficient observation values of each baseline in different polarization states, wherein an elliptical area is formed in a complex plane unit circle; and calculating the eccentricity of each elliptical area, and taking the baseline corresponding to the maximum eccentricity as a reference baseline, wherein the reference baseline is the one of all the N baselines which is least influenced by the time decorrelation factor.
Further, the method for calculating the vegetation height initial value, the extinction coefficient initial value and the surface elevation initial value by utilizing the RVoG model and the three-stage vegetation height inversion algorithm comprises the following steps:
step 1), straight line fitting:
randomly selecting a base line, and performing straight line fitting on the RVoG model in a complex plane by using complex coherence coefficient observation values of the base line in different polarization states to obtain a straight line shown in the following formula:
Figure GDA0003024000320000031
step 2), calculating the earth surface phase:
calculating 2 intersection points A and B of the straight line and the complex plane unit circle shown in the formula (2), and taking the intersection point, in which the distance between the intersection point and the point of the volume scattering dominant polarization mode is greater than the distance between the intersection point and the point of the ground scattering dominant polarization mode, as the earth surface phase point
Figure GDA0003024000320000032
Then using the earth's surface phase
Figure GDA0003024000320000033
Calculating the initial value of the surface elevation according to the following formula (3):
Figure GDA0003024000320000034
step 3), vegetation height and extinction coefficient estimation:
if the bulk scattering dominant polarization channel only contains the vegetation layer scattering contribution and the ground-to-volume amplitude ratio is 0, the expression of the pure-body coherence is
Figure GDA0003024000320000035
Figure GDA0003024000320000037
The complex coherence coefficient of the volume scattering dominant channel is related to the height h of the vegetationvAnd a function of the extinction coefficient sigma expressed as
Figure GDA0003024000320000036
By numerical calculation, by giving a series of vegetation heights hvAnd an extinction coefficient sigma, establishing a two-dimensional lookup table according to a formula (4), and finding a group of values with the minimum difference, namely the vegetation height hvAnd an estimated value of the extinction coefficient sigma, and the estimated value is used as an initial vegetation height value and an initial extinction coefficient value:
Figure GDA0003024000320000041
further, the baseline selected for the straight line fitting is the reference baseline.
Advantageous effects
The beneficial effects of the invention include:
1. the influence of the time decoherence factor which is inevitable for vegetation parameter inversion in the heavy-rail interference mode is compensated, so that the precision of the vegetation parameter obtained by the vegetation parameter inversion is improved;
2. by adopting observation value data of multiple baselines, the problem of rank deficiency when the RVoG extended model which considers time to remove coherent factors is applied to inversion of vegetation height in the prior art is solved, the resolving stability in parameter inversion solving is improved, and the resolving precision is improved;
3. according to the eccentricity of the ellipse represented by the observation value, a base line which is minimally affected by the time decorrelation factor is selected as a reference base line, and the difficulty of parameter inversion by using a RVoG expansion model through the reference base line can be reduced to the greatest extent under the condition that the inversion accuracy of vegetation parameters is guaranteed;
4. the vegetation parameter inversion method is simple and clear, is easy to implement, and can be widely applied to aspects such as large-range vegetation height under a long-time baseline, global carbon reserve estimation and the like in the future.
Drawings
FIG. 1 is a schematic flow chart of a vegetation height inversion method of the present invention;
FIG. 2 is a geometric representation of the magnitude ratio of the terrain under the RVoG model framework;
FIG. 3 shows the inversion result of vegetation height according to an embodiment of the present invention;
FIG. 4 is a cross-validation graph of vegetation height inversion results and LiDAR measurement results in accordance with an embodiment of the present invention.
Detailed Description
The following describes embodiments of the present invention in detail, which are developed based on the technical solutions of the present invention, and give detailed implementation manners and specific operation procedures to further explain the technical solutions of the present invention.
A multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors is shown in figure 1 and comprises the following steps:
step 0, constructing an RVoG extension model;
introducing a time decorrelation factor into the RVoG model to obtain the RVoG extended model considering the time decorrelation factor:
Figure GDA0003024000320000051
representing the vegetation scene using a RVoG extended model, wherein:
ω represents the polarization state relative to the baseline and polarization mode;
Figure GDA0003024000320000054
representing earth surface phase, geometrically related to the interference baseline, and different earth surface phases corresponding to different baselines;
γTVrepresenting the temporal decoherence factor, gamma, of the vegetation bodyTGRepresenting the time-of-the-earth decorrelation factor,
μ(ω)the earth body amplitude ratio corresponding to the polarization state omega is expressed and is related to the interference baseline and the polarization mode, and the earth body amplitude ratios of different interference baselines and the same polarization mode are generally different;
γvrepresenting the decoherence coefficient of a pure body, is the height h of the vegetationvAnd an extinction coefficient σ, the specific expression being as follows:
Figure GDA0003024000320000052
wherein:
hv: vegetation height;
σ: average extinction coefficient of vegetation body; in the RVoG model, a vegetation body is taken as a uniform medium, and the microwave obeys exponential distribution attenuation in the process of penetrating through vegetation, so that the extinction coefficient of the vegetation body is a fixed real number;
θ: a microwave incident angle;
kz: the effective wavenumber in the vertical direction is related to the geometrical configuration of the interference baseline, and is specifically expressed as follows:
Figure GDA0003024000320000053
wherein:
B⊥: a vertical baseline length;
r: the microwave transmission distance is the distance between a radar satellite and an observation target;
λ: a microwave wavelength.
Step 0.5, normalizing parameters;
earth amplitude ratio mu for interfering N base lines(ω)Normalizing in a complex plane to obtain the same normalized parameter x of the earth volume amplitude ratior(ii) a Wherein, the body amplitude ratio normalization parameter xrIt refers to the distance from the polarized complex coherence coefficient of the volume scattering dominant channel with reference to the baseline r to the pure volume decorrelation coefficient.
Because of the complex plane unit circle, the complex coherence of different polarization modes has different earth body amplitude ratio
Figure GDA0003024000320000061
The function of the method is to determine the position of the complex coherence on a coherent straight line, and obviously, a certain linear correlation exists. Therefore, based on the RVoG model, the terrain amplitude ratio is represented by the geometric relational expression shown in formula (8) and the terrain amplitude ratio geometric representation under the RVoG model framework shown in fig. 2:
Figure GDA0003024000320000062
wherein:
Figure GDA0003024000320000063
arbitrary polarization complex coherence coefficient;
Figure GDA0003024000320000064
the volume scattering dominates the polarization complex coherence coefficient of the channel;
γv: pure body decoherence coefficient;
x: polarization complex coherence coefficient of dominant channel of volume scattering
Figure GDA0003024000320000065
To pure body decoherence coefficient gammavThe distance of (c).
Wherein, in the formula (8)
Figure GDA0003024000320000066
D can be solved through the complex phase dry coefficient estimated by PolInSAR, so that the earth-to-body amplitude ratio of all polarization modes can be normalized to an unknown parameter x through a formula (8)rIs described in (1).
When the time decorrelation factor is considered, the vegetation scattering characteristic cannot be changed by the time decorrelation factor, namely, the earth volume amplitude ratio of the same polarization mode of different polarization channels is assumed to be unchanged, so that the formula (8) can be expanded to be the earth volume amplitude ratio expression under the condition of multiple base lines:
Figure GDA0003024000320000067
wherein:
j: superscript, representing baseline designation;
xr: the distance from the polarized complex coherence coefficient of the volume scattering dominant channel corresponding to the reference baseline to the pure body decorrelation coefficient;
according to equation (9), the RVoG extended model when considering the time decoherence factor, all polarization mode body amplitude ratios of different baselines are normalized to unknown parameter xrIs described in (1).
In this step 0.5, except that the earth-magnitude ratio is normalized to unknownParameter xrFurther comprising normalizing the surface elevation parameters:
according to InSAR principle, earth surface phase
Figure GDA0003024000320000071
Elevation h from the earth's surfacegAnd the effective wavenumber in the vertical direction. In a multi-baseline configuration, the surface phase can be expressed as the following equation (10):
Figure GDA0003024000320000072
the difference of interference geometries of different baselines causes the difference of earth surface phase, but the earth surface elevation does not change along with the change of the baselines, so that the earth surface phase of different baselines can pass through an unknown parameter hgAnd (4) expressing.
Step 1, polarization interference treatment;
the step mainly obtains a multi-baseline multi-polarization complex phase dry coefficient observation value gamma(ω)The method comprises the steps of registering, land leveling, multi-vision and polarization interference processing on a multi-baseline polarization interference SAR image of vegetation so as to generate an interferogram, namely, a multi-baseline multi-polarization complex phase coherence coefficient observation value gamma of the vegetation can be obtained from the interferogram(ω)(ii) a N base lines are generated, and each base line adopts M polarization modes.
During registration, the main image is one of the SAR images, and all the auxiliary images need to be registered with the same main image.
Step 2, estimating initial values of parameters;
because the body amplitude ratio normalization parameter is 0 in an ideal state, the initial value of the body amplitude ratio normalization parameter is set to be 0 when the initial value is set; the time decorrelation factor is 1 under the condition of not being influenced by the time decorrelation factor, so the initial values of the time decorrelation factors of the vegetation body and the earth surface are both set to be 1;
then, calculating an initial vegetation height value, an initial extinction coefficient value and an initial surface elevation value by using the RVoG model and a three-stage vegetation height inversion algorithm, wherein the specific method comprises the following steps:
step 1), straight line fitting:
randomly selecting a base line, and performing straight line fitting on the RVoG model in a complex plane by using complex coherence coefficient observation values of the base line in different polarization states to obtain a straight line shown in the following formula:
Figure GDA0003024000320000081
if the base line selected by straight line fitting is the reference base line, the initial values of the vegetation height, the extinction coefficient and the surface elevation can be closer to the initial value of the RVoG extended model, and therefore the inversion accuracy of the vegetation parameters is improved.
Step 2), calculating the earth surface phase:
2 intersection points exist between the straight line obtained by fitting in the step 1) and the unit circle of the complex plane, wherein 1 is a ground surface phase point. In the complex plane unit circle, the distance from the earth surface phase point to the in-vitro scattering dominant polarization mode point is larger than the distance from the earth surface scattering dominant polarization mode point, so the earth surface phase point can be obtained by judgment through a judgment formula (12), and the judgment formula is as follows:
Figure GDA0003024000320000082
wherein:
Figure GDA0003024000320000083
respectively represents the complex coherence coefficient of the body scattering dominant channel and the complex coherence coefficient of the ground scattering dominant channel.
Therefore, the method for calculating the earth's surface phase is as follows: calculating 2 intersection points A and B of the straight line and the complex plane unit circle shown in the formula (12), and taking the intersection point of which the distance between the intersection point and the point in the volume scattering dominant polarization mode is greater than the distance between the intersection point and the point in the ground scattering dominant polarization mode as the earth surface phase point
Figure GDA0003024000320000084
Then the earth's surface phase is reused
Figure GDA0003024000320000085
Calculating the initial value of the surface elevation according to the following formula (13):
Figure GDA0003024000320000086
step 3), vegetation height and extinction coefficient estimation:
if the bulk scattering dominant polarization channel only contains the vegetation layer scattering contribution and the ground-to-volume amplitude ratio is 0, the expression of the pure-body coherence is
Figure GDA0003024000320000087
Figure GDA0003024000320000088
The complex coherence coefficient of the volume scattering dominant channel is related to the height h of the vegetationvAnd a function of the extinction coefficient sigma expressed as
Figure GDA0003024000320000089
By numerical calculation, by giving a series of vegetation heights hvAnd an extinction coefficient sigma, establishing a two-dimensional lookup table according to a formula (14), and finding a group of values with the minimum difference, namely the vegetation height hvAnd an estimated value of the extinction coefficient sigma, and the estimated value is used as an initial vegetation height value and an initial extinction coefficient value:
Figure GDA00030240003200000810
step 3, selecting one of the N baselines as a reference baseline r, wherein the reference baseline has no influence of time decoherence, and the earth surface time decoherence factor gamma of the reference baseline TG1 and vegetation body time decoherence factor gamma TV1, and is fixed; in the present invention, other than the reference base lineThe ground surface time decorrelation factor and the vegetation body time decorrelation factor of the baseline refer to corresponding time decorrelation factors when interfering relative to a reference baseline.
Based on the RVoG model framework, the complex coherence coefficient, namely the PolInSAR observed value obtained in step 1, is theoretically distributed on a straight line, but the straight line is degraded into an elliptical area under the influence of various noise and other factors. For any baseline, the more oblate and prolate the elliptical area formed within the complex planar unit circle, the less it will be affected by factors such as temporal decoherence. Therefore, in the RVoG extended model considering the temporal decorrelation factor, the reference baseline is selected by using the elliptical eccentricity as a judgment index:
acquiring complex phase coherence coefficient observation values of each baseline in different polarization states, and representing the complex phase coherence coefficient observation values in a complex plane unit circle to form an elliptical area; and calculating the eccentricity of each elliptical area, taking the base line corresponding to the maximum eccentricity as a reference base line, and obtaining the reference base line which is the one of all the N base lines and is least influenced by the time decorrelation factor. The judgment formula is as follows:
Figure GDA0003024000320000091
wherein: ecc denotes elliptical regional eccentricity; 1,2, L, j in the superscript represents the baseline designation; max | | | represents taking the maximum value.
Step 4, multi-baseline PolInSAR vegetation parameter inversion considering time decoherence factors;
for each base line, matching the M observation values obtained in the step 1 with the RVoG expansion model constructed in the step 0 to obtain an observation equation, and splitting a real part and an imaginary part to obtain 2MN split observation equations;
solving 2(N-1) +4 unknown parameters in 2MN observation equations by adopting a nonlinear iterative algorithm according to the initial values obtained in the step 2 and the time decoherence factors of the earth surface and the vegetation body of the reference base line obtained in the step 3: vegetation height, extinction coefficient, surface elevation, surface-to-body amplitude ratio normalization parameter, surface time decorrelation factor of other N-1 baselines except reference base line and vegetation body time decorrelation factor.
The general formula of an observation equation obtained by matching the observation value with the RVoG extended model is as follows:
Figure GDA0003024000320000101
if the invention does not normalize the body-to-body amplitude ratio to an unknown parameter xrThe observation equation of any baseline includes 7 unknown parameters:
Figure GDA0003024000320000102
σ,hv,γTV,γTG,μ1,μ2(ii) a According to the common knowledge, a straight line equation can be uniquely determined by 2 points on a straight line, namely, 2 complex phase dry coefficients corresponding to any base line can uniquely determine the position of the fitted straight line on a complex plane unit circle in the step 2, namely, only 2 independent complex phase dry coefficient observation values exist in any base line, only 4 independent observation values are obtained after a real part and an imaginary part are disassembled, and the equation can not be solved due to rank.
In the normalization of the earth volume amplitude ratio to an unknown parameter xrThe unknown parameters can then be reduced to 6: h isg,σ,hv,γTV,γTG,xrStill greater than the number of independent observation equations 4, the equations still have a rank deficiency problem.
Under the assumption that the vegetation scattering characteristics are stable, the 2 nd base line is added, the other 4 independent complex observed quantities can be provided, 8 real independent observation equations (the real part and the imaginary part of a complex number are split), and at the moment, the unknown parameters become 8 hv,σ,hg,xr,
Figure GDA0003024000320000103
Although the number of observed equations is equal to the number of unknown parameters, it is lacking
Figure GDA0003024000320000104
Is determined based on the prior information of (c),the equation is still rank deficient and cannot be solved.
Similarly, the theory can be extended to the configuration situation of N base lines, and the equations all have the problem of rank deficiency.
In order to solve the problem that when the influence of time decorrelation factors exists, the configuration of multiple baselines is still rank deficient and vegetation parameters cannot be stably solved, the method selects a reference baseline r which is not influenced by the time decorrelation factors, and thus the unknown number existing in the reference baseline is hv,σ,hg,xr. If a baseline is added, x is determinedrCan be solved without prior information, then hv,σ,hgCan be uniquely determined. Substituting it into the expression of the newly added baseline, at this time
Figure GDA0003024000320000105
The solution can be performed without any prior information. In summary, when there are 2 baselines, assuming that the reference baseline is not affected by temporal decoherence and the vegetation scattering characteristics are stable, the equation is no longer rank deficient at this time, and the unknown parameter h can be solvedv,σ,hg,xr,
Figure GDA0003024000320000106
Wherein
Figure GDA0003024000320000107
Is a temporal decorrelation factor relative to a reference baseline r in which there is no temporal influence.
When the method is expanded to multi-baseline configuration, the number of observation equations is 2MN, the number of unknown parameters is 2N +2, and when N is not less than 2, the equations no longer have the rank deficiency problem and can be solved. All temporal decorrelation factors obtained at this time are temporal decorrelation factors for the baseline relative to the absence of a temporal influence baseline.
And finally obtaining vegetation parameters to be solved, such as vegetation height, extinction coefficient, surface elevation, land volume amplitude ratio normalization parameter, time decoherence factor and the like through a nonlinear iterative algorithm.
To further clearly illustrate the multi-baseline PolInSAR vegetation parameter inversion method considering the time decoherence factor, the algorithm is verified by adopting P-waveband multi-baseline full polarization data of the Remningstorp region (58 degrees 28 'N and 13 degrees 38' E) provided by the BioSAR 2007 project. Note that this is for illustration only, and the present invention is not limited to data sources.
The terrain of the experimental area is flat, and the elevation is 120-145 meters; the vegetation mainly comprises artificial forests, and the tree species comprise spruce, pine, birch and the like, and the height is between 5 and 35 meters.
And E-SAR P-band SAR data are adopted for experiments. The experimental data adopts 9 scene airborne P wave band full polarization data, and 8 interference pairs can be formed by taking the number 0411 image as a main image and other images as auxiliary images. And carrying out data preprocessing by using PolSARpro software released by the European Bureau.
The method comprises the following specific steps:
step 1, polarization interference treatment:
the number 0411 image is used as a main image, the number 0106, 0107, 0302, 0303, 0304, 0407, 0408 and 0409 images are used as auxiliary images to form 8 interference baselines, the PolSARpro software is used for removing the ground effect and performing multi-view processing (the azimuth direction is 2: the distance direction is 1) on the main image and the auxiliary image respectively, and then the polarization complex coherence coefficient is obtained.
In the calculation example, each baseline adopts 2 phase maximum separation coherent optimal polarization modes, namely PDHigh and PDLow polarization modes. Finally, 8 baselines were generated, and 16 different polarization modes (PDHigh1, PDHigh2, …, PDHigh 8; PDLow1, PDLow2, PDLow3, superscript for the baseline number) were all independently observed.
Step 2, generating initial parameter values:
the step is mainly to provide an initial value of an unknown parameter for a nonlinear iterative algorithm.
Firstly, setting the initial value of the body amplitude ratio normalization parameter as 0, and setting the initial values of the time decorrelation factors of the vegetation body and the earth surface as 1. Then calculating the initial values of vegetation height, extinction coefficient and earth surface elevation by using a three-stage algorithm which is widely applied in the field of PolInSAR vegetation parameter inversion at present, wherein the specific method comprises the following steps:
1) and (3) straight line fitting: selection of PDHigh under the RVoG model framework8And PDLow8The polarization mode is linear fit in the complex plane (the 8 th baseline is selected as an example in this embodiment, and any other reasonable baseline selection mode is possible, and a specific baseline is not limited).
2) And (3) earth surface phase estimation: 2 intersection points exist between the straight line obtained by fitting in the step 1) and the complex plane unit circle, wherein 1 is a ground surface phase point. Because the distance between the earth surface phase point and the point of the in-vitro scattering dominant polarization mode is larger than the distance between the earth surface phase point and the point of the in-vitro scattering dominant polarization mode in the complex plane unit circle, the earth surface phase can be determined by adopting the following judgment formula
Figure GDA0003024000320000111
Figure GDA0003024000320000112
Wherein the content of the first and second substances,
Figure GDA0003024000320000113
respectively represents the complex coherence coefficient of the body scattering dominant channel and the complex coherence coefficient of the ground scattering dominant channel.
After the earth surface phase is determined, the earth surface elevation h can be directly calculated by the following formulagInitial value of (d):
Figure GDA0003024000320000121
3) vegetation height and extinction coefficient estimation: assuming bulk scattering dominates the polarization channel
Figure GDA0003024000320000122
Only contains vegetation layer scattering contribution, the earth volume amplitude ratio is 0, and the expression of pure body coherence is
Figure GDA0003024000320000123
A numerical calculation method is adopted to realize the purpose of giving a series of vegetation heights hvAnd an extinction coefficient sigma according toEquation (17) establishes a two-dimensional lookup table, finds a set of values with the minimum difference, namely the vegetation height hvAnd an estimated value of the extinction coefficient sigma, and as an initial value of the vegetation height and an initial value of the extinction coefficient, formula (17) is:
Figure GDA0003024000320000124
step 3, determining a baseline which is not affected by the temporal decorrelation factor:
since the larger the eccentricity of the elliptical region is, the smaller the influence of the temporal decorrelation factor is, in this embodiment, the elliptical eccentricity is used as a determination index, and one baseline with the largest elliptical eccentricity among all 8 baselines is used as a reference baseline, and is considered to be not influenced by the temporal decorrelation factor, while the other baselines are influenced by the temporal decorrelation factor, and both refer to the corresponding temporal decorrelation factor when interfering with respect to the reference baseline. The judgment formula is as follows (18):
Figure GDA0003024000320000125
wherein: ecc denotes elliptical regional eccentricity; 1,2, L,8 in superscript are baseline designations; max | | | represents taking the maximum value.
Step 4, multi-baseline PolInSAR vegetation parameter inversion algorithm considering time decoherence factors
When there is a temporal decoherence factor effect, the RVoG model is extended to obtain an RVoG extended model as shown in equation (19):
Figure GDA0003024000320000126
according to RVoG extended model considering time decoherence factor and the ratio mu of the amplitude of the ground body(ω)Normalization is carried out to obtain a normalized parameter x of the earth volume amplitude ratio1And then, constructing a nonlinear observation equation to invert parameters to be solved:
Figure GDA0003024000320000131
in this embodiment, assume that the finally selected reference baseline is the 8 th baseline, and therefore there is
Figure GDA0003024000320000133
The formula (20) thus varies as:
Figure GDA0003024000320000132
from the formula (21), the unknown parameter of the present embodiment is fixed h in the parameter inversiong,σ,hv,x1And 14 temporal decorrelation factors for the 1 st to 7 th baselines. Utilizing the vegetation height initial value, the extinction coefficient initial value, the surface elevation initial value, the land body amplitude ratio normalization parameter initial value and the time decoherence factor initial value in the step 3 and adopting a nonlinear iterative algorithm to finally obtain the vegetation height to be solved, the extinction coefficient, the surface elevation and the land body amplitude ratio normalization parameter x1Vegetation parameters such as time decoherence factors.
The nonlinear iterative algorithm adopted in the invention is the prior art, and an optimal solution is searched in a given solution space from an initial parameter value by giving the solution space and the initial parameter value.
Fig. 3 shows the inversion result of vegetation height obtained by the algorithm of the present invention. For quantitative analysis, the LiDAR vegetation height product is used herein as a reference. 563 block samples of 51 × 51 pixel size were taken within the experimental area and the corresponding average vegetation height was calculated for accuracy verification. FIG. 4 shows a plot cross-check of the inversion results of vegetation height and LiDAR vegetation height products, the vegetation height results obtained by the algorithm of the present invention have a corresponding RMSE of 3.53 meters and a higher accuracy.
The above embodiments are preferred embodiments of the present application, and those skilled in the art can make various changes or modifications without departing from the general concept of the present application, and such changes or modifications should fall within the scope of the claims of the present application.

Claims (4)

1. A multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors is characterized by comprising the following steps:
step 0, introducing a time decorrelation factor into the RVoG model to obtain an RVoG extended model considering the time decorrelation factor:
Figure FDA0003024000310000011
the vegetation scene is expressed by using a RVoG extended model, wherein omega represents the polarization state related to the baseline and the polarization mode,
Figure FDA0003024000310000012
representing the phase of the earth surface, gammaTVRepresenting the temporal decoherence factor, gamma, of the vegetation bodyTGRepresenting the surface time decorrelation factor, mu(ω)Representing the earth body amplitude ratio corresponding to the polarization state omega; gamma rayvRepresenting a pure decoherence coefficient as a function of vegetation height and extinction coefficient;
step 0.5, parameter normalization: earth amplitude ratio mu for interfering N base lines(ω)Normalizing in a complex plane to obtain the same normalized parameter x of the earth volume amplitude ratiorAnd normalizing the surface elevation parameters; wherein, the body amplitude ratio normalization parameter xrThe distance from the polarized complex coherence coefficient of the volume scattering dominant channel of the reference baseline r to the pure body decorrelation coefficient is referred to;
step 1, carrying out registration, flat ground removal, multi-vision and polarization interference processing on a multi-baseline polarization interference SAR image of vegetation to generate an interference pattern, and acquiring a multi-baseline multi-polarization complex phase correlation coefficient observed value gamma of the vegetation(ω)(ii) a Setting and generating N base lines, wherein each base line adopts M polarization modes;
step 2, setting the initial value of the body amplitude ratio normalization parameter as 0, and setting the initial values of the time decorrelation factors of the vegetation body and the earth surface as 1; then calculating an initial value of vegetation height, an initial value of extinction coefficient and an initial value of surface elevation by using the RVoG model and a three-stage vegetation height inversion algorithm;
step 3, selecting one of the N baselines as a reference baseline r, wherein the reference baseline has no influence of time decoherence, and the earth surface time decoherence factor gamma of the reference baselineTG1 and vegetation body time decoherence factor gammaTV1, and is fixed; in the RVoG extended model, the earth surface time decorrelation factors and vegetation body time decorrelation factors of other baselines except the reference baseline refer to corresponding time decorrelation factors when interfering relative to the reference baseline;
step 4, multi-baseline PolInSAR vegetation parameter inversion considering time decoherence factors;
for each base line, matching the M observation values obtained in the step 1 with the RVoG expansion model constructed in the step 0 to obtain an observation equation, and splitting a real part and an imaginary part to obtain 2MN split observation equations;
solving 2(N-1) +4 unknown parameters in 2MN observation equations by adopting a nonlinear iterative algorithm according to the initial values obtained in the step 2 and the time decoherence factors of the earth surface and the vegetation body of the reference base line obtained in the step 3: vegetation height, extinction coefficient, surface elevation, surface-to-body amplitude ratio normalization parameter, surface time decorrelation factor of other N-1 baselines except reference base line and vegetation body time decorrelation factor.
2. The method of claim 1, wherein the reference baseline r is selected by: acquiring complex phase coherence coefficient observation values of each baseline in different polarization states, wherein an elliptical area is formed in a complex plane unit circle; and calculating the eccentricity of each elliptical area, and taking the baseline corresponding to the maximum eccentricity as a reference baseline, wherein the reference baseline is the one of all the N baselines which is least influenced by the time decorrelation factor.
3. The method of claim 1, wherein the method for calculating the initial vegetation height, the initial extinction coefficient and the initial surface elevation using the RVoG model and the three-stage vegetation height inversion algorithm comprises:
step 1), straight line fitting:
randomly selecting a base line, and performing straight line fitting on the RVoG model in a complex plane by using complex coherence coefficient observation values of the base line in different polarization states to obtain a straight line shown in the following formula:
Figure FDA0003024000310000021
step 2), calculating the earth surface phase:
calculating 2 intersection points A and B of the straight line and the complex plane unit circle shown in the formula (2), and taking the intersection point, in which the distance between the intersection point and the point of the volume scattering dominant polarization mode is greater than the distance between the intersection point and the point of the ground scattering dominant polarization mode, as the earth surface phase point
Figure FDA0003024000310000022
Then using the earth's surface phase
Figure FDA0003024000310000023
Calculating the initial value of the surface elevation according to the following formula (3):
Figure FDA0003024000310000024
step 3), vegetation height and extinction coefficient estimation:
if the bulk scattering dominant polarization channel only contains the vegetation layer scattering contribution and the ground-to-volume amplitude ratio is 0, the expression of the pure-body coherence is
Figure FDA0003024000310000025
Figure FDA0003024000310000026
The complex coherence coefficient of the dominant channel of the volume scattering is expressed byHeight h of vegetation of concernvAnd a function of the extinction coefficient sigma expressed as
Figure FDA0003024000310000027
By numerical calculation, by giving a series of vegetation heights hvAnd an extinction coefficient sigma, establishing a two-dimensional lookup table according to a formula (4), and finding a group of values with the minimum difference, namely the vegetation height hvAnd an estimated value of the extinction coefficient sigma, and the estimated value is used as an initial vegetation height value and an initial extinction coefficient value:
Figure FDA0003024000310000031
4. the method of claim 3, wherein the baseline selected for performing the line fit is a reference baseline.
CN201910968899.9A 2019-10-12 2019-10-12 Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors Active CN110703220B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910968899.9A CN110703220B (en) 2019-10-12 2019-10-12 Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910968899.9A CN110703220B (en) 2019-10-12 2019-10-12 Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors

Publications (2)

Publication Number Publication Date
CN110703220A CN110703220A (en) 2020-01-17
CN110703220B true CN110703220B (en) 2021-06-22

Family

ID=69198638

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910968899.9A Active CN110703220B (en) 2019-10-12 2019-10-12 Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors

Country Status (1)

Country Link
CN (1) CN110703220B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111965645B (en) * 2020-08-10 2022-04-05 中南大学 Multi-baseline vegetation height inversion method and device considering geometric constraints
CN112363161B (en) * 2020-10-27 2022-12-20 中南大学 Vegetation vertical structure and under-forest topography inversion method and device based on scattering mechanism decomposition
CN113945927B (en) * 2021-09-17 2022-09-06 西南林业大学 Forest canopy height inversion method through volume scattering optimization
CN113945926B (en) * 2021-09-17 2022-07-08 西南林业大学 Forest canopy height inversion method improved through underestimation compensation
CN117434517B (en) * 2023-12-21 2024-02-27 西南林业大学 Forest canopy height estimation method based on extinction coefficient optimization

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102401892A (en) * 2010-09-19 2012-04-04 中国科学院电子学研究所 System performance assessment method of polarized interferometric synthetic aperture radar
CN103235301A (en) * 2013-05-14 2013-08-07 中南大学 Polarimetric synthetic aperture radar interferometry (POLInSAR) vegetation height inversion method based on complex field adjustment theory
CN109738895A (en) * 2019-01-31 2019-05-10 中南大学 A kind of building and inversion method based on second-order Fourier gear-Legnedre polynomial vegetation height inverse model

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102401892A (en) * 2010-09-19 2012-04-04 中国科学院电子学研究所 System performance assessment method of polarized interferometric synthetic aperture radar
CN103235301A (en) * 2013-05-14 2013-08-07 中南大学 Polarimetric synthetic aperture radar interferometry (POLInSAR) vegetation height inversion method based on complex field adjustment theory
CN109738895A (en) * 2019-01-31 2019-05-10 中南大学 A kind of building and inversion method based on second-order Fourier gear-Legnedre polynomial vegetation height inverse model

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Estimation of Pine Forest Height and Underlying DEM Using Multi-Baseline P-Band PolInSAR Data;Haiqiang Fu 等;《REMOTE SENSING 》;20161005;第8卷(第10期) *
Underlying topography extraction over forest areas from multi-baseline PolInSAR data;Haiqiang Fu 等;《JOURNAL OF GEODESY》;20171125;第92卷(第7期) *
多基线POLINSAR植被高度与植被下地形估计;史磊;《中国博士学位论文全文数据库 基础科学辑》;20130915 *
星载重轨极化干涉SAR反演森林植被高度;李震 等;《中国科学:地球科学》;20140420 *

Also Published As

Publication number Publication date
CN110703220A (en) 2020-01-17

Similar Documents

Publication Publication Date Title
CN110703220B (en) Multi-baseline PolInSAR vegetation parameter inversion method considering time decoherence factors
CN110988879B (en) Vegetation parameter inversion method, terminal equipment and storage medium
Lavalle et al. Extraction of structural and dynamic properties of forests from polarimetric-interferometric SAR data affected by temporal decorrelation
Garestier et al. Pine forest height inversion using single-pass X-band PolInSAR data
CN113205475B (en) Forest height inversion method based on multi-source satellite remote sensing data
CN109738895B (en) Method for constructing and inverting vegetation height inversion model based on second-order Fourier-Legendre polynomial
CN111766577B (en) Power transmission line channel tree height inversion method based on three-stage algorithm P wave band
CN109388887B (en) Quantitative analysis method and system for ground settlement influence factors
CN109031300B (en) Method and system for monitoring deformation of dangerous rock mass by synthetic aperture radar
CN102401892B (en) System performance assessment method of polarized interferometric synthetic aperture radar
CN110487241B (en) Method for extracting elevation control points of building area by satellite laser height measurement
d’Alessandro et al. Interferometric ground cancellation for above ground biomass estimation
Yang et al. Tropical forest height retrieval based on P-band multibaseline SAR data
CN110378896B (en) TomosAR vegetation pest and disease damage monitoring method and device based on polarization coherence
CN111352109B (en) Vegetation height inversion method and device based on two-scene SAR (synthetic aperture radar) image
CN111965645B (en) Multi-baseline vegetation height inversion method and device considering geometric constraints
CN111273289A (en) Desert parameter inversion method, device, equipment and storage medium
Lavalle et al. Demonstration of repeat-pass POLINSAR using UAVSAR: The RMOG model
Shen et al. Orientation angle calibration for bare soil moisture estimation using fully polarimetric SAR data
Babu et al. Tree canopy height estimation using multi baseline RVOG inversion technique
Kugler et al. Forest characterisation by means of TerraSAR-X and TanDEM-X (polarimetric and) interferometric data
CN110399832B (en) TomosAR vegetation pest and disease damage monitoring method and device based on coherence
Prudente et al. Relationship between SAR/Sentinel-1 polarimetric and interferometric data with biophysical parameters of agricultural crops
Souissi et al. Polarimetric SAR data correction and terrain topography measurement based on the radar target orientation angle
Neumann et al. Improvement of vegetation parameter retrieval from polarimetric SAR interferometry using a simple polarimetric scattering model

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