A method for soil moisture inversion in vegetated areas considering polarimetric scattering information

Abstract

The present invention discloses a method for retrieving soil moisture in vegetated areas considering polarimetric scattering information, including extracting soil moisture retrieval characteristic parameters in vegetated areas based on polarimetric target decomposition; analyzing and obtaining an optimized combination of characteristic parameters; constructing and solving a soil moisture retrieval model integrating polarimetric information and radar backscattering information. In the method provided by the present invention, the construction of the model starts from the target scattering mechanism. Through the decomposition and correction of polarimetric characteristic parameters, the influence of vegetation coverage is effectively removed. By combining independent component analysis to obtain an optimized combination of retrieval characteristic parameters, the effective observation information of fully polarimetric SAR data is fully utilized, and reliable retrieval of soil moisture in vegetated areas is realized without the need for other auxiliary data.

Description

Technical Field

[0001] The present invention relates to the technical field of synthetic aperture radar for monitoring soil moisture, and particularly to a method for dynamically monitoring soil moisture in vegetation-covered areas considering multi-polarization SAR characteristics. Background Technique

[0002] Soil moisture is an important parameter characterizing the surface soil state and is widely used in fields such as agricultural soil moisture assessment, flood prediction, and meteorological forecasting. The ground monitoring method of soil moisture has the advantage of high precision. However, it can only obtain the soil water content at discrete points, and there are deficiencies in the dynamic monitoring of soil moisture, such as low update frequency and high consumption of manpower and material resources. Remote sensing technology, with its advantages of non-contact, large-scale, and high-timeliness observation, has been importantly applied in the quantitative monitoring of soil moisture. Especially, microwave remote sensing technology has the ability to observe all day and all weather, and there is a strict theoretical basis between the observed quantity and soil moisture, which provides a solid foundation for quantitatively retrieving soil moisture using microwave remote sensing data.

[0003] Synthetic Aperture Radar (SAR) has the advantages of high spatio-temporal resolution, multi-polarization, and multi-band earth observation. In particular, the SAR observation information is sensitive to the response of soil surface characteristics, can effectively characterize the soil surface parameter information, and effectively overcomes the deficiencies of traditional soil moisture monitoring methods. It has been widely studied in the inversion of soil surface parameters and provides an effective technical method for large-scale dynamic soil moisture monitoring. However, the inversion of soil water content using radar technology is affected by factors such as vegetation cover and soil surface roughness. Currently, significant achievements have been made in the inversion research of bare soil moisture. Applying these methods directly to the inversion analysis of soil moisture in vegetation-covered areas has certain errors. Therefore, for vegetation-covered areas, it is necessary to consider the influence of vegetation on radar scattering information to obtain the scattering information characterizing the soil surface characteristics, and then establish a corresponding soil moisture inversion model.

[0004] The different polarization radar backscattering information and target polarization scattering information obtained by multi-polarization SAR satellites are important components of SAR observations. Polarization scattering information can effectively characterize the geometric and physical properties of the observed target. By performing polarization target decomposition on full-polarization SAR data, polarization scattering information characterizing different scattering mechanisms of the target can be obtained, providing effective characteristic parameters for soil inversion in vegetated areas. Radar backscattering information is sensitive to the response of soil dielectric properties and has been widely used in the quantitative inversion of soil moisture as an important characteristic parameter characterizing the soil surface moisture. Aiming at the influence of vegetation cover, by separately extracting the vegetation and soil surface characteristic parameters through polarization scattering information, and effectively removing the vegetation information, the backscattering information and polarization information characterizing the soil surface properties are obtained, providing an important idea for quantitatively inverting soil surface moisture by integrating polarization information and radar backscattering information.

[0005] Although there are many existing methods for extracting soil water content using SAR observation information, in the face of the influence of vegetation cover, there is a lack of a method to effectively remove or suppress the influence of vegetation cover from the perspective of radar scattering mechanism to obtain radar scattering information characterizing the soil surface properties. And currently, single polarization scattering information or radar backscattering information is mostly used for soil moisture inversion research, lacking the collaborative application of polarization information and radar backscattering information to improve the reliability of soil moisture inversion. Therefore, there is a need in this field for a method that can effectively remove the influence of vegetation cover on soil moisture inversion, obtain the soil surface scattering characteristics in vegetated areas from the perspective of radar scattering mechanism, realize the optimal combination of inversion characteristic parameters by collaborating polarization scattering information and radar backscattering information, and construct an accurate and reliable method for quantitatively inverting soil moisture under vegetation cover conditions. Summary of the Invention

[0006] To solve the above technical problems, the object of the present invention is to provide a method for retrieving soil moisture in vegetated areas considering polarimetric scattering information, which is a method for retrieving soil moisture in vegetated areas based on the combined radar backscattering information and polarimetric scattering information of full-polarimetric SAR data; first, perform non-coherent target decomposition on the full-polarimetric radar data to obtain the surface scattering, even scattering, and volume scattering information of the targets in the vegetated area; starting from the polarimetric scattering mechanisms of vegetation and soil, considering that vegetation is mainly characterized by volume scattering characteristics in the polarimetric dimension, in order to effectively remove the influence of vegetation cover on soil moisture retrieval, it is necessary to correct the volume scattering information obtained by target decomposition. Therefore, use the multi-polarimetric SAR observation information to extract the normalized radar vegetation index, describe the contribution component of vegetation to the volume scattering information through the normalized radar vegetation index, and then use this index to correct the volume scattering information, thereby obtaining the polarimetric scattering characteristic information with the influence of vegetation removed. Secondly, use the water cloud model combined with the normalized radar vegetation index to extract the multi-polarimetric radar backscattering coefficients characterizing the soil surface characteristics. Integrate the radar backscattering coefficients and polarimetric scattering information that effectively characterize the soil surface scattering characteristics, use independent component analysis to obtain an optimized combination of soil inversion characteristic parameters, effectively remove the redundancy between the multi-polarimetric radar backscattering coefficients and polarimetric scattering information, improve the reliability and robustness of model construction, provide optimized characteristic parameters for the quantitative inversion of soil water content in vegetated areas, and finally obtain the soil moisture content in vegetated areas through a multi-variable non-linear regression model, providing technical support for the dynamic monitoring of soil moisture in vegetated areas under different phenological conditions supported by full-polarimetric SAR.

[0007] The object of the present invention is achieved by the following technical solutions:

[0008] A method for retrieving soil moisture in vegetated areas considering polarimetric scattering information, comprising the following steps:

[0009] Step 1: Extract the soil moisture retrieval characteristic parameters in vegetated areas based on polarimetric target decomposition;

[0010] Step 2: Analyze and obtain an optimized combination of characteristic parameters;

[0011] Step 3: Construct and solve a soil moisture retrieval model integrating polarimetric information and radar backscattering information.

[0012] Compared with the prior art, one or more embodiments of the present invention may have the following advantages:

[0013] The present invention is directed to the quantitative inversion of soil moisture in vegetated areas. From the perspective of the polarization scattering mechanism, the decomposition and optimization of the scattering mechanisms of vegetation and soil are carried out. The radar backscattering and polarization scattering information are fully utilized to characterize the surface characteristics of vegetation and soil. By using different polarimetric SAR observations, the influence of vegetation cover on the inversion of soil surface moisture is eliminated. Furthermore, an optimized combination of radar characteristic parameters for characterizing soil surface characteristics is obtained. A soil moisture inversion model is constructed using simulated data and measured data, and reliable soil surface moisture information is obtained by combining the input of multiple optimized characteristic parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 is a flowchart of a method for inverting soil moisture in vegetated areas considering polarization scattering information;

[0015] Figure 2 is a schematic diagram of a multi-characteristic parameter inversion model combining polarization information and radar backscattering information;

[0016] Figure 3 is a schematic diagram of the decomposition of the scattering mechanism of the vegetated ground surface. DETAILED DESCRIPTION OF THE INVENTION

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with embodiments and the accompanying drawings.

[0018] As Figure 1 shown, a method for inverting soil moisture in vegetated areas considering polarization scattering information includes the following steps:

[0019] Step 1: Optimized extraction of characteristic parameters for inverting soil moisture in vegetated areas based on polarization target decomposition;

[0020] Step 1.1: For fully polarimetric SAR data, an incoherent polarization target decomposition method is used to obtain polarization scattering information characterizing different characteristics of the target

[0021] Polarization target decomposition is a method for extracting typical scattering characteristics of fully polarimetric radar data, which can extract characteristic components representing different scattering mechanisms. Among them, the model-based incoherent target decomposition method, Freeman decomposition, can decompose fully polarimetric data into surface scattering, double-bounce scattering, and volume scattering information. Using the polarization target decomposition method to obtain the scattering information describing the characteristics of soil and vegetation, from the perspective of the scattering mechanism, the radar scattering information in the vegetated area includes vegetation scattering, soil scattering, and vegetation-soil superposition scattering. The decomposition of the surface scattering mechanism in the vegetated area is as Figure 3 shown. By performing polarization target decomposition processing on the fully polarimetric SAR data, the initial polarization characteristic parameters for soil moisture inversion are obtained.

[0022] Fully polarimetric SAR data can acquire the backscattering information of the surface corresponding to different polarization modes. The fully polarimetric scattering matrix S is composed of different polarization complex scattering coefficients S pq , where p and q represent horizontal and vertical transmit / receive polarizations.

[0023]

[0024] In the formula, S hh represents the scattering component of the horizontal transmit and horizontal receive polarization mode. The definitions of S hv , S vh and S vv are similar to that of S hh .

[0025] The coherence matrix T3 characterizing the polarization characteristics is constructed by the Pauli basis vectors k and is used to describe the polarization scattering information of the target:

[0026]

[0027] T3 = <k·k *T > (3)

[0028] In the formula, the superscript *, the superscript T and <> represent complex conjugate, matrix transpose and averaging processing respectively.

[0029] In order to extract the separate scattering contributions of the target ground objects, the Freeman polarization target decomposition method is used to process the coherence matrix, and the coherence matrix is decomposed into three components: surface scattering, double-bounce scattering and volume scattering. The Freeman decomposition components are used to characterize the polarization scattering information of the target:

[0030]

[0031] In the formula, f s , f d and f v represent the surface scattering, double-bounce scattering and volume scattering amplitudes respectively, and β and α represent the surface scattering and double-bounce scattering parameters respectively;

[0032] The trace of the coherence matrix is used to describe the scattering intensities of the surface scattering, double-bounce scattering and volume scattering components.

[0033] P s = f s (1 + |β| 2 ) (5)

[0034] P d = f d (1 + |α| 2 ) (6)

[0035] P v = fv (7)

[0036] In the formula, Ps, Pd and Pv respectively represent the surface scattering, even-order scattering and volume scattering characteristics of the target.

[0037] Step 1.2: For the vegetation-covered area, use the fully polarized radar observation information to extract the normalized radar vegetation index

[0038] Vegetation cover is an important factor affecting the SAR inversion of soil surface moisture. To address this problem, index information characterizing vegetation characteristics is extracted using fully polarized SAR data. The normalized radar vegetation (Normalized Radar Vegetation Index, NRVI) is defined to extract the vegetation scattering information in the vegetation-covered area. The normalized radar vegetation index is expressed as:

[0039] NRVI = (RVI - RVI min ) / (RVI max - RVI min ) (8)

[0040] In the formula, RVI is the radar vegetation index, and RVI max and RVI min respectively represent the maximum and minimum radar vegetation indices extracted within the region. The normalized radar vegetation index has good applicability for extracting the vegetation cover information within the region.

[0041] Based on the fully polarized SAR data, the radar vegetation index sensitive to the vegetation cover information is extracted through polarization feature combination, and the expression relationship is as follows:

[0042] RVI = f(σ HH , σ HV , σ VH , σ VV ) (9)

[0043] In the formula, σ HH and σ VV respectively represent the horizontal polarization and vertical polarization radar backscattering information, and σ HV and σ VH respectively represent the cross-polarization radar backscattering information (intensity) of horizontal polarization transmitting and vertical polarization receiving and vertical polarization transmitting and horizontal polarization receiving.

[0044] The NRVI can describe the vegetation coverage information of each pixel in the area. The smaller the NRVI value, the lower the vegetation coverage, and the larger the NRVI value, the higher the vegetation coverage. As the vegetation coverage increases, the NRVI gradually increases. Thus, the effective extraction of vegetation coverage information can be directly achieved using fully polarized SAR images, avoiding the need for other auxiliary data, while ensuring the temporal consistency and spatial scale consistency of vegetation information extraction. Using the NRVI to analyze the influence of vegetation on radar polarization scattering information and backscattering coefficient, and then correcting and processing the radar scattering information in the vegetation coverage area, separating the vegetation contribution part, and obtaining the polarization scattering information and radar backscattering coefficient characterizing the soil surface characteristics, so as to be used for soil moisture inversion analysis in the vegetation coverage area.

[0045] Step 1.3 Based on the polarization scattering mechanism characteristics of vegetation and soil surface, the vegetation scattering information is mainly characterized by the volume scattering component. Combining the normalized radar vegetation index NRVI to eliminate the contribution of vegetation to the volume scattering characteristics, obtaining the scattering mechanism characterizing the soil polarization scattering characteristics, and then used for soil moisture quantitative inversion analysis. The processing flow is as Figure 2 shown.

[0046] The normalized radar vegetation index NRVI can effectively characterize the vegetation scattering information, which is calculated from multi-polarization radar data and can describe the vegetation scattering characteristics of each pixel. In the radar scattering mechanism, vegetation is mainly characterized by volume scattering information and acts as a noise factor in soil surface moisture inversion. Therefore, the NRVI parameter is used to process the vegetation volume scattering information and correct the characteristic parameters for soil moisture inversion. Define the volume scattering information after eliminating the vegetation effect as Pv’, and the volume scattering information obtained from the initial polarization target decomposition as Pv. Use the NRVI to correct the volume scattering information, remove the vegetation volume scattering contribution, and the corrected volume scattering characteristics are expressed as the following non-linear conversion relationship:

[0047] Pv’ = (1 - NRVI) / (1 + NRVI) × Pv (10)

[0048] Thus, through multi-polarization SAR observation information, the influence of vegetation coverage is eliminated, and the polarization scattering information characterizing the soil surface characteristics is extracted to obtain effective inversion characteristic parameters.

[0049] Through the above optimization processing of polarization characteristic parameters, the polarization parameter information characterizing the soil surface scattering characteristics is obtained on the basis of effectively eliminating the influence of vegetation coverage.

[0050] Step 2 Use independent component analysis to obtain an optimized combination of characteristic parameters

[0051] Using the radar backscattering coefficient and polarization scattering information that characterize the soil surface properties as the input of the inversion characteristic parameters, and using the independent component analysis method to extract the optimal combination of characteristic parameters that are sensitive to soil moisture response.

[0052] Step 2.1 Use the normalized radar vegetation index NRVI to eliminate the influence of vegetation on the multi-polarization radar backscattering coefficient, and obtain the multi-polarization radar backscattering coefficient that characterizes the soil surface characteristics.

[0053] The total radar backscattering in the vegetation-covered area is described as the superposition of the soil surface scattering contribution, the soil scattering under vegetation cover, and the vegetation scattering contribution, as Figure 3 shown, where part of the soil surface scattering is affected by the attenuation of the vegetation layer. The specific expression of this model is:

[0054] σ o = NRVI×(σ veg + τ 2 σ soil )+(1 - NRVI)×σ soil (11)

[0055] In the formula, σ o represents the total backscattering information of the vegetation-covered ground surface, σ veg and σ soil represent the vegetation scattering and soil surface scattering information respectively, and τ represents the vegetation attenuation coefficient. By introducing the normalized radar vegetation index NRVI, the soil surface scattering, vegetation scattering, and soil surface scattering characteristics under vegetation cover in the vegetation-covered area are fully considered, and the model has good applicability for both densely vegetated areas and sparsely vegetated areas. When the vegetation is completely covered, this model is the representation form of the traditional water cloud model for the vegetation-covered area; when the ground surface is bare soil, the total backscattering information of the ground surface is the soil surface scattering information. Through the above processing, the independent decomposition of the radar backscattering information in the vegetation-covered area is realized, and then the radar backscattering coefficient that characterizes the soil surface scattering characteristics is obtained.

[0056] For the full-polarization data of China's L-band differential interferometric SAR satellite, in the vegetation-covered area, the vegetation effects are removed for different polarization observables respectively, and the multi-polarization radar backscattering coefficient that characterizes the soil surface scattering characteristics is obtained.

[0057]

[0058] In the formula, σ HHo , σ HVo , σ VHo and σ VVo represent the total radar backscattering coefficients corresponding to different polarization modes in the vegetation-covered area respectively, σ HHveg , σ HVveg , σVHveg and σ VVveg respectively represent the vegetation scattering information of different polarization modes, and NRVI is the normalized radar vegetation index; the multi-polarized radar backscattering coefficient σ of the soil surface after removing the vegetation influence is obtained by using formula (12) HHsoil , σ HVsoil , σ VHsoil and σ VVsoil are used to construct the soil moisture inversion model.

[0059] Step 2.2 uses the polarization scattering information after removing the vegetation influence and the multi-polarized radar backscattering coefficient as inputs, and through independent component analysis, extracts an optimized combination of characteristic parameters sensitive to soil moisture response, effectively removing the redundancy between the polarization scattering information and the multi-polarized radar backscattering information, and providing effective inputs for the construction of the inversion model.

[0060] Using fully polarized L-band differential interferometric SAR satellite data, the multi-polarized radar backscattering coefficient and polarization scattering information characterizing the soil surface properties after removing the vegetation influence are extracted. Aiming at the information redundancy between multi-dimensional input parameters, independent component analysis is used to obtain an optimized combination of characteristic parameters for soil moisture inversion.

[0061] [Par(1), Par(i)...]=F(σ HHsoil , σ HVsoil , σ VHsoil , σ VVsoil , Ps, Pd, Pv')(13)

[0062] In the formula, Par(1) and Par(i) respectively represent the optimized characteristic parameters for soil moisture inversion obtained by independent component analysis extraction, σ HHsoil , σ HVsoil , σ VHsoil and σ VVsoil respectively represent the multi-polarized radar backscattering coefficients of the soil surface after removing the vegetation influence, and Ps, Pd and Pv' respectively represent the polarization target decomposition information of the soil surface scattering, even-order scattering and volume scattering after vegetation removal treatment. Through the above processing, an inversion characteristic parameter combination integrating the radar backscattering coefficient and polarization scattering information of the soil surface is obtained, providing important input characteristics for the reliable acquisition of soil moisture.

[0063] In the above processing, F represents the independent component analysis process. Facing the radar backscattering coefficients of different polarization modes and the polarization target decomposition information for removing the influence of vegetation, and aiming at the redundancy between multi-dimensional information and the influence of noise, the input parameters are processed by independent component analysis. A combination of characteristic parameters sensitive to soil moisture response is extracted, the potential noise influence and parameter redundancy information are effectively removed, and independent characteristic parameter information is obtained, providing effective input parameters for soil moisture inversion.

[0064] Step 3: Construction and solution of the soil moisture inversion model integrating radar backscattering information and polarization information

[0065] With the optimized characteristic parameters input integrating radar backscattering information and polarization target decomposition information, a function model between soil surface parameters and the combination of SAR characteristic parameters is constructed. Through the model training with measured data, reliable acquisition of regional soil surface moisture is realized. The overall technical process is as Figure 2 shown.

[0066] Step 3.1: Using the optimized characteristic parameters integrating polarization scattering information and radar backscattering information, construct a multivariate nonlinear function model between soil surface moisture and SAR characteristic parameters.

[0067] m v = G[Par(1), Par(i) ...] (14)

[0068] In the formula, m v represents the soil moisture content on the ground surface, and G represents the nonlinear conversion relationship between the characteristic combination of radar backscattering information and polarization information and soil moisture.

[0069] Step 3.2: Using the measured soil surface parameters on-site as the model driving data, construct the conversion relationship between the optimized characteristic parameters and soil moisture;

[0070] Through the optimized extraction of the radar backscattering coefficients and polarization scattering information of the full-polarization SAR data, effective input parameters are provided for the construction of the soil moisture inversion model. Combining the measured data of the soil moisture at the sample points, a function model between multi-dimensional radar observation information and soil moisture is constructed to complete the training and construction of the soil moisture inversion model.

[0071] Step 3.3: Using the optimized characteristic parameters integrating the full-polarization SAR backscattering information and polarization information as the input, through the multivariate nonlinear driving model, obtain the soil surface moisture information, and realize the reliable inversion of the soil surface moisture in the vegetation-covered area by synergistically using the polarization scattering information and radar backscattering information.

[0072] Taking the multi-polarization and multi-dimensional radar characteristic parameters in the vegetation-covered area as input and combining with a multivariate non-linear model to achieve the acquisition of soil surface moisture in the vegetation-covered area, providing effective soil moisture data products for farmland soil moisture monitoring, weather forecasting, flood prediction, etc.

[0073] Although the embodiments disclosed in the present invention are as above, the above content is only an embodiment adopted for the convenience of understanding the present invention and is not intended to limit the present invention. Any person skilled in the art within the technical field to which the present invention pertains may make any modifications and changes in the form of implementation and details without departing from the spirit and scope disclosed by the present invention. However, the scope of patent protection of the present invention shall still be subject to the scope defined by the appended claims.

Claims

1. A method for retrieving soil moisture in vegetated areas considering polarimetric scattering information, characterized in that, The method includes the following steps: Step 1: Extract the vegetation-covered area soil moisture inversion characteristic parameters based on polarimetric target decomposition; Step 2: Analyze and obtain the optimized combination of characteristic parameters; Step 3: Construct and solve the soil moisture inversion model that synthesizes polarimetric information and radar backscattering information; The said Step 1 includes: Step 1.1: For fully polarimetric SAR data, adopt non-coherent polarimetric target decomposition to obtain the polarimetric scattering information of different surface characteristics, and from the perspective of scattering mechanism, perform polarimetric decomposition on the radar echo information in the vegetation-covered area. The radar echo information in the vegetation-covered area includes vegetation body scattering information, soil surface scattering information, and soil scattering information attenuated by vegetation; Step 1.2: For the vegetation-covered area, utilize the fully polarimetric radar observation information to extract the normalized radar vegetation index; Step 1.3: Aiming at the difference in polarimetric scattering characteristics between vegetation and the soil surface, based on the volume scattering characteristics of vegetation and combined with the normalized radar vegetation index, correct the volume scattering characteristic parameters to obtain the polarimetric scattering information characterizing the soil surface characteristics; In Step 1.3, use the NRVI parameter to process the vegetation body scattering information, correct the characteristic parameters for soil moisture inversion, define the volume scattering information excluding the vegetation effect as Pv’, the volume scattering information obtained by the initial polarimetric target decomposition as Pv, use NRVI to correct the volume scattering information, remove the vegetation body scattering contribution, and the corrected volume scattering characteristic is expressed as the following non-linear conversion relationship: Pv’ = (1 - NRVI) / (1 + NRVI) × Pv (10) Thus, through the multi-polarimetric SAR observation information, the influence of vegetation coverage is removed, the polarimetric scattering information characterizing the soil surface characteristics is extracted, and effective inversion characteristic parameters are obtained; The said Step 2 includes: Step 2.1: Use the normalized radar vegetation index NRVI to remove the influence of vegetation on the multi-polarimetric radar backscattering coefficient, and obtain the multi-polarimetric radar backscattering coefficient characterizing the soil surface characteristics; Step 2.2: Take the polarimetric scattering information and multi-polarimetric radar backscattering coefficient after removing the vegetation influence as inputs, analyze and extract the optimized combination of characteristic parameters sensitive to soil moisture response, and remove the redundancy between the polarimetric scattering information and multi-polarimetric radar backscattering information; In the said Step 2.2, use the fully polarimetric L-band differential interferometric SAR satellite data to extract the multi-polarimetric radar backscattering coefficient and polarimetric scattering information characterizing the soil surface characteristics after removing the vegetation influence; for the information redundancy between multi-dimensional input parameters, use independent component analysis to obtain the optimized characteristic parameter combination for soil moisture inversion: [Par(1), Par(i)...] = F(σ HHsoil , σ HVsoil , σ VHsoil , σ VVsoil , Ps, Pd, Pv') wherein, Par(1) and Par(i) respectively represent the optimized characteristic parameters for soil moisture inversion obtained by independent component analysis extraction, σ HHsoil , σ HVsoil , σ VHsoil and σ VVsoil respectively represent the soil surface multi-polarized radar backscattering coefficients after removing the vegetation influence, and Ps, Pd and Pv' respectively represent the polarization target decomposition information of soil surface scattering, even-order scattering and volume scattering after vegetation removal treatment; The said Step 3 includes: Step 3.1: Use the optimized characteristic parameters that synthesize polarimetric scattering information and radar backscattering information to construct a multivariate non-linear function model between the soil surface moisture and SAR characteristic parameters; Step 3.2: Take the measured soil surface parameters as the model driving data to construct the conversion relationship between the optimized characteristic parameters and soil moisture; Step 3.3 uses the optimized characteristic parameters of the full-polarization SAR backscattering information and polarization information as inputs, and through a multivariate non-linear driving model, obtains the soil surface moisture information, realizing the reliable inversion of the soil surface moisture in the vegetation-covered area for the collaborative polarization scattering information and radar backscattering information.

2. The method for retrieving soil moisture in vegetated areas considering polarimetric scattering information according to claim 1, characterized in that In step 1.1, the full-polarization SAR data can obtain the surface backscattering information corresponding to different polarization modes. The full-polarization scattering matrix S is composed of different polarization complex scattering coefficients S pq where p and q represent horizontal and vertical transmit / receive polarizations: Where, S hh represents the scattering component of the horizontal transmission and horizontal reception polarization mode, and the definitions of S hv , S vh and S vv are similar to those of S hh ; The coherence matrix T3 characterizing the polarization characteristics is constructed by the Pauli basis vector k and is used to describe the polarization scattering information of the target: T3 = <k·k *T > (3) In the formula, the superscripts *, T, and < > represent complex conjugate, matrix transpose, and averaging processing respectively; In order to extract the individual scattering contributions of the target ground objects, the Freeman polarization target decomposition method is used to process the coherence matrix, and the coherence matrix is decomposed into three components: surface scattering, double-bounce scattering, and volume scattering; the Freeman decomposition components are used to characterize the polarization scattering information of the target: where f s , f d and f v represent the surface scattering, even-order scattering, and volume scattering amplitudes respectively, and β and α represent the surface scattering and even-order scattering parameters respectively; The trace of the coherence matrix is used to describe the scattering intensities of the surface scattering, double-bounce scattering, and volume scattering components; P s = f s (1 + |β| 2 ) (5) P d = f d (1 + |α| 2 ) (6) P v = f v (7) In the formula, Ps, Pd, and Pv represent the surface scattering, double-bounce scattering, and volume scattering characteristic information of the target respectively.

3. The method for retrieving soil moisture in vegetated areas considering polarimetric scattering information according to claim 1, characterized in that In the said Step 1.2, the vegetation scattering information in the vegetation-covered area is extracted by defining the normalized radar vegetation, and the normalized radar vegetation index is expressed as: NRVI=(RVI - RVI min ) / (RVI max -RVI min ) (8) where RVI is the radar vegetation index, and RVI max and RVI min represent the maximum and minimum radar vegetation indices extracted within the region, respectively; Based on the full-polarization SAR data, the radar vegetation index sensitive to the vegetation coverage information is extracted through polarization feature combination, and the expression relationship is as follows: RVI = f(σ HH , σ HV , σ VH , σ VV ) (9) where σ HH and σ VV represent the radar backscattering information of horizontal polarization and vertical polarization respectively, and σ HV and σ VH represent the cross-polarization radar backscattering information of horizontal polarization transmitting and vertical polarization receiving, and vertical polarization transmitting and horizontal polarization receiving respectively.

4. The method for retrieving soil moisture in vegetation-covered areas considering polarimetric scattering information according to claim 1, characterized in that In Step 2.1, the total radar backscattering in the vegetation-covered area is described as the superposition of the soil surface scattering contribution, the soil scattering under the vegetation coverage, and the vegetation scattering contribution. Among them, part of the soil surface scattering is affected by the attenuation of the vegetation layer, and the soil moisture inversion model is specifically expressed as: σ o = NRVI × (σ veg + τ 2 σ soil ) + (1 - NRVI) × σ soil (11) where, σ o represents the total backscattering information of the vegetated surface, σ veg and σ soil represent the vegetation scattering and soil surface scattering information respectively, and τ represents the vegetation attenuation coefficient; Facing the full-polarization data of the L-band differential interferometric SAS satellite, the vegetation effects are removed from different polarization observables in the vegetation-covered area respectively, and the multi-polarization radar backscattering coefficients characterizing the soil surface scattering characteristics are obtained: σ HHo = NRVI × (σ HHveg + τ 2 σ HHsoil ) + (1 - NRVI) × σ HHsoil σ HVo = NRVI × (σ HVveg + τ 2 σ HVsoil ) + (1 - NRVI) × σ HVsoil σ VHo = NRVI × (σ VHveg + τ 2 σ VHsoil ) + (1 - NRVI) × σ VHsoil σ VVo = NRVI × (σ VVveg + τ 2 σ VVsoil ) + (1 - NRVI) × σ VVsoil where σ HHo , σ HVo , σ VHo and σ VVo represent the total radar backscattering coefficients corresponding to different polarization modes in the vegetation-covered area, respectively. σ HHveg , σ HVveg , σ VHveg and σ VVveg represent the vegetation scattering information of different polarization modes, respectively. NRVI is the normalized radar vegetation index. The multi-polarization radar backscattering coefficients σ HHsoil , σ HVsoil , σ VHsoil and σ VVsoil of the soil surface after removing the vegetation influence are obtained by using this formula, and then are used for constructing the soil moisture inversion model.

5. The method for retrieving soil moisture in vegetation-covered areas considering polarimetric scattering information according to claim 1, characterized in that, The calculation formula of the soil water content in the said Step 3.1 is: m v = G[Par(1), Par(i)...] wherein, m v represents the surface soil water content, G represents the non-linear conversion relationship between the combined characteristics of radar backscattering information and polarization information and soil moisture, and Par(1) and Par(i) respectively represent the optimized characteristic parameters for soil moisture inversion obtained by independent component analysis.

Images