A Soil Salinity Inversion Method and System Based on Polarization Remediation and Active-Passive Counterbalancing

Abstract

This invention proposes a soil salinity inversion method based on polarization remediation and active-passive offsetting, comprising: S1, constructing a normalized polarization difference rate reflecting the anisotropy of surface electromagnetic radiation and a spatial gradient factor reflecting the spatial non-uniformity of the surface; constructing a polarization driving factor characterizing the composite extinction capacity of the land surface; S2, constructing a dynamically coupled coherent compensation operator; performing energy compensation remediation on the initial reflectance to obtain the corrected reflectance analytically; S3, converting the corrected reflectance into the total dielectric constant modulus characterizing the comprehensive polarization intensity of the soil medium; S4, analyzing the real part of the dielectric constant; extracting the imaginary part of the dielectric constant; S5, converting the imaginary part of the dielectric constant into soil salinity content. This invention also proposes a corresponding soil salinity inversion system based on polarization remediation and active-passive offsetting. This invention achieves high-precision, highly generalized quantitative inversion of soil salinity without the need for external prior data constraints.

Description

Technical Field

[0001] This invention relates to the field of soil salinity inversion technology, and in particular to a soil salinity inversion method and system based on polarization remediation and active-passive offsetting. Background Technology

[0002] Soil salinization is a key factor hindering sustainable agricultural development. Remote sensing technology, with its advantages of high timeliness, wide coverage, and low cost, has become the main technical means for obtaining dynamic information on soil salinity over large areas. Currently, research on soil salinity retrieval based on remote sensing mainly focuses on optical remote sensing and microwave remote sensing. Optical remote sensing is mostly limited to establishing empirical models based on the correlation between spectral reflectance and measured soil salinity, which suffers from insufficient mechanistic rationality, poor model universality, and is easily constrained by meteorological conditions. Active microwave remote sensing is limited by the complex relationship between soil salinity and backscattering coefficient, especially the interference of factors such as vegetation and land surface roughness on the backscattering coefficient, which limits the accuracy of soil salinity retrieval. Passive microwave remote sensing has low spatial resolution, making it difficult to meet the requirements for higher spatial resolution soil salinity retrieval accuracy.

[0003] In recent years, with the rapid development of Global Navigation Satellite System Reflectometry (GNSS-R), it has become possible to use GNSS-R technology to continuously and accurately monitor soil salinity over a wide area.

[0004] Extracting high signal-to-noise ratio land surface reflectance is a prerequisite for GNSS-R soil salinity inversion. However, due to the complexity of the land surface environment, the reflected signal is easily modulated by vegetation cover and land surface roughness during propagation and scattering. This modulation not only masks the soil salinity-related features in the reflectance but also significantly reduces its physical sensitivity to salinity changes.

[0005] To address the aforementioned electromagnetic losses and scattering interference, existing technologies primarily employ empirical parameters and semi-empirical models for correction. However, the following technical bottlenecks still exist in complex land surface environments:

[0006] First, there are the limitations of empirical parameter correction in terms of "physical representation" and "spatiotemporal dynamic mismatch." Existing schemes heavily rely on external prior parameters such as vegetation optical thickness and surface roughness. In the dimension of vegetation extinction, existing models are limited by a single scalar value of water content, neglecting the modulation effect of vegetation volume density and asymmetric geometry on beam polarization, resulting in significant deviations in coherent energy reduction under complex vegetation cover. In the dimension of roughness characteristics, parameters are mostly derived from low-frequency offline lookup tables or coarse-scale remote sensing products, making it difficult to capture in real time the micro-topographical evolution caused by transient factors such as cultivation and rainfall. This logical mismatch between "static priors" and "highly dynamic evolution" leads to a severe disconnect between the correction operator and the real-time reflection signal, with residual incoherent clutter components masking the intrinsic electrical characteristics of reflectivity.

[0007] Second, the semi-empirical model correction suffers from "oversimplification of physical mechanisms" and "salt response passivation." Existing schemes are mostly based on the first-order radiative transfer approximation theory, treating passive microwave brightness temperature as a single linear correction factor and viewing vegetation extinction and surface scattering as algebraically additive, isolated processes. However, in real electromagnetic propagation, electromagnetic waves travel along a complex "penetration-reflection-repenetration" path, and there is a strong second-order nonlinear interference loss between vegetation extinction and random surface roughness. This oversimplified physical description prevents the model from accurately identifying and removing incoherent clutter in complex spatial scattering fields. This directly leads to a significant passivation of the physical sensitivity of the corrected reflectivity to the imaginary part of the soil's complex permittivity (salt characteristics), severely limiting the robustness and universality of the inversion algorithm in complex, high-salt scenarios. Summary of the Invention

[0008] To address the problems in the background technology, this invention proposes a soil salinity inversion method based on polarization remediation and active-passive offsetting, including: S1, simultaneously acquiring the initial reflectance obtained from observations by a spaceborne GNSS-R payload. IR Microwave radiometer vertical polarization brightness temperature T B_V and horizontal polarization brightness temperature T B_H ;based on T B_V and T B_H Construct a normalized polarization difference rate that reflects the anisotropy of surface electromagnetic radiation MPD And the spatial gradient factor reflecting the spatial heterogeneity of the Earth's surface ▽ T B ; Utilizing the MPD and ▽ T B Nonlinear coupling was performed to construct polarization driving factors characterizing the composite extinction capability of the land surface. c S2, based on the polarization driving factor c With the angle of incidence observed iConstruct a dynamically coupled coherent compensation operator with active and passive physical component feedback mechanism F loss Using the operator F loss Without requiring external a priori data on surface roughness and vegetation optical thickness, the polarization driving factor... c Real-time offsetting of the second-order nonlinear interference loss between beam depolarization induced by vegetation multipath scattering and random surface roughness, affecting the initial reflectivity. IR Energy compensation and repair were performed, and the corrected reflectivity was obtained through analysis. IR cor S3, Based on the Fresnel reflection principle, a circular polarization state conversion mapping relationship is established, and an iterative step-based convergence discrimination mechanism is introduced to correct the reflectivity. IR cor Converted to the total dielectric constant modulus characterizing the overall polarization intensity of the soil medium e S4, based on external soil moisture data m υ As a physical boundary constraint, the real part of the dielectric constant driven by the polarization of water dipoles is analyzed based on the dielectric model of the hybrid dielectric. e' Using the orthogonal decomposition logic of the complex permittivity vector space, from the total permittivity modulus e Middle stripping solid portion e' Extract the imaginary part of the dielectric constant to characterize the conductivity loss in the ionic state. eh" S5, Based on the electrochemical physics mapping model, the imaginary part of the extracted dielectric constant is... eh" Converted to soil salinity S This invention achieves integrated and coordinated restoration of vegetation extinction and surface roughness interference by constructing a polarization gradient-driven second-order nonlinear compensation operator. Combined with complex dielectric vector orthogonal stripping logic, it improves the physical sensitivity and accuracy of soil salinity inversion under complex environments.

[0009] This invention also proposes a soil salinity inversion system based on polarization remediation and active / passive offsetting, comprising:

[0010] Feature space construction module: configured to synchronously acquire the initial reflectivity obtained from observations by the spaceborne GNSS-R payload. IR Microwave radiometer vertical polarization brightness temperature T B_V and horizontal polarization brightness temperature T B_H ;based on T B_V and T B_H Construct a normalized polarization difference rate that reflects the anisotropy of surface electromagnetic radiation MPDAnd the spatial gradient factor reflecting the spatial heterogeneity of the Earth's surface ▽ T B ; Utilizing the MPD and ▽ T B Nonlinear coupling was performed to construct polarization driving factors characterizing the composite extinction capability of the land surface. c ;

[0011] Dynamic coupling compensation module: configured to be based on the polarization driving factor c With the angle of incidence observed i Construct a dynamically coupled coherent compensation operator with active and passive physical component feedback mechanism F loss Using the operator F loss Without requiring external a priori data on surface roughness and vegetation optical thickness, the polarization driving factor... c Real-time offsetting of the second-order nonlinear interference loss between beam depolarization induced by vegetation multipath scattering and random surface roughness, affecting the initial reflectivity. IR Energy compensation and repair were performed, and the corrected reflectivity was obtained through analysis. IR cor ;

[0012] Electromagnetic mapping inversion module: configured to establish a circular polarization state transformation mapping relationship based on the Fresnel reflection principle, and introduce a convergence discrimination mechanism based on iterative steps to adjust the corrected reflectivity. IR cor Converted to the total dielectric constant modulus characterizing the overall polarization intensity of the soil medium e ;

[0013] Physics field decoupling module: configured with external soil moisture data m υ As a physical boundary constraint, the real part of the dielectric constant driven by the polarization of water dipoles is analyzed based on the dielectric model of the hybrid dielectric. e' Using the orthogonal decomposition logic of the complex permittivity vector space, from the total permittivity modulus e Middle stripping solid portion e' Extract the imaginary part of the dielectric constant to characterize the conductivity loss in the ionic state. eh" ;

[0014] Salt content quantitative inversion module: Configured with an electrochemical physics mapping model, it extracts the imaginary part of the dielectric constant. eh" Converted to soil salinity S .

[0015] The beneficial effects of this invention are as follows: This invention utilizes a polarization gradient-driven second-order nonlinear compensation operator to overcome the mechanistic limitations of traditional schemes based on first-order physical approximations, achieving integrated and collaborative restoration of vegetation structure and transient micro-topographical disturbances on the land surface. This method significantly improves the signal-to-noise ratio and coherence purity of the extracted reflectance; combined with complex dielectric vector orthogonal stripping logic, it effectively suppresses the mathematical ill-conditioning and physical ambiguity of salinity inversion under complex environments, greatly enhancing the physical sensitivity of reflectance to changes in the imaginary part of dielectric material driven by salinity. This invention provides robust mechanistic support and algorithmic assurance for high-precision dynamic monitoring of soil salinity in large-scale, complex land surface environments using spaceborne GNSS-R. Attached Figure Description

[0016] To facilitate understanding of the invention, it will be described in more detail with reference to the specific embodiments shown in the accompanying drawings. These drawings depict only typical embodiments of the invention and should not be considered as limiting the scope of protection of the invention.

[0017] Figure 1 This is a flowchart of the method of the present invention.

[0018] Figure 2 This is a flowchart of the system of the present invention.

[0019] Figure 3 The graphs show the correlation between initial and corrected reflectance and moisture content.

[0020] Figure 4 This is a scatter plot showing the relationship between soil salinity inversion values ​​and measured values.

[0021] Figure 5 This is a comparison chart of the accuracy of soil salinity inversion between the method of this invention and the traditional method.

[0022] Figure 6 This is a monthly spatial distribution map of soil salinity in the test area. Detailed Implementation

[0023] The embodiments of the present invention are described below with reference to the accompanying drawings to enable those skilled in the art to better understand and implement the present invention. However, the listed embodiments are not intended to limit the present invention. In the absence of conflict, the following embodiments and the technical features in the embodiments can be combined with each other, wherein the same components are indicated by the same reference numerals.

[0024] This invention provides a soil salinity retrieval method based on polarization coherence restoration and active / passive component offsetting through deep data fusion of spaceborne GNSS-R observation data (such as CYGNSS and Fengyun-3 series satellites) and passive microwave data (such as SMAP and SMOS satellite data). This embodiment first uses bistatic radar equations to physically restore the raw power of the spaceborne GNSS-R data to obtain the initial reflectivity; simultaneously, it utilizes passive brightness temperature to construct a normalized polarization difference rate. MPD With spatial gradient factor ▽ T B ,pass MPD With ▽ T B Nonlinear coupling was used to construct a surrogate parameter for the composite extinction capacity of the land surface—the polarization driving factor. c; Subsequently, the polarization driving factor was utilized. c Construct a dynamically coupled coherent compensation operator with the observed incident angle F loss This is used to offset the second-order nonlinear interference loss between beam depolarization induced by vegetation multipath scattering and random surface roughness in real time. This is achieved through the dynamically coupled coherent compensation operator. F loss Coherence gain compensation was applied to the initial reflectivity to achieve a high signal-to-noise ratio coherent reflectivity restoration that closely approximates the physical properties of bare soil. Building upon this, this embodiment further incorporates a real-imaginary part separation technique for complex dielectric properties. Utilizing the specific sensitivity of salt to the imaginary part of the dielectric constant, it eliminates the interference of soil moisture on the real part of the dielectric while achieving accurate inversion of soil salinity information. This embodiment significantly improves the accuracy and universality of soil salinity inversion under complex and heterogeneous surface conditions through physical decoupling and multi-source data offsetting.

[0025] Specifically, such as Figure 1 As shown, the method of the present invention includes:

[0026] S1, synchronously acquire the initial reflectivity obtained from observations by the spaceborne GNSS-R payload. IR Microwave radiometer vertical polarization brightness temperature T B_V and horizontal polarization brightness temperature T B_H ;based on T B_V and T B_H Construct a normalized polarization difference rate that reflects the anisotropy of surface electromagnetic radiation MPD And the spatial gradient factor reflecting the spatial heterogeneity of the Earth's surface ▽ T B ; Utilizing the MPD and ▽ T BNonlinear coupling was performed to construct polarization driving factors characterizing the composite extinction capability of the land surface. c .

[0027] This embodiment first acquires the delayed Doppler image (DDM) generated by the spaceborne GNSS-R. By extracting the latitude and longitude of the specular reflection point and the observation time from the DDM metadata, it retrieves and matches corresponding passive microwave radiometer data (such as SMAP and SMOS satellite data) within a preset spatial grid and time window. The spatial grid uses the global standard EASE-Grid 2.0 projection grid, with a resolution set to 9km to match the standard pixel size of the microwave radiometer. Preferably, to ensure physical consistency, a spatial index lookup table is constructed to map the coordinates of the GNSS-R specular reflection point to the grid index and lock the center pixel. A bilinear interpolation algorithm is used for spatial resampling to address sampling offsets, ensuring that each observation point obtains a physically consistent equivalent brightness temperature value. T B_V and T B_H .

[0028] In one implementation, to address the geometric attenuation of GNSS-R signals during propagation, this embodiment employs bistatic radar equations to physically reconstruct the original DDM power. A search algorithm is used to pinpoint the peak power in the DDM spectrum. This algorithm first identifies the maximum power grid point through a global traversal to locate the discrete maximum coordinates. Then, a 3×3 pixel local feature window is defined centered on this point, and a two-dimensional quadratic surface fitting algorithm is used to solve for the extreme points within this region where the partial derivatives are zero. This eliminates the quantization error caused by discrete sampling and extracts the true physical peak power at the sub-pixel level. Simultaneously, a background noise power benchmark is introduced. N The benchmark N The power mean of the DDM (Digital Delayed Mirror) is obtained by sampling the non-signal-correlated regions (such as background grid points far from the center of the time-delayed Doppler). The initial reflectivity is calculated using the following formula. IR :

[0029] (1),

[0030] In the formula, P peak For DDM peak power, P t For transmission power, G t For the transmit antenna gain, G r For receiving antenna gain, R t The bistatic distance between the transmitter and the specular reflection point. R rThe bistatic distance between the specular reflection point and the receiver. l The carrier wavelength is used. Through the above compensation, the influence of satellite trajectory and geometric gain variations on reflected energy is eliminated, thus improving the initial reflectivity. IR It can accurately characterize the scattering properties of the land surface.

[0031] This embodiment further utilizes the aligned vertical polarization brightness temperature. T B_V and horizontal polarization brightness temperature T B_H Construct a normalized polarization difference rate that reflects the anisotropy of surface radiation. MPD The calculation formula is as follows:

[0032] (2),

[0033] MPD The value is limited to [0, 1], directly reflecting the ratio of coherent to incoherent scattering within the observed footprint. This parameter eliminates the absolute value interference of surface physical temperature fluctuations through normalization operations, thus purifying the polarization characteristics of electromagnetic radiation.

[0034] Furthermore, to capture the non-uniform scattering characteristics within the observed footprint, this embodiment constructs a 3×3 neighborhood pixel window (covering an area of ​​approximately 27km×27km) using the aforementioned locked center pixel as the core. A spatial second-order difference operation is performed on the brightness temperature matrix to extract the spatial gradient factor ▽, which reflects the degree of drastic change in land cover. T B :

[0035] (3),

[0036] The partial derivative term is implemented using the central difference operator, whose calculation step size is the grid spacing (9 km). This factor is used to quantify transient energy fluctuations caused by topographic relief, land-water boundaries, or vegetation patch distribution, characterizing the micro-topographic heterogeneity within the first Fresnel zone.

[0037] Furthermore, this embodiment will MPD With ▽ T B Multidimensional nonlinear coupling is performed to generate polarization driving factors, which characterize the composite extinction surrogate parameters of the land surface. c The calculation formula is as follows:

[0038] (4),

[0039] Where norm(·) is a normalization function based on the global brightness temperature extremes; oh 1 and oh 2 For dynamic weighting coefficients ( oh 1 + oh 2 =1).

[0040] The weighting coefficients are adjusted in real time based on the preset land use classification (LUCC). In vegetated areas, the weighting coefficients are increased. oh 1 The weighting is increased to enhance the response to vegetation extinction loss; in areas with complex micro-topography, the weighting is increased. oh 2 The weights are used to accurately capture incoherent scattering caused by roughness.

[0041] Through the refined S1 steps described above, this invention achieves precise physical matching of active and passive observation data within a 9km standard spatial grid. This scheme achieves this through real-time capture. MPD The polarization convergence phenomenon and ▽ T B The spatial energy fluctuations were dynamically quantified, and the "extinction loss environment" within the observation footprint point was determined, providing a basis for the use of the dynamically coupled coherent compensation operator in step S2. F loss Achieving second-order nonlinear hedging provides a high-confidence physical driving input.

[0042] S2, based on the polarization driving factor c With the angle of incidence observed i Construct a dynamically coupled coherent compensation operator with active and passive physical component feedback mechanism F loss Using the operator F loss Without requiring external a priori data on surface roughness and vegetation optical thickness, the polarization driving factor... c Real-time offsetting of the second-order nonlinear interference loss between beam depolarization induced by vegetation multipath scattering and random surface roughness, affecting the initial reflectivity. IR Energy compensation and repair were performed, and the corrected reflectivity was obtained through analysis. IR cor .

[0043] The core of step S2 in this embodiment lies in constructing a dynamic physical compensation mechanism capable of providing real-time feedback on the complex heterogeneous loss characteristics of the land surface. This step achieves energy repair through the deep coupling of active and passive polarization characteristics without requiring prior data on external surface roughness and vegetation optical thickness.

[0044] First, using the driving factors obtained in step S1 c With the angle of incidence observed iA second-order nonlinear coupling mapping with active and passive physical component feedback mechanism is established, and a dynamic coupling coherent compensation operator is constructed. F loss The specific calculation formula is as follows:

[0045] (5),

[0046] In the formula, i To observe the angle of incidence; α It is a second-order interference modulation index used to adjust the coupling weight between the vegetation layer and the surface roughness. Its value is usually set between 0.5 and 3.0, and it is calibrated according to the electromagnetic wave penetration characteristics under a specific frequency band (such as L-band).

[0047] The physical essence of this operator lies in its ability to capture and quantify the energy extinction and phase degradation of reflected signals in the "penetration-reflection-repeating" path in real time by utilizing the polarization state sensitivity of passive microwave radiation to surface roughness and vegetation density. Through exponential mapping, this operator can generate extremely strong nonlinear fitting capabilities, effectively characterizing the second-order scattering and interference effects of electromagnetic waves in complex heterogeneous underlying surfaces.

[0048] Furthermore, in constructing the dynamically coupled coherent compensation operator... F loss Subsequently, this embodiment, based on physical hedging logic, utilizes this operator to adjust the initial reflectivity in step S1. IR Inverse energy gain restoration is performed. By offsetting the observed reflectivity with the compensation operator on a physical scale, incoherent noise components are removed and the coherent characteristics of the reflected signal are restored, resulting in the analytically corrected reflectivity. IR cor The calculation formula is as follows:

[0049] (6),

[0050] In this process, the operator F loss Through driving factors c Real-time capture of depolarization physical characteristics within footprint points enables "feedback" offset compensation for coherent energy loss caused by the coupling of vegetation multipath scattering and random surface roughness.

[0051] Compared with traditional first-order physical approximations (such as...) t-o The model may be linearly simplified, or the correction scheme may depend on prior parameters. This embodiment uses... F lossThe operator enables real-time mechanistic adaptation between the correction model and the reflected signal under highly dynamic environments. This physically constrained dynamic hedging logic not only synergistically repairs the nonlinear interference between vegetation cover loss and incoherent surface scattering, but also improves the corrected reflectivity. IR cor By forcibly regressing the statistical properties to the coherent reflection benchmark defined by Fresnel's law of reflection, the intrinsic sensitivity of reflectivity to the imaginary part of the soil complex permittivity (the ion conductivity loss characteristic controlled by salinity) is greatly enhanced.

[0052] S3, a circular polarization state conversion mapping relationship is established based on the Fresnel reflection principle, and a convergence discrimination mechanism based on iterative steps is introduced to correct the reflectivity. IR cor Converted to the total dielectric constant modulus characterizing the overall polarization intensity of the soil medium e .

[0053] In this embodiment, since the satellite-borne GNSS-R transmitted signal (such as the GPS L1 band) is right-hand circularly polarized (RHCP), but after reflection from the land surface, it mainly exhibits left-hand circularly polarized (LHCP), this embodiment establishes a circular polarization conversion mapping based on the Fresnel reflection principle, and its corrected reflectivity... IR cor The polarization difference relationship between the vertically polarized reflection coefficient and the horizontally polarized reflection coefficient must be satisfied:

[0054] (7),

[0055] In the formula, C H and C V Let F represent the surface reflectivity of horizontal and vertical polarization, respectively, and their components follow the classical Fresnel formula:

[0056] (8),

[0057] (9),

[0058] By substituting formulas (8) and (9) into the polarization state conversion formula (7), this embodiment further derives the corrected reflectivity. IR cor With the total dielectric constant modulus e Explicit analytical equation:

[0059] (10)

[0060] Using this higher-order nonlinear analytical equation, combined with the results obtained in step S2 IR corWith the angle of incidence observed i The dielectric modulus, which characterizes the integrated polarization of soil water and salt, was analyzed in real time using a numerical inverse algorithm. e .

[0061] To address the potential nonlinear singularities in formula (10) when processing extreme reflectivity data, this embodiment introduces a convergence discrimination mechanism during the solution process. This mechanism, through... e The physical value range is dynamically limited to ensure that it meets the constraints. e min ≤ e ≤ e max ;in, e min It is usually set at 2.0-3.0 (dry soil baseline). e max The range is set to 40.0-80.0 (high-concentration salinization baseline). If the calculated result exceeds this range, the system automatically initiates an iterative optimization program to fine-tune the incident angle. i The characteristic deviation guides the numerical solution to converge toward the physically reasonable interval.

[0062] The total dielectric constant modulus obtained by analysis e This comprehensively reflects the polarization response intensity of the soil medium under electromagnetic field excitation. Its modulus is jointly determined by the dipole polarization induced by soil moisture (real part) and the ion conduction loss induced by soil salinity (imaginary part). This purification step eliminates the interference of surface geometric heterogeneity, providing pure physical parameters for the subsequent "water-salt" decoupling using active-passive component counterbalancing logic in step S4.

[0063] S4, based on external soil moisture data m υ As a physical boundary constraint, the real part of the dielectric constant driven by the polarization of water dipoles is analyzed based on the dielectric model of the hybrid dielectric. e' Using the orthogonal decomposition logic of the complex permittivity vector space, from the total permittivity modulus e Middle stripping solid portion e' Extract the imaginary part of the dielectric constant to characterize the conductivity loss in the ionic state. eh" .

[0064] In this embodiment, step S4 introduces external multi-source prior information to solve the signal ambiguity problem caused by the synergistic contribution of soil water and salt. Soil moisture data. m υThe data is derived from publicly available soil moisture remote sensing products, including but not limited to spaceborne passive microwave radiometer inversion products (such as SMAP L3 / L4 level moisture data and SMOS moisture products) and active radar scattering coefficient inversion products. This data is resampled using temporal window matching and spatial EASE-Grid to provide a physical reference for decoupled calculations of the reflection footprint points.

[0065] Surface temperature T Publicly available land surface temperature products acquired simultaneously, including but not limited to land surface temperature products from the Moderate Resolution Imaging Spectroradiometer (MODIS) and infrared temperature products integrated from multiple satellite searches (such as global land surface temperature grid data), are used to drive temperature dependency corrections in dielectric models.

[0066] Soil texture data includes soil bulk density r b Sand content G With clay content C The data are derived from external global soil prior databases, including but not limited to Soil Grids, HWSD, or regional soil texture datasets.

[0067] Based on the acquired moisture data m υ In this embodiment, the real part of the dielectric constant driven by moisture is analyzed using the Dobson-S hybrid dielectric model. e' The calculation formula is as follows:

[0068] (11),

[0069] In the formula, α The shape factor, which characterizes the geometric arrangement and pore structure of soil particles, ranges from 0.5 to 0.8, preferably 0.65. e s is the dielectric constant of the soil solid phase material, usually taken as 4.7; r s This refers to soil density, typically taken as 2.65 g / cm³. (Parameter) β′ This is used to correct for the differential contribution of moisture to polarization under different textures, based on sand content. G With clay content C Dynamic calculation: .

[0070] Based on Debye relaxation theory, using surface temperature T For the dielectric real part of free water e fω ′ Dynamic correction is performed to eliminate the interference of ambient temperature fluctuations on dielectric properties:

[0071] (12)

[0072] (13)

[0073] (14)

[0074] In the formula, f The carrier operating frequency of the spaceborne GNSS-R satellite (e.g., 1.575 GHz). e ω∞ It is the high-frequency limiting dielectric constant, usually set to 4.9.

[0075] Obtaining the real part of the dielectric constant e' Subsequently, this embodiment utilizes the orthogonal decomposition logic of the complex permittivity vector space to process the total permittivity modulus obtained in step S3. e Phase space stripping is performed. Based on the orthogonal characteristics of moisture (displacement current) and salt (conduction current) in the soil medium in terms of physical response mechanisms, the imaginary part of the dielectric constant, which characterizes the conductivity loss of the ionic state, is extracted using the following vector stripping formula. e '':

[0076] (15)

[0077] Through the orthogonal decomposition based on existing publicly available moisture products, this embodiment successfully eliminated signal artifacts caused by soil moisture fluctuations. The resulting imaginary part of the dielectric constant... e "The interference of the 'water-salt co-effect' phenomenon is eliminated at the physical source, providing a pure feature input for the subsequent step S5 to perform high-confidence salt quantification inversion."

[0078] S5, based on the electrochemical physics mapping model, the imaginary part of the extracted dielectric constant is... eh" Converted to soil salinity S .

[0079] This embodiment establishes the imaginary part of the dielectric constant. e The physical correlation between salinity and soil solution conductivity enables quantitative characterization of salinity. The soil salinity content... S The specific calculation formula is as follows:

[0080] (16)

[0081] In the formula, A It depends on the types of salt ions contained in the soil (such as...) NaCl、Na 2 SO 4 etc The structural constants of ) are typically defined between 0.7 and 1.0; g For S-based togrin The first-order fitting coefficient of soil solution conductivity and salinity in the model is usually taken as 0.14. e 0 This is the free-space conductivity, typically 8.854 × 10⁻⁶. -12 F / m; x This is the temperature compensation coefficient; β″ To correct for the differential contribution of moisture to polarization under different textures, it is based on sand content. G With clay content C Dynamic calculation: The physical meaning and source of the other parameters are the same as those described in formula (11).

[0082] The temperature compensation coefficient χ is used to correct the migration rate deviation caused by the thermal motion of ions, and the calculation formula is as follows:

[0083] (17)

[0084] In the formula T The surface temperature (degrees Celsius) is obtained from the same source as in step S4.

[0085] Through the combined operation of formulas (16) and (17), this embodiment utilizes the pure imaginary part extracted in step S4. e'' As the core input, by deeply offsetting and stripping away the mechanisms of "water dipole polarization" and "salt ion conduction," the nonlinear interference of water content fluctuations on the salinity inversion results is completely eliminated. The final soil salinity output is... S It can accurately characterize the true degree of salinization within the footprint point, fundamentally solving the low sensitivity problem caused by "water and salt co-effect" in traditional GNSS-R salinity inversion.

[0086] Based on the unified inventive concept, this application also provides a system corresponding to the soil salinity inversion method based on polarization coherent repair and active-passive component offsetting. Since the problem-solving principle of the system in this application is similar to the above-mentioned method in this application, the implementation of the system can participate in the implementation of the method.

[0087] like Figure 2 The diagram shown is a schematic of a soil salinity inversion system based on polarization coherent remediation and active / passive component offsetting, including: a feature space construction module 201, a dynamic coupling compensation module 202, an electromagnetic mapping inversion module 203, a physical field decoupling module 204, and a salinity quantitative inversion module 205. Among them:

[0088] Feature space construction module 201 is configured to synchronously acquire the initial reflectivity obtained from observations by the spaceborne GNSS-R payload. IRMicrowave radiometer vertical polarization brightness temperature T B_V and horizontal polarization brightness temperature T B_H ;based on T B_V and T B_H Construct a normalized polarization difference rate that reflects the anisotropy of surface electromagnetic radiation MPD And the spatial gradient factor reflecting the spatial heterogeneity of the Earth's surface ▽ T B ; Utilizing the MPD and ▽ T B Nonlinear coupling was performed to construct polarization driving factors characterizing the composite extinction capability of the land surface. c ;

[0089] The dynamic coupling compensation module 202 is configured to be based on the polarization driving factor. c With the angle of incidence observed i Construct a dynamically coupled coherent compensation operator with active and passive physical component feedback mechanism F loss Using the operator F loss Without requiring external a priori data on surface roughness and vegetation optical thickness, the polarization driving factor... c Real-time offsetting of the second-order nonlinear interference loss between beam depolarization induced by vegetation multipath scattering and random surface roughness, affecting the initial reflectivity. IR Energy compensation and repair were performed, and the corrected reflectivity was obtained through analysis. IR cor ;

[0090] The electromagnetic mapping inversion module 203 is configured to establish a circular polarization state transformation mapping relationship based on the Fresnel reflection principle, and introduces a convergence discrimination mechanism based on iterative steps to adjust the corrected reflectivity. IR cor Converted to the total dielectric constant modulus characterizing the overall polarization intensity of the soil medium e ;

[0091] The physics field decoupling module 204 is configured to use external soil moisture data. m υ As a physical boundary constraint, the real part of the dielectric constant driven by the polarization of water dipoles is analyzed based on the dielectric model of the hybrid dielectric. e' Using the orthogonal decomposition logic of the complex permittivity vector space, from the total permittivity modulus e Middle stripping solid portion e' Extract the imaginary part of the dielectric constant to characterize the conductivity loss in the ionic state. e″ ;

[0092] The salt content quantitative inversion module 205 is configured to use an electrochemical physical mapping model to extract the imaginary part of the dielectric constant. eh" Converted to soil salinity S .

[0093] This embodiment conducted a full-process empirical study in a typical salinized area. Through the following specific execution steps, the superior performance of the soil salinity inversion model (PRD) driven by polarization coherent remediation and active-passive component offsetting was verified.

[0094] Step 1: Acquisition and preprocessing of multi-source data.

[0095] This embodiment acquired the CYGNSS L1 dataset from January to December 2023 through the NASA Earth Data platform, totaling 2548 data files, and analyzed them based on signal-to-noise ratio (SNR>0), incident angle (…). i (<65°) and quality labels underwent rigorous screening to ensure the reliability of the raw signals. Auxiliary parameters were primarily derived from the concurrent SMAP enhanced L3 radiometer soil moisture product, with a total of 3650 data files downloaded, providing soil moisture content. m υ and surface temperature T Physical benchmarks were used. In addition, the 2018 version of the basic attributes data from the China High-Resolution National Soil Information Grid was introduced. This dataset has a high spatial resolution of 250 m, providing the model with accurate soil bulk density data. r b Sand content G With clay content C Parameters such as these.

[0096] Step 2: Surface reflectance extraction and dynamic compensation repair effect verification.

[0097] This embodiment extracts the initial reflectance based on delayed Doppler images (DDM) observed by CYGNSS satellites. IR And call the dynamically coupled coherent compensation operator constructed in step S2. F loss Energy compensation remediation was performed. Correlation analysis was conducted between the reflectance before and after remediation and publicly available soil moisture products (SMAP). The experimental results showed (see...) Figure 3 The correlation coefficient between the corrected reflectance and moisture content increased significantly from 0.69 to 0.84, resulting in a signal effectiveness gain of over 21%. This result fully demonstrates that the present invention, through dynamic coupling coherent compensation operators... F loss It effectively removes incoherent scattering noise induced by extinction of terrestrial vegetation and random roughness, making the reflectivity more intrinsically map the electrical characteristics of the soil medium.

[0098] Step 3: Performance of water-salt decoupling inversion and its comparative verification.

[0099] This embodiment uses reflectivity correction. IR cor As the driving source, analyze the total dielectric constant modulus. e The imaginary part of the dielectric is extracted by combining the Dobson-S model with the orthogonal decomposition logic of the vector space. eh" To achieve soil salinity S The inversion was performed. The results of the RPD model inversion were compared with synchronous surface measurement data (see...). Figure 4 The results show its coefficient of determination. R 2 The root mean square error reached 0.83. RMSE As low as 0.53 g / kg. This proves that the model ensures the physical purity of the inversion input through physical hedging, enabling it to still have strong physical generalization ability without relying on a large number of experimental prior samples.

[0100] Furthermore, the PRD model of this invention is compared with traditional empirical models (ERD) and semi-empirical models (SERD). The comparative analysis shows (see...) Figure 5 The ERD model, relying entirely on prior data such as vegetation optical thickness and surface roughness coefficient, struggles to quantify the highly dynamic transient characteristics of land surface physical degradation, resulting in the weakest performance. R 2 Only 0.64, RMSE The result reached 0.82 g / kg. Although the SERD model incorporates some physical parameters, its accuracy improvement is limited due to its failure to effectively decouple the beam depolarization effect. R 2 It is 0.69. RMSE It is 0.72 g / kg.

[0101] In contrast, the PRD model of this invention uses dynamically coupled coherent compensation operators. F loss It achieves physical restoration of reflectivity, its R 2 It improved performance by approximately 29.7% compared to the ERD model and by approximately 20.3% compared to the SERD model; meanwhile RMSE Compared to ERD, it reduced noise by approximately 35.4%, and compared to SERD, it reduced noise by approximately 26.4%. The experimental results fully demonstrate that the present invention, by removing incoherent noise and performing vector orthogonal decomposition, successfully overcomes the bottleneck of the traditional model's dependence on external static parameters, producing unexpected technical effects.

[0102] Step 4: Mapping the spatial distribution of soil salinity and analyzing its spatiotemporal evolution.

[0103] This embodiment uses corrected reflectance data for the entire year of 2023 as input. IR cor A monthly spatial distribution map of soil salinity in the test area was generated. Figure 6 ).from Figure 6 As can be seen, the test area exhibits typical climate-driven seasonal variation characteristics: soil salinity peaks in May and reaches its lowest point in July. Meanwhile, high-salinity areas are spatially stable along the coastal zone and inland low-lying areas, demonstrating significant structural and fixed characteristics. The method of this invention can capture dynamic changes on the land surface with high sensitivity, accurately characterizing the spatial heterogeneity and temporal evolution trend of soil salinity, and verifying the robustness and temporal fidelity of the model under complex underlying surfaces.

[0104] In summary, this invention, through rigorous physical logic, integrates the dynamically coupled coherent compensation operator. F loss As the "golden key" to solving the interference of vegetation and roughness, it ensures the physical fidelity of reflectivity from the source. Combined with the orthogonal decomposition technology of complex permittivity vector space, it achieves deep decoupling of water and salt effects. It not only successfully drives the accuracy of subsequent inversion processes, but also provides a leading physical mechanism-driven solution for large-scale dynamic monitoring of soil salinization with its advantages of zero prior dependence, high robustness and strong generalization ability.

[0105] The embodiments described above are merely preferred embodiments of the present invention. The terms "in one embodiment," "in another embodiment," "in yet another embodiment," or "in still another embodiment" used in this specification all refer to one or more of the same or different embodiments according to this disclosure. Ordinary variations and substitutions made by those skilled in the art within the scope of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for soil salinity inversion based on polarization remediation and active / passive offsetting, characterized in that, include: S1, synchronously acquire the initial reflectivity obtained from observations by the spaceborne GNSS-R payload. IR Microwave radiometer vertical polarization brightness temperature T B_V and horizontal polarization brightness temperature T B_H Based on vertical polarization brightness temperature T B_V and horizontal polarization brightness temperature T B_H Construct a normalized polarization difference rate that reflects the anisotropy of surface electromagnetic radiation MPD And the spatial gradient factor reflecting the spatial heterogeneity of the Earth's surface ▽ T B ; Using the normalized polarization difference rate MPD With spatial gradient factor ▽ T B Nonlinear coupling was performed to construct polarization driving factors characterizing the composite extinction capability of the land surface. γ ; S2, based on the polarization driving factor γ With the angle of incidence observed θ Construct a dynamically coupled coherent compensation operator with active and passive physical component feedback mechanism F loss ; Using the dynamic coupling coherent compensation operator F loss Without requiring external a priori data on surface roughness and vegetation optical thickness, the polarization driving factor... γ Real-time offsetting of the second-order nonlinear interference loss between beam depolarization induced by vegetation multipath scattering and random surface roughness, affecting the initial reflectivity. IR Energy compensation and repair were performed, and the corrected reflectivity was obtained through analysis. IR cor ; S3, a circular polarization state conversion mapping relationship is established based on the Fresnel reflection principle, and a convergence discrimination mechanism based on iterative steps is introduced to correct the reflectivity. IR cor Converted to the total dielectric constant modulus characterizing the overall polarization intensity of the soil medium ε ; S4, based on external soil moisture data m υ As a physical boundary constraint, the real part of the dielectric constant driven by the polarization of water dipoles is analyzed based on the dielectric model of the hybrid dielectric. ε′ Using the orthogonal decomposition logic of the complex permittivity vector space, from the total permittivity modulus ε Middle stripping solid portion ε′ Extract the imaginary part of the dielectric constant to characterize the conductivity loss in the ionic state. ε″ ; S5, based on the electrochemical physics mapping model, the imaginary part of the extracted dielectric constant is... ε″ Converted to soil salinity S .

2. The soil salinity inversion method based on polarization remediation and active / passive offsetting as described in claim 1, characterized in that, Step S1 includes: S11: Obtain the delayed Doppler image generated by the spaceborne GNSS-R payload, extract the peak power at its mirror reflection point, and perform antenna gain correction, spatial propagation loss compensation, and propagation distance normalization based on the bistatic radar equations to obtain the initial reflectivity characterizing the original surface reflection properties. IR ; S12, utilizing vertical polarization brightness temperature T B_V With horizontal polarization brightness temperature T B_H Perform standardized ratio calculations to construct normalized polarization difference rates. MPD =( T B_V - T B_H ) / ( T B_V + T B_H ); using normalized polarization difference rate MPD Extract electromagnetic anisotropy characteristics induced by surface vegetation extinction and physical roughness; S13, using spatial difference or gradient convolution operators, extracts the polarization-weighted brightness temperature spatial gradient factor between the target reflection pixel and its neighboring pixels. T B It is used to characterize the non-uniform heterogeneity of the ground space within the footprint of the reflection area; S14, through the dynamic composite mapping function γ = f ( MPD ,▽ T B Constructing a system based on the normalized polarization difference rate MPD For polarization sensitive terms and spatial gradient factors ▽ T B Polarization driving factor for spatial constraint terms γ The polarization driving factor γ As a core proxy parameter for simulating the composite extinction capacity of the earth's surface, it is used to establish a reflectivity restoration feedback mechanism without the need for external vegetation and prior roughness data.

3. The soil salinity inversion method based on polarization remediation and active / passive offsetting as described in claim 2, characterized in that, Step S11 includes: The initial reflectivity is calculated using the following formula. IR : ， In the formula, P peak For DDM peak power, N As a background noise power reference, P t For transmission power, G t For the transmit antenna gain, G r For receiving antenna gain, R t The bistatic distance between the transmitter and the specular reflection point. R r The bistatic distance between the specular reflection point and the receiver. λ The carrier wavelength.

4. The soil salinity inversion method based on polarization remediation and active / passive offsetting as described in claim 3, characterized in that, Step S13 includes: The spatial gradient factor ▽ is calculated using the following formula. T B : ， The partial derivative terms are implemented using the central difference operator, whose calculation step size is the grid spacing.

5. The soil salinity inversion method based on polarization remediation and active / passive offsetting as described in claim 1, characterized in that, Step S2 includes: S21, the polarization driving factor γ With the angle of incidence observed θ cosine component cosθ Establish a nonlinear coupling mapping relationship and construct a dynamic coupled coherent compensation operator for quantitatively characterizing the intensity of electromagnetic energy attenuation. F loss =exp[ Φ ( γ·cosθ In the formula, Φ A mapping operator to characterize the gain in coherent energy loss caused by land surface composite extinction; S22, based on physical hedging logic, utilizes the polarization driving factor γ The polarization-sensitive component offsets the beam depolarization loss induced by vegetation multipath scattering in real time, and its spatial constraint component compensates for the incoherent scattering component induced by random surface roughness; the initial reflectivity IR The dynamic coupling coherent compensation operator F loss Nonlinear analysis was performed to calculate the decoupled corrected reflectivity. IR cor = IR / F loss .

6. The soil salinity inversion method based on polarization remediation and active / passive offsetting as described in claim 1, characterized in that, Step S3 includes: Based on the Fresnel reflection principle, construct a corrected reflectivity IR cor The left- and right-hand circular polarization mapping relationship between the soil medium and the complex permittivity characterizes the soil medium as having a total permittivity modulus with comprehensive polarization intensity. ε The mapping relationship is expressed as follows: ， In the formula, θ The angle of incidence of the spaceborne GNSS-R signal at the mirror reflection point. Γ V and Γ H These are the Fresnel reflection coefficients under vertical and horizontal polarization, respectively. An iterative step-based convergence discrimination mechanism is introduced, utilizing a preset soil dielectric physical threshold range to determine the convergence. ε The value of is dynamically limited to lock the unique total dielectric constant modulus with physical meaning. ε .

7. The soil salinity inversion method based on polarization remediation and active / passive offsetting as described in claim 1, characterized in that, Step S4 includes: S41, based on external soil moisture data m υ As a physical boundary constraint, the real part of the dielectric constant driven by the polarization of water dipoles is analyzed based on the dielectric model of the hybrid dielectric. In the formula, α For shape factor, ρ s For soil density, ρ b For soil bulk density, ε s The dielectric constant of the soil solid phase material. β′ This is a correction factor for sand and clay content. ε fω ′ The dielectric real part of free water; S42, using the orthogonal decomposition logic of the complex permittivity vector space, the total permittivity modulus obtained in step S3 is... ε Middle stripping solid portion ε′ Extract the imaginary part of the dielectric constant to characterize the conductivity loss in the ionic state. By increasing the total electrical strength ε Component stripping is performed along the real and imaginary axes to ensure the extraction of the imaginary part of the dielectric constant. ε″ It is driven solely by the conductivity loss generated by ion migration.

8. The soil salinity inversion method based on polarization remediation and active / passive offsetting as described in claim 1, characterized in that, Step S5 includes: Based on the ion-state conductivity loss mechanism, the imaginary part of the soil dielectric constant extracted in step S4 is... ε″ The equivalent conductivity, which characterizes the electrical conductivity of the soil solution, is converted into an analytical value. This value is then combined with the carrier operating frequency, temperature compensation coefficient, and soil structure parameters to determine the target soil salinity. In the formula, f The carrier operating frequency is used to analyze the dispersion loss characteristics of electromagnetic waves in salt-containing media. χ This is the temperature compensation coefficient, used to correct the migration rate deviation caused by the thermal motion of ions; ζ These are the fitting coefficients. A This is a constant that depends on the types of salt ions contained in the soil. ε 0 The free vacuum permittivity is 1. α For shape factor, ρ s For soil density, ρ b For soil bulk density, β′′ This is a correction factor for sand and clay content. m υ This refers to soil moisture content.

9. A soil salinity inversion system based on polarization remediation and active / passive offsetting, characterized in that, include: Feature space construction module: configured to synchronously acquire the initial reflectivity obtained from observations by the spaceborne GNSS-R payload. IR Microwave radiometer vertical polarization brightness temperature T B_V and horizontal polarization brightness temperature T B_H ;based on T B_V and T B_H Construct a normalized polarization difference rate that reflects the anisotropy of surface electromagnetic radiation MPD And the spatial gradient factor reflecting the spatial heterogeneity of the Earth's surface ▽ T B ; Using the normalized polarization difference rate MPD With spatial gradient factor ▽ T B Nonlinear coupling was performed to construct polarization driving factors characterizing the composite extinction capability of the land surface. γ ; Dynamic coupling compensation module: configured to be based on the polarization driving factor γ With the angle of incidence observed θ Construct a dynamically coupled coherent compensation operator with active and passive physical component feedback mechanism F loss ; Using the dynamic coupling coherent compensation operator F loss Without requiring external a priori data on surface roughness and vegetation optical thickness, the polarization driving factor... γ Real-time offsetting of the second-order nonlinear interference loss between beam depolarization induced by vegetation multipath scattering and random surface roughness, affecting the initial reflectivity. IR Energy compensation and repair were performed, and the corrected reflectivity was obtained through analysis. IR cor ; Electromagnetic mapping inversion module: configured to establish a circular polarization state transformation mapping relationship based on the Fresnel reflection principle, and introduce a convergence discrimination mechanism based on iterative steps to adjust the corrected reflectivity. IR cor Converted to the total dielectric constant modulus characterizing the overall polarization intensity of the soil medium ε ; Physics field decoupling module: configured with external soil moisture data m υ As a physical boundary constraint, the real part of the dielectric constant driven by the polarization of water dipoles is analyzed based on the dielectric model of the hybrid dielectric. ε′ Using the orthogonal decomposition logic of the complex permittivity vector space, from the total permittivity modulus ε Middle stripping solid portion ε′ Extract the imaginary part of the dielectric constant to characterize the conductivity loss in the ionic state. ε″ ; Salt content quantitative inversion module: Configured with an electrochemical physics mapping model, it extracts the imaginary part of the dielectric constant. ε″ Converted to soil salinity S .

10. The soil salinity inversion system based on polarization remediation and active / passive offsetting as described in claim 9, characterized in that, The feature space construction module, dynamic coupling compensation module, electromagnetic mapping inversion module, physical field decoupling module, and salinity quantitative inversion module include a computer program that, when executed, implements the method as described in any one of claims 2-8.

Images