Active and passive microwave joint inversion method and system combined with Bayesian probability inversion

Abstract

The invention provides an active and passive microwave joint inversion method and system combined with Bayesian probability inversion, and the method comprises the steps: carrying out the sampling in a priori numerical range of a to-be-inverted parameter, obtaining an initial guess value, and carrying out the simulation based on the initial guess value, and obtaining active simulation data. And performing active inversion by combining the active simulation data and the active microwave observation data, and determining an effective roughness parameter required by passive inversion based on the roughness parameter obtained by the active inversion and a pre-constructed relation function. And performing passive inversion by combining the effective roughness parameter, the initial guess value of the soil moisture and the passive microwave observation data. The active inversion result and the passive inversion result respectively comprise an optimal estimation value and an uncertainty quantitative index of the to-be-inverted parameter. And obtaining a joint inversion result by combining the active inversion result and the passive inversion result. According to the scheme, active and passive microwave observation is combined to accurately estimate the soil moisture and quantify the uncertainty in the inversion algorithm, and the soil moisture estimation precision and reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of microwave remote sensing technology, and more specifically, to a method and system for joint active and passive microwave inversion that combines Bayesian probabilistic inversion. Background Technology

[0002] Soil moisture plays a crucial role in the terrestrial water cycle, and it is of paramount importance to research in hydrology, ecology, meteorology, and climate change. Active and passive microwave remote sensing technologies, with their unique penetrating power and all-weather observation capabilities unaffected by weather conditions, have become effective means of estimating and monitoring soil moisture.

[0003] Passive microwave remote sensing technology has successfully generated a series of global soil moisture products, such as AMSR-E, AMSR2, SMOS, Aquarius, and SMAP, but these products have low spatial resolution. While active microwave sensors can improve spatial resolution, they are susceptible to surface roughness and vegetation, which introduces many uncertainties when retrieving soil moisture from active microwave observation data. Summary of the Invention

[0004] The purpose of this invention is to provide a combined active and passive microwave inversion method and system that combines Bayesian probabilistic inversion to accurately estimate soil moisture and quantify the uncertainty in the inversion algorithm, thereby improving the accuracy and reliability of soil moisture estimation.

[0005] In a first aspect, the present invention provides a method for joint active and passive microwave inversion combining Bayesian probabilistic inversion, the method comprising: Collect active and passive microwave observation data of the target area; Samples are taken within the prior value range of the parameters to be inverted to obtain preliminary guesses. Active simulation data is obtained based on the preliminary guesses. The parameters to be inverted include soil moisture and roughness parameters. By combining the active simulation data and the active microwave observation data, active Bayesian probabilistic inversion is performed to obtain the active inversion result; The effective roughness parameters required for passive inversion are determined based on the roughness parameters obtained by active Bayesian probabilistic inversion and the pre-constructed relational function. By combining the effective roughness parameter, the initial guess value obtained from the prior value range of soil moisture, and the passive microwave observation data, a passive Bayesian probabilistic inversion is performed to obtain the passive inversion result; wherein, the active inversion result and the passive inversion result respectively include the optimal estimate value of the parameter to be inverted and the uncertainty quantification index. By combining the active inversion results and the passive inversion results, a joint inversion result is generated.

[0006] In an optional implementation, the step of performing active Bayesian probabilistic inversion by combining the active simulation data and active microwave observation data to obtain the active inversion result includes: Based on the active simulation data and active microwave observation data, a minimum cost function is constructed. Active Bayesian probabilistic inversion is performed under the guidance of the minimum cost function. When the inversion iteration meets the preset requirements, the active inversion result is obtained.

[0007] In an optional implementation, the step of obtaining the active inversion result when the inversion iteration meets the preset requirements includes: When the inversion iteration meets the preset requirements, the inversion iterations in which the function value of the cost function is less than the preset threshold are selected. Maximum likelihood estimation is performed based on the initial guesses used in the selected inversion iterations to obtain the optimal estimate and uncertainty quantification index of the parameters to be inverted.

[0008] In an optional implementation, the step of performing passive Bayesian probabilistic inversion by combining the effective roughness parameter, the initial guess value obtained from the prior value range of soil moisture, and the passive microwave observation data to obtain the passive inversion result includes: Passive simulation data is obtained based on the effective roughness parameters and the initial guesses obtained by sampling from the prior value range of soil moisture. By combining the passive simulation data and the passive microwave observation data, a passive Bayesian probabilistic inversion is performed to obtain the passive inversion result.

[0009] In an optional implementation, the step of performing passive Bayesian probabilistic inversion by combining the passive simulation data and the passive microwave observation data to obtain the passive inversion result includes: Based on the passive simulation data and passive microwave observation data, a minimum cost function is constructed. Passive Bayesian probabilistic inversion is performed under the guidance of the minimum cost function. When the inversion iteration meets the preset requirements, the passive inversion result is obtained.

[0010] In an optional implementation, the step of combining the active inversion result and the passive inversion result to generate a joint inversion result includes: Based on the optimal estimates and uncertainty quantification index of the parameters to be inverted in the active inversion results and the passive inversion results, the joint optimal estimate of the parameters to be inverted is calculated. Based on the uncertainty quantification index of the parameters to be inverted in the active inversion results and the passive inversion results, the joint uncertainty quantification index of the parameters to be inverted is calculated.

[0011] In an optional implementation, the method further includes: Based on the measured values ​​of soil moisture and roughness parameters in the measured data and the optimal estimated values ​​of soil moisture and roughness parameters in the joint inversion results, the index values ​​of each of the multiple error indices are calculated. The reliability of the joint inversion result is determined based on the index values ​​of each error index.

[0012] In an optional implementation, the method further includes a preprocessing step for the active microwave observation data and the passive microwave observation data, the step comprising: Perform one or more of the following processes on the active microwave observation data and the passive microwave observation data: geometric registration, time synchronization, outlier removal, and missing data completion.

[0013] In an optional implementation, the step of performing missing data completion processing on the active microwave observation data and the passive microwave observation data is achieved in the following way: Obtain time-series data of active and passive microwave observation data collected within a set time period; For time points where missing data exists, time interpolation is performed using existing data in the time series data to obtain the missing data for the corresponding time point.

[0014] Secondly, embodiments of the present invention provide a joint active-passive microwave inversion system combining Bayesian probabilistic inversion, the system comprising: The acquisition module is used to acquire active microwave observation data and passive microwave observation data of the target area; The processing module is used to sample within the prior value range of the parameters to be inverted to obtain preliminary guesses, and to obtain active simulation data based on the preliminary guesses. The parameters to be inverted include soil moisture and roughness parameters. The processing module is also used to perform active Bayesian probability inversion by combining the active simulation data and active microwave observation data to obtain active inversion results; The processing module is also used to determine the effective roughness parameters required for passive inversion based on the roughness parameters inverted by active Bayesian probabilistic inversion and the pre-constructed relational function. The processing module is also used to perform passive Bayesian probabilistic inversion by combining the effective roughness parameter, the initial guess value obtained by sampling from the prior value range of soil moisture, and the passive microwave observation data to obtain the passive inversion result; wherein, the active inversion result and the passive inversion result respectively include the optimal estimate value of the parameter to be inverted and the uncertainty quantification index. The processing module is also used to combine the active inversion results and the passive inversion results to generate joint inversion results.

[0015] This invention provides a method and system for joint active and passive microwave inversion combining Bayesian probabilistic inversion. Initial guesses are obtained by sampling within the prior value range of the parameters to be inverted. Active simulation is then performed based on these initial guesses to obtain active simulation data. Active inversion is performed by combining the active simulation data and active microwave observation data. Based on the roughness parameters and relational functions obtained from the active inversion, the effective roughness parameters required for passive inversion are determined. Passive inversion is then performed by combining the effective roughness parameters, the initial guesses of soil moisture, and the passive microwave observation data. The active and passive inversion results each include the optimal estimate of the parameters to be inverted and an uncertainty quantification index. The joint inversion results are obtained by combining the active and passive inversion results. In this scheme, the combination of active and passive microwave observations accurately estimates soil moisture and quantifies the uncertainty in the inversion algorithm, improving the accuracy and reliability of soil moisture estimation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart of the active and passive microwave joint inversion method provided in the embodiments of the present invention; Figure 2 A schematic diagram of the logical framework for the joint active and passive microwave inversion provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the results of active Bayesian probability inversion in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the results of joint inversion in an embodiment of the present invention; Figure 5 This is one of the schematic diagrams showing the results of passive Bayesian probability inversion in an embodiment of the present invention; Figure 6 This is the second schematic diagram of the result of passive Bayesian probability inversion in an embodiment of the present invention; Figure 7 This is the third schematic diagram of the result of passive Bayesian probability inversion in an embodiment of the present invention; Figure 8 This is a functional block diagram of the active and passive microwave joint inversion system provided in an embodiment of the present invention; Figure 9 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention.

[0019] Please see Figure 1 The above is a flowchart of the active-passive microwave joint inversion method combining Bayesian probabilistic inversion provided in an embodiment of the present invention. This active-passive microwave joint inversion method can be executed by an active-passive microwave joint inversion system, which can be implemented by software and / or hardware and can be configured in an electronic device, such as a computer device, server, programmable logic controller, etc. The detailed steps of the active-passive microwave joint inversion method are described below.

[0020] S11, collect active microwave observation data and passive microwave observation data of the target area.

[0021] S12, sampling is performed within the prior value range of the parameters to be inverted to obtain preliminary guesses. Active simulation data is obtained based on the preliminary guesses. The parameters to be inverted include soil moisture and roughness parameters.

[0022] S13. Active Bayesian probabilistic inversion is performed by combining active simulation data and active microwave observation data to obtain active inversion results.

[0023] S14. Based on the roughness parameters obtained from the active Bayesian probabilistic inversion and the pre-constructed relational function, determine the effective roughness parameters required for passive inversion.

[0024] S15. By combining the effective roughness parameter, the initial guess value obtained from the prior value range of soil moisture, and the passive microwave observation data, passive Bayesian probability inversion is performed to obtain the passive inversion result.

[0025] The active inversion results and passive inversion results include the optimal estimate of the parameters to be inverted and the uncertainty quantification index, respectively.

[0026] S16. Combine the active inversion results and the passive inversion results to generate joint inversion results.

[0027] In this embodiment, active microwave observation data is obtained by transmitting microwave signals through a data acquisition device and receiving signals reflected back from the target object. This data acquisition device can be, for example, synthetic aperture radar, a scatterometer, or a microwave radar.

[0028] Passive microwave observation data is obtained by acquiring microwave signals emitted or reflected by a target object using acquisition equipment. This acquisition equipment can be, for example, a microwave radiometer or a passive microwave sensor.

[0029] The collected active and passive microwave observation data can be preprocessed. Specifically, one or more of the following processes can be performed on the active and passive microwave observation data: geometric registration, time synchronization, outlier removal, and missing data completion.

[0030] Among them, geometric registration is performed by placing microwave radiation observation and corresponding soil and vegetation parameter observation at the same observation point during the experiment. That is, corresponding synchronous observation points for soil, vegetation, etc. are set up within the field of view of the microwave radiometer.

[0031] Time synchronization processing is achieved by keeping the observation frequency of the microwave radiometer consistent with the observation frequency of the soil, or by time aggregation during data processing.

[0032] Outlier removal can be achieved by detecting whether an observation exceeds a reasonable range. For example, the average value of time series data can be calculated, and a preset range can be set by adding or subtracting a set multiple of the standard deviation. If the value exceeds the preset range, it is identified as an outlier and removed.

[0033] Missing data completion can be achieved in the following ways: The system obtains time-series data of active and passive microwave observation data collected within a set time period. For time points with missing data, it uses existing data in the time-series data for time interpolation to obtain the missing data for the corresponding time points.

[0034] In other words, for missing data, time interpolation can be used to determine the missing data, thereby completing the missing data.

[0035] Preprocessing active and passive microwave observation data using any one or more of the above-mentioned preprocessing methods can ensure the integrity and reliability of the data.

[0036] Based on this, active Bayesian probabilistic inversion and passive Bayesian probabilistic inversion are performed using active microwave observation data and passive microwave observation data.

[0037] Currently, there are also methods using active and passive microwave remote sensing to improve the accuracy of soil moisture estimation. Existing combined active and passive methods can be mainly summarized into three categories: The first category assumes a linear relationship between soil moisture and active microwave observations, then uses passive microwave observation data to calibrate this relationship, thereby achieving soil moisture retrieval; the second category estimates soil moisture and other surface parameters by minimizing a cost function that reflects the difference between the observed data and the simulation results of the forward model. However, this process is relatively complex when interpreting the contributions of different observation data to the results; the third category first estimates vegetation and surface roughness parameters from active microwave observation data, then substitutes these parameters into the passive microwave forward model to estimate soil moisture. However, this sequential method, to some extent, limits the full utilization of active microwave observation data.

[0038] In addition, effective parameterization of surface roughness in the passive model is crucial in the coordinated process of active and passive microwave remote sensing. However, existing parameterization schemes have many problems, such as insufficient estimation accuracy and limited applicability. Moreover, existing algorithms do not fully exploit the complementarity between active and passive data in different frequency bands. Meanwhile, uncertainty quantification is a significant challenge in forward modeling and inversion algorithms for microwave remote sensing soil moisture. Most existing inversion methods neglect the estimation of uncertainty, which can easily lead to significant inversion errors.

[0039] Based on the above research findings, in this embodiment, active Bayesian probabilistic inversion and passive Bayesian probabilistic inversion are performed simultaneously. That is, active Bayesian probabilistic inversion and passive Bayesian probabilistic inversion each include multiple inversion iterations. After each iteration step of active Bayesian probabilistic inversion, the required parameters can be provided for the corresponding iteration steps of passive Bayesian probabilistic inversion, thereby reducing the number of uncertain parameters in the passive Bayesian inversion process and thus reducing uncertainty.

[0040] Specifically, preliminary values ​​can be obtained by sampling within the prior value range of the parameters to be inverted. The prior value range is the soil moisture m known based on prior knowledge. s and roughness parameter (H) rms For L, an autocorrelation function (usually exponential) needs to be set, specifying its numerical range. Sampling can be done randomly or in a specific order. This allows for the sampling of multiple pairs of soil moisture and roughness parameters. Please refer to [reference needed]. Figure 2 The initial guesses for the soil moisture and roughness parameters obtained from each sampling can be denoted as m. s,init H rms,init and L init .

[0041] Given soil moisture and roughness parameters obtained from each sampling, and combining these with soil temperature (ts), sand volume percentage (sand), clay content percentage (clay), radar incident angle (theta), and radar frequency (f), active simulation data is obtained, i.e., actively simulated backscattering. Based on this, active Bayesian probabilistic inversion is performed by combining the active simulation data and active microwave observation data to obtain the active inversion result. This step can be achieved through the following method: Based on active simulation data and active microwave observation data, a minimum cost function is constructed. Active Bayesian probabilistic inversion is performed under the guidance of minimizing the cost function. When the inversion iteration meets the preset requirements, the active inversion result is obtained.

[0042] In this embodiment, active Bayesian probabilistic inversion is achieved using active microwave observation data, such as L-band synthetic aperture radar observation data, and the obtained active simulation data.

[0043] The active microwave observation data consists of backscattered data from active observations. Based on this, active Bayesian probabilistic inversion is performed using both actively observed and actively simulated backscattered data. During the active Bayesian probabilistic inversion, the inversion process is guided by a constructed cost function, which is shown below:

[0044] in, and Let represent the backscattering from the active simulation and the backscattering from the active observation of the i-th channel, respectively, where n is the total number of channels.

[0045] Guided by the constructed minimum cost function, active Bayesian probabilistic inversion is performed. When the inversion iterations meet preset requirements, the active inversion result is obtained. Specifically, this can be achieved in the following way: When the inversion iteration meets the preset requirements, the inversion iterations in which the function value of the cost function minimized in multiple inversion iterations is less than the preset threshold are selected; based on the initial guess value used in the selected inversion iterations, the maximum likelihood estimation is performed to obtain the optimal estimate value of the parameter to be inverted and the uncertainty quantification index.

[0046] The preset requirement could be that the number of inversion iterations reaches a preset maximum, say 10,000. In each inversion iteration, the function value that minimizes the cost function is obtained. The function value is compared with a preset threshold to filter out inversion iterations where the function value is less than the preset threshold. For example, inversion iterations where the function value is less than the preset threshold can be selected out of 5,000 iterations.

[0047] Maximum likelihood estimation (MLE) is performed based on the initial guesses used in the selected inversion iterations, which generates posterior distributions, the optimal estimate, and uncertainty. The posterior distribution is the posterior probability distribution in Bayes' theorem. The optimal estimate is the best value of the posterior distribution, which can be the mean, median, or, in this embodiment, the maximum likelihood estimate of the generalized mechanism distribution. Uncertainty can be the standard deviation of the posterior probability distribution.

[0048] In this way, the optimal values ​​and uncertainty indices of the parameters to be inverted can be obtained. Specifically, this includes the optimal values ​​of soil moisture and roughness parameters. and their respective uncertainty quantification indicators .

[0049] The roughness parameters obtained from active Bayesian probabilistic inversion can be provided to the passive inversion process. This reduces the number of uncertain parameters in the passive inversion process, thereby lowering uncertainty.

[0050] Specifically, the effective roughness parameters required for passive inversion are determined based on the roughness parameters obtained by active Bayesian probabilistic inversion and the pre-constructed relational function.

[0051] The relational function f can be an exponential relational function, which can be fitted in advance based on the active and passive roughness parameters obtained from active and passive inversion tests, respectively.

[0052] In the application phase, the roughness parameters obtained by active inversion can be substituted into the relational function to obtain the effective roughness parameters required for passive inversion.

[0053] In addition, the soil moisture required for passive inversion is also obtained by sampling from the prior data range of soil moisture.

[0054] In passive inversion, the sampled soil moisture and the obtained effective roughness parameter are respectively expressed as m s,init H R,init and .

[0055] Based on this, passive Bayesian probabilistic inversion is performed by combining the effective roughness parameter, the initial guess value obtained from the prior value range of soil moisture, and passive microwave observation data to obtain the passive inversion result. Specifically, this step can be achieved in the following way: Passive simulation data is obtained based on the effective roughness parameter and the initial guess value obtained by sampling from the prior value range of soil moisture. Passive Bayesian probabilistic inversion is performed by combining the passive simulation data and passive microwave observation data to obtain the passive inversion result.

[0056] Similarly, a minimum cost function is constructed based on passive simulation data and passive microwave observation data. Passive Bayesian probabilistic inversion is performed under the guidance of minimizing the cost function. When the inversion iteration meets the preset requirements, the passive inversion result is obtained.

[0057] In this embodiment, the minimized cost function constructed here is similar to the minimized cost function constructed in the above-mentioned active Bayesian probability inversion process, the difference being that... and These represent the passive simulation data (passively simulated brightness and temperature data) and the active microwave observation data (observed brightness and temperature data) of the i-th channel, respectively.

[0058] Similarly, passive Bayesian probabilistic inversion is performed with the minimization of the cost function as the guiding principle.

[0059] After multiple inversion iterations, the iteration stops when the cost function reaches its minimum and no longer changes, or when the number of iterations reaches a preset maximum, yielding the final passive inversion result. The final passive inversion result specifically includes the optimal estimates of soil moisture and roughness parameters. and their respective uncertainty quantification indicators .

[0060] Based on this, a joint inversion result is generated by combining the active and passive inversion results. The purpose of estimating soil moisture in this embodiment is mainly based on the optimal estimate and uncertainty quantification index of soil moisture from the active inversion results, as well as the optimal estimate and uncertainty quantification index of soil moisture from the passive inversion results, to obtain the final joint inversion result. Specifically, this can be achieved in the following ways: Based on the optimal estimates and uncertainty quantification indices of the parameters to be inverted in the active and passive inversion results, the joint optimal estimates of the parameters to be inverted are calculated; based on the uncertainty quantification indices of the parameters to be inverted in the active and passive inversion results, the joint uncertainty quantification indices of the parameters to be inverted are calculated.

[0061] In this embodiment, the parameter to be inverted mainly refers to soil moisture. The optimal estimate of soil moisture and its uncertainty quantification index in the active inversion result are expressed as follows: In the passive inversion results, the optimal estimate of soil moisture and its uncertainty quantification index are expressed as follows: The final joint optimal estimate of soil moisture It can be calculated using the following formula:

[0062] The final joint uncertainty quantification index of soil moisture It can be calculated using the following formula:

[0063] Based on this, the active and passive microwave joint inversion method provided in this embodiment may further include the following steps: Based on the measured values ​​of soil moisture and roughness parameters in the actual data and the optimal estimates of soil moisture and roughness parameters in the joint inversion results, the index values ​​of each error index among multiple error indices are calculated; based on the index values ​​of each error index, the reliability of the joint inversion results is determined.

[0064] In this embodiment, the measured values ​​of soil moisture in the measured data can be obtained. Combined with the inverted values ​​of soil moisture obtained through joint inversion, the index values ​​of various error indicators can be obtained. Among them, the error indicators may include correlation coefficient, deviation value, root mean square error, unbiased root mean square error, etc.

[0065] By analyzing the values ​​of the various error indices, it is possible to determine whether the inversion values ​​of soil moisture pairs obtained through joint inversion are reliable, and thus evaluate the accuracy of the joint inversion.

[0066] To provide a clearer explanation of the implementation and effects of the active and passive microwave joint inversion method provided in this embodiment, the following explanation is combined with specific examples.

[0067] Using a pasture within a specific region as the study area, L-band multipolar backscattering data were acquired using GBSAR. After screening, aggregation, and synchronization processing, multiple sets of soil-backscattering data pairs were obtained over a certain time period, for example, 25 sets, for the active part of the CAP-BPI algorithm. L-band brightness temperature data were collected using an RPG-6CH-DP microwave radiometer. Brightness temperature data collected every 0.5 hours were processed daily by averaging, resulting in multiple sets of soil-brightness temperature data pairs, for example, 32 sets, for the passive part. Simultaneously, parameters such as soil moisture, temperature, surface roughness, soil texture, and vegetation water content were measured. In this method, active and passive Bayesian probabilistic inversion are performed simultaneously. Soil moisture is jointly inverted using the active microwave backscattering coefficient and the passive microwave brightness temperature observation. Relevant parameters are initialized for both the active and passive microwave components separately. Through simulation model calculations, minimizing the cost function, MCMC (Markov Chain Monte Carlo) updates, and MLE (Maximum Likelihood Estimation), the optimal estimates and uncertainty quantification indices for each component are obtained. Then, the optimal estimates and uncertainty quantification indices for both components are combined to calculate the joint optimal estimate and joint uncertainty quantification index. Finally, the performance of the inversion algorithm in soil moisture estimation is verified by comparing it with ground-measured values.

[0068] To examine the impact of the values ​​of parameters (A, B, and α) in the WCM on the inversion results, the inversion results for all land uses, grasslands, pastures, and winter wheat across four different land surfaces were compared with ground-measured results for verification. Figure 3 As shown, Figure 3 The subplots, from left to right and top to bottom, represent the soil moisture retrieval results based on active Bayesian probabilistic inversion using different land surface parameters (A, B, α). The horizontal axis of each subplot represents the observed soil moisture value (Observed SM), and the vertical axis represents the estimated SM value (Estimated SM). The values ​​represent all land use types, pasture, grassland, and winter wheat, respectively. The points in the figures represent scatter plots of the estimated and observed values.

[0069] The comparison results show that active Bayesian probabilistic inversion can effectively retrieve soil moisture under different surface conditions. Correlation coefficient (R) analysis shows that the R values ​​for the four surface types range from 0.779 to 0.813, indicating a strong positive correlation between the estimated and measured values. This means that active Bayesian probabilistic inversion can capture the changing trends of soil moisture well. Among them, the grassland surface inversion result has the highest R value, reaching 0.813, indicating that under this surface condition, active Bayesian probabilistic inversion is most effective in retrieving soil moisture and can more accurately reflect the actual soil moisture situation. Meanwhile, the bias values ​​for the four surface types range from -0.046 to -0.048, all negative, indicating that the active Bayesian probabilistic inversion results slightly underestimate soil moisture. However, this underestimation is relatively small and has a limited impact on the overall accuracy of the inversion results. Root mean square error (RMSE) and unbiased root mean square error (ubRMSE) also reflect the accuracy of the inversion results. The RMSE values ​​are between 0.069 and 0.07, and the ubRMSE values ​​are between 0.051 and 0.052, indicating that the error in the inversion process is within an acceptable range and the random error is relatively small.

[0070] Figure 4The figure shows the soil moisture inversion results using L-band dual-polarization brightness temperature data and backscattering data. Specifically, it shows the results of soil moisture inversion using L-band dual-polarization brightness temperature data and backscattering data based on a parallel strategy within the CAP-BPI framework, and the comparison with ground measurements. The results show that the soil moisture inverted under this strategy has a strong correlation with the measured values ​​(R = 0.973), but there is a certain degree of underestimation (Bias = -0.031). The root mean square error (RMSE = 0.053) and unbiased root mean square error (ubRMSE = 0.043) are relatively small. Overall, it can effectively capture changes in soil moisture and the inversion effect is good.

[0071] The results of soil moisture inversion using passive Bayesian probabilistic inversion and comparison with ground measurements are as follows: Figure 5 The figures show the corresponding results in the L-band, C-band, and X-band, respectively. The results indicate that the L-band has the highest correlation coefficient (R = 0.977), and the smallest bias (Bias = -0.022), root mean square error (RMSE = 0.046), and unbiased root mean square error (ubRMSE = 0.04), resulting in the best inversion performance. The C-band has a correlation coefficient of 0.854, with all error indicators increasing compared to the L-band, indicating the second best inversion performance. The X-band has a correlation coefficient of only 0.728, and the largest bias, RMSE, and ubRMSE among the three, resulting in the worst inversion performance.

[0072] Soil moisture inversion was performed using multi-frequency (L, C, X band) brightness temperature data via passive Bayesian probabilistic inversion, and the results were compared with ground measurements, as shown below. Figure 6 As shown in the figure. The results show that the correlation coefficient (R = 0.89) of the V-polarized brightness temperature data inversion is higher than that of the H-polarized data (R = 0.878), while the bias (Bias = -0.051), root mean square error (RMSE = 0.065), and unbiased root mean square error (ubRMSE = 0.04) are also lower than those of the H-polarized data (-0.059, 0.074, and 0.045, respectively). This indicates that in the passive BPI soil moisture inversion based on multi-frequency brightness temperature data, the V-polarized brightness temperature data inversion effect is better than that of the H-polarized data.

[0073] Soil moisture inversion was performed using dual-frequency / multi-frequency dual-polarization brightness temperature data via passive Bayesian probabilistic inversion, and the results were compared with ground measurements, as follows: Figure 7As shown in the figure. The results indicate that the soil moisture retrieved by each frequency combination is correlated with the measured values ​​to some extent (R values ​​are all greater than 0.82), and there are varying degrees of underestimation. Among them, the L-C band combination has the highest correlation coefficient (R = 0.922), the smallest bias (Bias = -0.045) and root mean square error (RMSE = 0.061), and the best retrieval effect. The C-X band combination has a relatively weak correlation, and relatively high RMSE and bias. Overall, multi-frequency combinations have application potential in soil moisture retrieval, but different combinations show different performance characteristics.

[0074] The active and passive microwave joint inversion method provided in this embodiment combines active and passive microwave observations, fully utilizing the advantages of both. Compared to using active or passive microwave observation data alone, it significantly improves the accuracy of soil moisture estimation. Furthermore, it effectively quantifies uncertainty; the Bayesian probabilistic inversion method can quantify the uncertainty in soil moisture inversion, providing more comprehensive and reliable estimation results and enhancing the understanding and application value of the inversion results.

[0075] Based on the same inventive concept, please refer to Figure 8 This invention also provides a functional module diagram of a combined active and passive microwave inversion system incorporating Bayesian probabilistic inversion. This embodiment divides the active and passive microwave joint inversion system into functional modules based on the above-described method embodiments. For example, each function can be divided into its own functional modules, or two or more functions can be integrated into a single processing module. The integrated modules can be implemented in hardware or as software functional modules. It should be noted that the module division in this invention embodiment is illustrative and represents only one logical functional division; other division methods may be used in actual implementation.

[0076] For example, when dividing functional modules according to their respective functions, Figure 8 The active-passive microwave joint inversion system shown is only a schematic diagram of a device. The active-passive microwave joint inversion system may include an acquisition module and a processing module. The functions of each module of this active-passive microwave joint inversion system will be described in detail below.

[0077] The acquisition module is used to acquire active microwave observation data and passive microwave observation data of the target area; The processing module is used to sample within the prior value range of the parameters to be inverted to obtain preliminary guesses, and to obtain active simulation data based on the preliminary guesses. The parameters to be inverted include soil moisture and roughness parameters. The processing module is also used to perform active Bayesian probability inversion by combining the active simulation data and active microwave observation data to obtain active inversion results; The processing module is also used to determine the effective roughness parameters required for passive inversion based on the roughness parameters inverted by active Bayesian probabilistic inversion and the pre-constructed relational function. The processing module is also used to perform passive Bayesian probabilistic inversion by combining the effective roughness parameter, the initial guess value obtained by sampling from the prior value range of soil moisture, and the passive microwave observation data to obtain the passive inversion result; wherein, the active inversion result and the passive inversion result respectively include the optimal estimate value of the parameter to be inverted and the uncertainty quantification index. The processing module is also used to combine the active inversion results and the passive inversion results to generate joint inversion results.

[0078] The active-passive microwave joint inversion system provided in this embodiment can be used to execute the active-passive microwave joint inversion method under any of the above embodiments. For details not covered in this embodiment, please refer to the corresponding descriptions in the above embodiments. This embodiment will not elaborate further here.

[0079] Please see Figure 9 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. The electronic device can be a computer device, server, or similar component in a back-end analysis platform. The electronic device includes a memory, a processor, and a communication module. The memory, processor, and communication module are electrically connected directly or indirectly to each other to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0080] The memory is used to store computer programs or data. Memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc.

[0081] The processor is used to read / write data or programs stored in the memory and to execute the active and passive microwave joint inversion method provided in any embodiment of the present invention.

[0082] The communication module is used to establish communication connections between electronic devices and other communication terminals via a network, and to send and receive data via the network.

[0083] It should be understood that, Figure 9 The structure shown is only a schematic diagram of an electronic device; the electronic device may also include components that are larger than those shown. Figure 9 The more or fewer components shown, or having the same Figure 9 The different configurations shown.

[0084] Furthermore, embodiments of the present invention also provide a computer-readable storage medium storing machine-executable instructions, which, when executed, implement the active-passive microwave joint inversion method provided in the above embodiments.

[0085] Specifically, the computer-readable storage medium can be a general-purpose storage medium, such as a removable disk or hard disk. When the computer program on the computer-readable storage medium is run, it can execute the aforementioned active and passive microwave joint inversion method. The processes involved in the execution of the executable instructions on the computer-readable storage medium can be referred to the relevant descriptions in the above method embodiments, and will not be detailed here.

[0086] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0087] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0088] Furthermore, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0089] It should be noted that if the functionality is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0090] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0091] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for joint active and passive microwave inversion combining Bayesian probabilistic inversion, characterized in that, The method includes: Collect active and passive microwave observation data of the target area; Samples are taken within the prior value range of the parameters to be inverted to obtain preliminary guesses. Active simulation data is obtained based on the preliminary guesses. The parameters to be inverted include soil moisture and roughness parameters. By combining the active simulation data and the active microwave observation data, active Bayesian probabilistic inversion is performed to obtain the active inversion result; The effective roughness parameters required for passive inversion are determined based on the roughness parameters obtained by active Bayesian probabilistic inversion and the pre-constructed relational function. By combining the effective roughness parameter, the initial guess value obtained from the prior value range of soil moisture, and the passive microwave observation data, a passive Bayesian probabilistic inversion is performed to obtain the passive inversion result; wherein, the active inversion result and the passive inversion result respectively include the optimal estimate value of the parameter to be inverted and the uncertainty quantification index. By combining the active inversion results and the passive inversion results, a joint inversion result is generated.

2. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 1, characterized in that, The step of performing active Bayesian probabilistic inversion by combining the active simulation data and active microwave observation data to obtain the active inversion result includes: Based on the active simulation data and active microwave observation data, a minimum cost function is constructed. Active Bayesian probabilistic inversion is performed under the guidance of the minimum cost function. When the inversion iteration meets the preset requirements, the active inversion result is obtained.

3. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 2, characterized in that, The step of obtaining the active inversion result when the inversion iteration meets the preset requirements includes: When the inversion iteration meets the preset requirements, the inversion iterations in which the function value of the cost function is less than the preset threshold are selected. Maximum likelihood estimation is performed based on the initial guesses used in the selected inversion iterations to obtain the optimal estimate and uncertainty quantification index of the parameters to be inverted.

4. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 1, characterized in that, The step of performing passive Bayesian probabilistic inversion by combining the effective roughness parameter, the initial guess value obtained from the prior value range of soil moisture, and the passive microwave observation data to obtain the passive inversion result includes: Passive simulation data is obtained based on the effective roughness parameters and the initial guesses obtained by sampling from the prior value range of soil moisture. By combining the passive simulation data and the passive microwave observation data, a passive Bayesian probabilistic inversion is performed to obtain the passive inversion result.

5. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 4, characterized in that, The step of performing passive Bayesian probabilistic inversion by combining the passive simulation data and passive microwave observation data to obtain the passive inversion result includes: Based on the passive simulation data and passive microwave observation data, a minimum cost function is constructed. Passive Bayesian probabilistic inversion is performed under the guidance of the minimum cost function. When the inversion iteration meets the preset requirements, the passive inversion result is obtained.

6. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 1, characterized in that, The step of combining the active inversion results and the passive inversion results to generate joint inversion results includes: Based on the optimal estimates and uncertainty quantification index of the parameters to be inverted in the active inversion results and the passive inversion results, the joint optimal estimate of the parameters to be inverted is calculated. Based on the uncertainty quantification index of the parameters to be inverted in the active inversion results and the passive inversion results, the joint uncertainty quantification index of the parameters to be inverted is calculated.

7. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 1, characterized in that, The method further includes: Based on the measured values ​​of soil moisture and roughness parameters in the measured data and the optimal estimated values ​​of soil moisture and roughness parameters in the joint inversion results, the index values ​​of each of the multiple error indices are calculated. The reliability of the joint inversion result is determined based on the index values ​​of each error index.

8. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 1, characterized in that, The method further includes a preprocessing step for the active microwave observation data and the passive microwave observation data, which includes: Perform one or more of the following processes on the active microwave observation data and the passive microwave observation data: geometric registration, time synchronization, outlier removal, and missing data completion.

9. The active-passive microwave joint inversion method combining Bayesian probabilistic inversion according to claim 8, characterized in that, The step of performing missing data completion processing on the active microwave observation data and the passive microwave observation data is implemented in the following way: Obtain time-series data of active and passive microwave observation data collected within a set time period; For time points where missing data exists, time interpolation is performed using existing data in the time series data to obtain the missing data for the corresponding time point.

10. A joint active-passive microwave inversion system combining Bayesian probabilistic inversion, characterized in that, The system includes: The acquisition module is used to acquire active microwave observation data and passive microwave observation data of the target area; The processing module is used to sample within the prior value range of the parameters to be inverted to obtain preliminary guesses, and to obtain active simulation data based on the preliminary guesses. The parameters to be inverted include soil moisture and roughness parameters. The processing module is also used to perform active Bayesian probability inversion by combining the active simulation data and active microwave observation data to obtain active inversion results; The processing module is also used to determine the effective roughness parameters required for passive inversion based on the roughness parameters inverted by active Bayesian probabilistic inversion and the pre-constructed relational function. The processing module is also used to perform passive Bayesian probabilistic inversion by combining the effective roughness parameter, the initial guess value obtained by sampling from the prior value range of soil moisture, and the passive microwave observation data to obtain the passive inversion result; wherein, the active inversion result and the passive inversion result respectively include the optimal estimate value of the parameter to be inverted and the uncertainty quantification index. The processing module is also used to combine the active inversion results and the passive inversion results to generate joint inversion results.

Images