GNSS-R soil humidity inversion method, equipment and medium

Abstract

The invention discloses a GNSS-R soil humidity inversion method and device and a medium, and relates to the field of data processing.The method comprises the steps that the soil reflectivity and the reflected signal phase of a research area are determined based on a reflected signal of a global navigation satellite system, and a normalized vegetation index is obtained; determining a temperature drought vegetation index based on the normalized vegetation index; performing space-time matching on the soil reflectivity, the reflected signal phase and the temperature drought vegetation index according to the timestamp of the reflected signal of the global navigation satellite system and the longitude and latitude information of the research area to obtain input data; performing normalization processing on the input data to obtain normalized data; constructing an NARX neural network, optimizing parameters of the NARX neural network by adopting a swarm intelligence optimization algorithm, and obtaining an inversion model based on the NARX neural network after parameter optimization; and inputting the normalized data into an inversion model to obtain a soil humidity inversion result. According to the invention, the inversion precision and the time resolution can be improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a GNSS-R soil moisture inversion method, device and medium. Background Art

[0002] Soil moisture is an important geophysical parameter. Continuously and accurately obtaining soil moisture data is of great significance to agricultural production, ecological environment, water resources management and meteorological forecasting. At present, with the development of the global satellite navigation system, the Global Navigation Satellite System-Reflectometry (GNSS-R) remote sensing technology has gradually matured. It has been widely used because it can achieve all-weather, high-temporal and spatial resolution, and low-cost earth observation based on different platforms.

[0003] When GNSS-R is used for soil moisture monitoring, the inversion accuracy is not ideal due to the complex surface reflection conditions and the interference of vegetation, surface roughness and other factors on the signal. In addition, environmental factors, weather conditions and terrain characteristics will affect the quality of GNSS-R data to a certain extent, and will also affect the inversion accuracy and spatiotemporal resolution. In addition, the existing signal reflection model is affected by factors such as soil surface roughness and vegetation coverage, and cannot achieve accurate inversion of soil moisture. Summary of the invention

[0004] The purpose of this application is to provide a GNSS-R soil moisture inversion method, device and medium, which can improve the inversion accuracy and time resolution.

[0005] To achieve the above objectives, this application provides the following solutions: In a first aspect, the present application provides a GNSS-R soil moisture inversion method, comprising: Based on the global navigation satellite system reflection signal, the soil reflectivity and reflection signal phase of the study area are determined, and the Normalized Difference Vegetation Index (NDVI) is obtained; the global navigation satellite system reflection signal includes: GNSS scattered signal, direct signal complex correlation value, GNSS (Global Navigation Satellite System) signal wave number, satellite altitude angle, direct signal receiving antenna gain in the direction of mirror reflection point, reflected signal receiving antenna gain in the direction of mirror reflection point, distance from mirror reflection point to receiver, and satellite to direct antenna spacing; Determining a Temperature Vegetation Dryness Index (TDVI) based on the NDVI; According to the timestamp of the global navigation satellite system reflection signal and the longitude and latitude information of the study area, the soil reflectivity, the reflection signal phase and the temperature drought vegetation index are temporally and spatially matched to obtain input data; Normalizing the input data to obtain normalized data; Constructing a NARX (nonlinear auto-regressive model with exogenous inputs) neural network, optimizing the parameters of the NARX neural network using a swarm intelligence optimization algorithm, and obtaining an inversion model based on the NARX neural network after parameter optimization; The normalized data is input into the inversion model to obtain a soil moisture inversion result.

[0006] Optionally, determining soil reflectivity and reflection signal phase of a study area based on a global navigation satellite system reflection signal includes: Determine a correlation power peak ratio of a direct and reverse signal based on a complex correlation value of the GNSS scattered signal and the direct signal; Determine an antenna gain correction factor based on a gain of a direct signal receiving antenna in the direction of a mirror reflection point and a gain of a reflected signal receiving antenna in the direction of a mirror reflection point; Determine the propagation path compensation factor based on the distance from the mirror reflection point to the receiver and the distance from the satellite to the direct antenna; Determine the soil reflectivity based on the surface root mean square height, the GNSS signal wave number, the satellite altitude angle, the antenna gain correction factor, the propagation path compensation factor, and the correlation power peak ratio of the direct and reverse signals; The carrier and pseudo code of the reflected signal are stripped by using the local carrier and pseudo code generated by the direct signal tracking link to obtain the reflected signal phase.

[0007] Optionally, a swarm intelligence optimization algorithm is used to optimize the parameters of the NARX neural network, and an inversion model is obtained based on the NARX neural network after the parameter optimization, including: Optimizing the parameters of the NARX neural network using a swarm intelligence optimization algorithm to obtain the NARX neural network after parameter optimization; Get a sample dataset; The sample data set is used to train the NARX neural network with optimized parameters until a set number of training times or a set error threshold is reached, and the trained NARX neural network with optimized parameters is used as the inversion model.

[0008] Optionally, obtain a sample data set, including: Acquire soil reflectivity historical data, reflection signal phase historical data and temperature drought vegetation index historical data, and generate sample input data based on the soil reflectivity historical data, the reflection signal phase historical data and the temperature drought vegetation index historical data; Acquire soil moisture historical data corresponding to the sample input data; Performing time-space matching and normalization processing on the sample input data to obtain normalized sample data; A training sample pair is formed based on the normalized sample data and the soil moisture historical data to obtain the sample data set.

[0009] Optionally, generating sample input data based on the soil reflectivity historical data, the reflection signal phase historical data and the temperature drought vegetation index historical data includes: Pre-classifying the soil reflectivity historical data and the reflection signal phase historical data by land type to obtain a pre-classification result; The pre-classification result and the temperature drought vegetation index historical data are used as sample input data.

[0010] Optionally, determining a temperature drought vegetation index based on the normalized difference vegetation index comprises: Using the dry-wet edge fitting equation, the maximum surface temperature and the minimum surface temperature are determined based on the normalized difference vegetation index; The real surface temperature is obtained, and the temperature drought vegetation index is determined based on the real surface temperature, the maximum surface temperature and the minimum surface temperature.

[0011] Optionally, the dry-wet edge fitting equation is expressed as: ; In the formula, is the maximum surface temperature, is the lowest surface temperature, are the coefficients of the linear fitting equations of the dry and wet edges, Represents the Normalized Difference Vegetation Index.

[0012] Optionally, the swarm intelligence optimization algorithm adopted is a particle swarm optimization algorithm.

[0013] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the GNSS-R soil moisture inversion method provided above.

[0014] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the GNSS-R soil moisture inversion method provided above.

[0015] According to the specific embodiments provided in this application, this application has the following technical effects: The present application provides a GNSS-R soil moisture inversion method, device and medium, which uses the parameter-optimized NARX neural network as the inversion model, so that a delay and feedback mechanism is added to the entire inversion process to enhance the memory of historical data and improve the time resolution. In addition, by using a swarm intelligence optimization algorithm to optimize the parameters of the NARX neural network, and using the soil reflectivity, reflection signal phase and temperature drought vegetation index after time-space matching as the input data of the inversion model, the inversion accuracy can be significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 A schematic diagram of a flow chart of a GNSS-R soil moisture inversion method provided in one embodiment of the present application; Figure 2 A schematic diagram of signal scattering on a smooth surface provided in one embodiment of the present application; Figure 3 A schematic diagram of signal scattering on a micro-rough surface provided by another embodiment of the present application; Figure 4 A schematic diagram of signal scattering on an extremely rough surface provided in one embodiment of the present application; Figure 5 A schematic diagram of the phase of the surface reflection coefficient provided in an embodiment of the present application; Figure 6 A schematic diagram of the sensitivity of the surface reflection coefficient phase to soil moisture changes provided by another embodiment of the present application; Figure 7 A flowchart of optimizing NARX neural network parameters using a particle swarm optimization algorithm provided in another embodiment of the present application; Figure 8 A schematic diagram of the structure of a NARX neural network provided in another embodiment of the present application; Fig. 9 A schematic diagram of GNSS-R soil moisture inversion results for a first time period provided in another embodiment of the present application; Fig.10 A schematic diagram of GNSS-R soil moisture inversion results for a second time period provided in another embodiment of the present application; Fig.11 A schematic diagram of GNSS-R soil moisture inversion results for a third time period provided in another embodiment of the present application; Fig.12 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments.

[0019] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of this application.

[0020] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0021] In an exemplary embodiment, the present application provides a GNSS-R soil moisture inversion method, which is executed by a computer device, and specifically can be executed by a computer device such as a terminal or a server alone, or can be executed by a terminal and a server together.

[0022] In the embodiment of the present application, the method is applied to a server as an example for explanation. Figure 1 As shown, the method includes: Step 100: Determine the soil reflectivity and reflected signal phase of the study area based on the global navigation satellite system reflected signal, and obtain the normalized vegetation index. The global navigation satellite system reflected signal includes: GNSS scattered signal, direct signal complex correlation value, GNSS signal wave number, satellite altitude angle, direct signal receiving antenna gain in the direction of the mirror reflection point, reflected signal receiving antenna gain in the direction of the mirror reflection point, the distance from the mirror reflection point to the receiver, and the distance from the satellite to the direct antenna.

[0023] Step 101: Determine a temperature drought vegetation index based on a normalized difference vegetation index.

[0024] Step 102: According to the timestamp of the global navigation satellite system reflection signal and the longitude and latitude information of the study area, the soil reflectivity, the reflection signal phase and the temperature drought vegetation index are matched in time and space to obtain input data.

[0025] Step 103: normalize the input data to obtain normalized data.

[0026] Step 104: construct a NARX neural network, use a swarm intelligence optimization algorithm to optimize the parameters of the NARX neural network, and obtain an inversion model based on the NARX neural network after parameter optimization.

[0027] Step 105: Input the normalized data into the inversion model to obtain the soil moisture inversion result.

[0028] In another exemplary embodiment of the present application, the reflecting surface is a non-smooth plane, and the ground surface should be regarded as an uneven rough surface. Therefore, when the signal is reflected, the reflected signal does not propagate along a single theoretical direction, and part of the signal will be diffusely reflected. In this process, the reflected signal propagating along the theoretical direction (i.e., the mirror direction) is the coherent component, that is, the required useful signal. The diffusely reflected signal is the incoherent component.

[0029] Ideally, all reflected signals are coherent components. The greater the surface roughness, the fewer the coherent components. Figure 2-Figure 4 As shown, the signal received by the antenna is the sum of various signals in the Fresnel reflection zone, among which the coherent components are called coherent components of the reflected signal, and the incoherent components are called incoherent components.

[0030] Based on this, the GNSS scattered signal is : (1) In the formula, For relevant components, are incoherent components, namely: (2) In the formula, is the navigation satellite transmission power, is the transmitting antenna gain in the direction of the mirror reflection point, is the gain of the reflected signal receiving antenna in the direction of the mirror reflection point, and Respectively represent the distance from the mirror reflection point to the receiver and the navigation satellite, is the root mean square height of the ground surface, is the GNSS signal wave number, is the satellite altitude angle, is the soil reflectivity of the bare surface, is the wavelength, χ is the amplitude of the Woodward fuzzy function, is the position vector relative to the mirror reflection point, is the incoherent surface RMS height component, is the time delay between each scattering surface element in the effective observation area and the current local signal The difference, The frequency of each scattering element in the effective observation area and the current local signal carrier frequency The difference between . and All are intermediate parameters.

[0031] Since most of the signal components reflected by the ground are dominated by coherent components, in practical applications, only the coherent components are taken for calculation, which is: (3) In order to calculate the soil reflectivity of the exposed surface, first calculate the relative power peak ratio of the direct and reflected signals: (4) In the formula, is the peak power ratio of the direct and reverse signals, peak(·) is the peak operator, is the complex correlation value of the direct signal. When the tracking error is ignored, we have: (5) In the formula, is the gain of the direct signal receiving antenna in the direction of the mirror reflection point, is the distance from the satellite to the direct antenna.

[0032] Substituting formula (3) and formula (5) into formula (4), we can obtain: (6) Determine the antenna gain correction factor and propagation path compensation factor ,have: (7) The soil reflectivity of the bare surface can finally be expressed as: (8) Furthermore, in GNSS-R soil moisture remote sensing, the local carrier and pseudo code generated by the direct signal tracking link are usually used to directly strip the reflected signal carrier and pseudo code. At this time, the reflected signal phase The direct signal phase component in Removed.

[0033] The reflected signal phase can be obtained for: (9) In the formula, is the phase of the surface reflection coefficient, which is related to the dielectric properties of the reflecting surface. The phase difference is caused by the difference in signal propagation path, which is determined by the height between the antenna and the reflecting surface.

[0034] based on Figure 5 and Figure 6 The reflection coefficient phase and its sensitivity to soil moisture changes can fully prove that the reflection coefficient phase is highly correlated with soil moisture and can be used as the main input data for soil moisture inversion.

[0035] Based on the above description, the implementation process of step 100 provided above in the present application may include: Step 1001: Determine the correlation power peak ratio of the direct and reflected signals based on the complex correlation value of the GNSS scattered signal and the direct signal.

[0036] Step 1002: Determine an antenna gain correction factor based on the gain of the direct signal receiving antenna in the direction of the mirror reflection point and the gain of the reflected signal receiving antenna in the direction of the mirror reflection point.

[0037] Step 1003: Determine the propagation path compensation factor based on the distance from the mirror reflection point to the receiver and the distance from the satellite to the direct antenna.

[0038] Step 1004: Determine soil reflectivity based on the surface RMS height, GNSS signal wave number, satellite altitude angle, antenna gain correction factor, propagation path compensation factor, and the relative power peak ratio of direct and reflected signals.

[0039] Step 1005: Use the local carrier and pseudo code generated by the direct signal tracking link to strip the carrier and pseudo code of the reflected signal to obtain the reflected signal phase.

[0040] In another exemplary embodiment of the present application, in order to further improve the accuracy of inversion, in this embodiment, the implementation process of the above step 104 can be described as: Step 1041: Optimize the parameters of the NARX neural network using a swarm intelligence optimization algorithm to obtain a NARX neural network with optimized parameters.

[0041] Step 1042: Obtain a sample data set.

[0042] Step 1043: Use the sample data set to train the NARX neural network with optimized parameters until a set number of training times or a set error threshold is reached, and then use the trained NARX neural network with optimized parameters as an inversion model.

[0043] In the implementation process, in order to obtain better network initial weights and thresholds and improve model accuracy, the particle swarm optimization algorithm can be used as a group intelligence optimization algorithm to adjust the parameters of the NARX neural network. The specific optimization process is as follows: Figure 7 shown.

[0044] Particle swarm optimization (PSO) is an improved global optimization algorithm that originated from the study of the foraging behavior of biological groups such as bird flocks or fish schools in nature.

[0045] PSO is widely used because it can adaptively adjust the search strategy and show good robustness in different optimization environments.

[0046] The update formula for the velocity and position of each particle is: (10) In the formula, and are learning factors, and are all random numbers in the interval (0,1). represents the velocity of the particle, express The position of the particle at time and They are The individual and global historical optimal position values ​​at each moment. for The inertia weight factor at the moment, is the initial inertia weight, is the inertia weight when it iterates to the maximum number of times, The maximum number of iterations, g is the global position, for The speed of the particle at that moment.

[0047] Furthermore, the implementation process of the above step 1042 includes: (1) Obtain historical data on soil reflectivity, reflection signal phase, and temperature drought vegetation index.

[0048] The sample input data is generated based on the soil reflectance historical data, the reflection signal phase historical data and the temperature drought vegetation index historical data.

[0049] (2) Obtain historical soil moisture data corresponding to the sample input data.

[0050] (3) Perform spatiotemporal matching and normalization on the sample input data to obtain normalized sample data.

[0051] (4) Based on the normalized sample data and soil moisture historical data, training sample pairs are formed to obtain a sample data set.

[0052] Because land type will greatly affect the inversion results, for example, soil moisture in desert areas is generally lower and vegetation has less impact.

[0053] Wetlands generally have high soil moisture or even saturation, while woodland types are greatly affected by vegetation.

[0054] In order to solve this problem, in this embodiment, land type classification is performed in advance, which can effectively improve the noise resistance of the inversion model and improve the inversion accuracy.

[0055] Based on this, the implementation process of the above step (1) can be described as: (11) The soil reflectivity historical data and the reflection signal phase historical data are pre-classified using land type to obtain pre-classification results.

[0056] In order to further improve the inversion accuracy, pre-classification can be performed according to IGBP (International Geosphere-Biosphere Program).

[0057] The IGBP land type classification is a classification system widely used in global change research. It divides the earth's surface into different land cover types based on vegetation characteristics, climate characteristics and land use types.

[0058] Each set of soil reflectivity historical data and reflection signal phase historical data is assigned different labels according to the category.

[0059] Among them, the IGBP classification after removing water bodies and snow is shown in Table 1.

[0060] Table 1 IGBP classification table

[0061] (12) The pre-classified input data and the historical data of the temperature drought vegetation index are used as sample input data.

[0062] In another exemplary embodiment of the present application, the process of obtaining NDVI data and calculating TDVI data of the study area includes: The normalized difference vegetation index (NDVI) is calculated using the following formula: (11) Where NIR is the reflectivity in the near infrared band, and RED is the reflectivity in the red band.

[0063] The linear fitting is performed by taking the NDVI of the study area extracted as the horizontal coordinate and the maximum and minimum surface temperature as the vertical coordinate to form a The slope of the feature space is taken as the contour line of TVDI. TVDI rises from bottom to top. The larger the absolute value of TVDI, the smaller the moisture content of the soil surface. The smaller the absolute value of TVDI, the greater the moisture content of the soil surface. Therefore, the dry-wet edge equation of the feature space is fitted in the study area, and the drought index of each pixel can be obtained. The calculation formula of TVDI is: (12) Where TVDI is the temperature drought vegetation index, is the true surface temperature, The highest surface temperature, is the lowest surface temperature. The larger the TVDI value, the lower the soil surface moisture content and the greater the drought degree. The dry-wet edge fitting equation of TVDI is: (13) In the formula, are the coefficients of the linear fitting equation for the dry and wet edges, respectively.

[0064] In another exemplary embodiment of the present application, the NARX neural network is a neural network model commonly used in time series prediction and system modeling. Unlike traditional feedforward neural networks, it also introduces past output values ​​into the input of the next moment at the same time, simulating a nonlinear autoregressive relationship. Its training method is divided into two stages: forward propagation and back propagation. Among them, the forward propagation stage is to pass historical input and output data to the network and generate the current predicted output. The back propagation stage is to update the weights and biases of the network according to the prediction error to improve the accuracy of the model.

[0065] To construct Figure 8 Taking the NARX neural network structure shown in FIG. 1 as an example, the implementation process of the above step 1043 is described. The NARX neural network includes an input layer, a hidden layer, and an output layer. The number of nodes in the hidden layer is 8, and the activation function selects the hyperbolic tangent function tanh. Figure 8 middle, is the delay operator. Based on this, the NARX neural network can be expressed as: (14) In the formula, and The neural network is Input and output at any time. and are the input and output delay orders respectively. is the nonlinear function obtained by network fitting. It can be seen that the output at the next moment The size of depends on the output of the previous moment And input .

[0066] Soil reflectivity and reflection coefficient phase are the main input data, and TDVI is the auxiliary data for data fusion. After NARX neural network fusion training, the output data, namely soil moisture (SMC), can be predicted.

[0067] Based on the above description, the training and inversion process of the NARX neural network after parameter optimization can be described as: 1. Filter out data with the same land type label. Due to the nature of the dual-antenna method, data with incident angles less than 20 degrees and greater than 65 degrees are deleted, and then spatiotemporal matching is performed based on the timestamps of the three input data and the latitude and longitude information to eliminate the negative impact of erroneous data on the model.

[0068] 2. Normalize the input data (scale to [0,1]) to ensure that the impact of data of different dimensions on network training is not unbalanced.

[0069] 3. Input the matched input data into the NARX neural network after parameter optimization. The input layer includes three types of input data at the current moment: and the soil moisture value predicted by the model at the L-order delay ( ).

[0070] 4. Using the open-loop structure to train the NARX neural network with optimized parameters, the hidden layer performs weighted and activation function processing on the three input data to capture the interaction relationship and time series characteristics between the input variables: (1) The parameter-optimized NARX neural network captures the dynamic patterns of reflectivity and reflectivity phase over time that affect humidity and extracts features related to soil moisture.

[0071] (2) The hidden layer learns the combined features of reflectivity, reflectance phase, and TVDI, and automatically adjusts the weights according to the implicit vegetation information to continuously perform self-correction.

[0072] 5. The output layer outputs the last inverted soil moisture value at the current moment , and based on the unique delayed feedback mechanism of the NARX neural network after parameter optimization, the predicted value of the previous L moments It enters the input layer through direct or delayed feedback and participates in the calculation of subsequent moments.

[0073] 6. When the gradient change is less than the threshold, stop training.

[0074] For example, we selected common farmland type data (winter wheat), used 80% of the input data as training data and 20% as input data, compared the output results of the in-situ soil moisture meter, and selected three time periods for inversion experiments. The inversion results are as follows: Figure 9-11 shown.

[0075] In summary, the present application weakens the impact of vegetation by introducing the temperature drought vegetation index model (TVDI). Because soil moisture generally changes continuously over time, the humidity value at the previous moment is highly correlated with the current moment and can be regarded as a time series. Therefore, the present application adopts NARX neural network to construct the inversion model, so that delay and feedback mechanisms are added in the whole inversion process to enhance the memory ability of historical data, and has higher spatiotemporal resolution than traditional methods. Moreover, through a series of data processing methods, the influence of noise can be weakened, and the accuracy of soil moisture inversion can be further improved.

[0076] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Fig.12 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store GNSS-R soil moisture inversion data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a GNSS-R soil moisture inversion method is implemented.

[0077] Those skilled in the art will understand that Fig.12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0078] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0079] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0080] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0081] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0082] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0083] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A GNSS-R soil moisture inversion method, characterized in that: The GNSS-R soil moisture inversion method includes: The soil reflectivity and reflected signal phase of the study area are determined based on the global navigation satellite system reflected signal, and the normalized vegetation index is obtained; the global navigation satellite system reflected signal includes: GNSS scattered signal, direct signal complex correlation value, GNSS signal wave number, satellite altitude angle, direct signal receiving antenna gain in the direction of the mirror reflection point, reflected signal receiving antenna gain in the direction of the mirror reflection point, distance from the mirror reflection point to the receiver, and satellite to direct antenna spacing; Determining a temperature drought vegetation index based on the normalized difference vegetation index; According to the timestamp of the global navigation satellite system reflection signal and the longitude and latitude information of the study area, the soil reflectivity, the reflection signal phase and the temperature drought vegetation index are temporally and spatially matched to obtain input data; Normalizing the input data to obtain normalized data; Constructing a NARX neural network, optimizing parameters of the NARX neural network using a swarm intelligence optimization algorithm, and obtaining an inversion model based on the NARX neural network after parameter optimization; The normalized data is input into the inversion model to obtain a soil moisture inversion result.

2. The GNSS-R soil moisture inversion method according to claim 1, characterized in that: The soil reflectivity and reflected signal phase of the study area are determined based on the GNSS reflected signal, including: Determine a correlation power peak ratio of a direct and reverse signal based on a complex correlation value of the GNSS scattered signal and the direct signal; Determine an antenna gain correction factor based on a gain of a direct signal receiving antenna in the direction of a mirror reflection point and a gain of a reflected signal receiving antenna in the direction of a mirror reflection point; Determine the propagation path compensation factor based on the distance from the mirror reflection point to the receiver and the distance from the satellite to the direct antenna; Determine the soil reflectivity based on the surface root mean square height, the GNSS signal wave number, the satellite altitude angle, the antenna gain correction factor, the propagation path compensation factor, and the correlation power peak ratio of the direct and reverse signals; The carrier and pseudo code of the reflected signal are stripped by using the local carrier and pseudo code generated by the direct signal tracking link to obtain the reflected signal phase.

3. The GNSS-R soil moisture inversion method according to claim 1, characterized in that: The swarm intelligence optimization algorithm is used to optimize the parameters of the NARX neural network, and an inversion model is obtained based on the NARX neural network after the parameter optimization, including: Optimizing the parameters of the NARX neural network using a swarm intelligence optimization algorithm to obtain the NARX neural network after parameter optimization; Get a sample dataset; The sample data set is used to train the NARX neural network with optimized parameters until a set number of training times or a set error threshold is reached, and the trained NARX neural network with optimized parameters is used as the inversion model.

4. The GNSS-R soil moisture inversion method according to claim 3, characterized in that: Get a sample dataset, including: Acquire soil reflectivity historical data, reflection signal phase historical data and temperature drought vegetation index historical data, and generate sample input data based on the soil reflectivity historical data, the reflection signal phase historical data and the temperature drought vegetation index historical data; Acquire soil moisture historical data corresponding to the sample input data; Performing time-space matching and normalization processing on the sample input data to obtain normalized sample data; A training sample pair is formed based on the normalized sample data and the soil moisture historical data to obtain the sample data set.

5. The GNSS-R soil moisture inversion method according to claim 4, characterized in that: Generating sample input data based on the soil reflectivity historical data, the reflection signal phase historical data and the temperature drought vegetation index historical data, including: Pre-classifying the soil reflectivity historical data and the reflection signal phase historical data by land type to obtain a pre-classification result; The pre-classification result and the temperature drought vegetation index historical data are used as sample input data.

6. The GNSS-R soil moisture inversion method according to claim 1, characterized in that: Determining a temperature drought vegetation index based on the normalized difference vegetation index includes: Using the dry-wet edge fitting equation, the maximum surface temperature and the minimum surface temperature are determined based on the normalized difference vegetation index; The real surface temperature is obtained, and the temperature drought vegetation index is determined based on the real surface temperature, the maximum surface temperature and the minimum surface temperature.

7. The GNSS-R soil moisture inversion method according to claim 6, characterized in that: The dry-wet edge fitting equation is expressed as: ； In the formula, is the maximum surface temperature, is the lowest surface temperature, are the coefficients of the linear fitting equations of the dry and wet edges, Represents the Normalized Difference Vegetation Index.

8. The GNSS-R soil moisture inversion method according to claim 1, characterized in that: The swarm intelligence optimization algorithm adopted is the particle swarm optimization algorithm.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the GNSS-R soil moisture inversion method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the GNSS-R soil moisture inversion method described in any one of claims 1 to 8 is implemented.

Images