GNSS-R Soil Moisture Inversion Method in Mountainous Areas Incorporating Terrain Humidity Index

Abstract

The present invention discloses a GNSS-R soil moisture inversion method for mountainous areas that integrates a terrain moisture index. The method comprises the following steps: obtaining remote sensing data from target mountainous areas other than water bodies and soil moisture data from each observation station; selecting altitude, slope, aspect, longitude and latitude, vegetation index, clay content, and signal-to-noise ratio as influencing factors for soil moisture inversion in mountainous areas; calculating the corrected signal-to-noise ratio and terrain moisture index; constructing a mountainous area soil moisture inversion model, training it, and validating it; inputting a test set into the trained model to obtain daily soil moisture prediction data with a 1km resolution, taking the monthly average, and obtaining monthly soil moisture prediction data. The present invention's integration of the terrain moisture index with CYGNSS data can fully leverage the advantages of terrain factors and remote sensing technology, further improving the accuracy of soil moisture distribution and the ability to predict spatiotemporal changes, and better revealing the dynamic changes in soil moisture in complex terrain areas such as mountainous areas.

Description

Technical Field

[0001] The present invention relates to a GNSS-R soil moisture inversion method in mountainous areas integrated with a terrain moisture index, and belongs to the technical field of surface soil moisture inversion. Background Art

[0002] Soil moisture, a key variable describing the exchange of water and energy between the Earth's surface and the atmosphere, is a crucial parameter influencing hydrological, ecological, and biogeochemical processes. It is an essential factor influencing research in areas such as climate prediction, water resources management, agrometeorology, and flood disasters. Traditional soil moisture monitoring methods are limited in time and space, and the acquisition process and data are cumbersome and complex. In contrast, specialized remote sensing satellites offer greater comprehensiveness, timeliness, and regional continuity. Advances in satellite navigation technology have led to the availability of over a hundred satellites in the Global Navigation Satellite System (GNSS), represented by the United States' Global Positioning System (GPS), Russia's GLONASS, the European Union's GALILEO, and China's BeiDou Navigation Satellite System. This has given rise to the spaceborne Global Navigation Satellite System Reflectometry (GNSS-R), which uses the reception and processing of signals reflected from the Earth's surface to determine the status of a target. Spaceborne GNSS-R, a recently emerging microwave remote sensing technology, utilizes receivers placed on low-orbit satellites for observation. It features passive detection, low power consumption, low deployment costs, and a wide range. Its revisit period can be as low as several hours, and its temporal resolution is high. Furthermore, GNSS signals operate in the L-band, offering excellent penetration and enabling all-day, all-weather observations. Among these, the Cyclone Global Navigation Satellite System (CYGNSS), launched by the United States in 2016 and a member of the GNSS-R system, is a particularly popular research topic in the field of soil moisture.

[0003] Currently, there are still many problems and challenges in the practical application of the algorithm for inverting soil moisture using CYGNSS data:

[0004] (1) Most mountainous areas are rain-fed agricultural areas, and water resources transportation is difficult. Agricultural development in mountainous areas urgently needs timely and accurate soil moisture data. The effective acquisition of soil moisture in mountainous areas will provide a more reliable scientific decision-making basis for water resource management, drought warning, and agricultural irrigation. However, the existing GNSS-R soil moisture inversion algorithm is mainly designed for large areas and lacks optimization for the specific terrain conditions in mountainous areas, resulting in the application effect of the constructed algorithm model in mountainous areas being weakened compared to the overall effect. Most current algorithm studies only consider conventional terrain elements such as elevation and slope, and simply use them as input features of machine learning models. This cannot effectively solve the problem of GNSS-R soil moisture inversion in mountainous areas.

[0005] (2) In the field of soil moisture research, the Topographic Wetness Index (TWI) is widely used to characterize the spatial variation characteristics of soil moisture distribution in mountainous areas. TWI takes into account topographic factors such as slope, aspect, and hydrological processes in the watershed, and can therefore effectively reflect the impact of topography on soil moisture. Especially in mountainous areas, where the terrain is complex and the spatial distribution of soil moisture often has significant non-uniformity, TWI provides an important quantitative tool for analyzing and predicting soil moisture in these areas. However, current research simply uses the Topographic Wetness Index as an influencing factor input for machine learning, and has not truly achieved integration with GNSS-R data. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a GNSS-R soil moisture inversion method in mountainous areas that integrates terrain moisture index, thereby effectively improving the inversion accuracy of spaceborne GNSS-R in mountainous areas.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] The GNSS-R soil moisture inversion method in mountainous areas integrated with the terrain moisture index includes the following steps:

[0009] Step 1: Acquire remote sensing data from areas other than water bodies in the target mountainous area, including Cyclone Global Navigation Satellite System data, digital elevation model data, medium-resolution imaging spectrometer observation data, and SoilGrids world soil information data; simultaneously acquire soil moisture data from each observation station in the target mountainous area; and preprocess the acquired remote sensing data and soil moisture data.

[0010] Step 2: Select altitude, slope, aspect, longitude and latitude, vegetation index, clay content, and signal-to-noise ratio as influencing factors for GNSS-R soil moisture inversion in mountainous areas. Obtain altitude, slope, aspect, and longitude and latitude data from preprocessed digital elevation model data. Calculate vegetation index using preprocessed medium-resolution imaging spectrometer observation data. Obtain clay content data from preprocessed SoilGrids world soil information data.

[0011] Step 3: Correct the signal-to-noise ratio (SNR) of the pre-processed Cyclone GNSS data and calculate the terrain wetness index (TWI) using the pre-processed digital elevation model (DEM) data. The altitude, slope, aspect, latitude and longitude, vegetation index, clay content, corrected SNR, TWI, and pre-processed soil moisture data from the observation station are spatiotemporally matched to daily data at a 1 km resolution grid.

[0012] Step 4: Obtain daily data on the signal-to-noise ratio and terrain humidity index corresponding to 1 km resolution grids whose signal-to-noise ratio is not null, use the signal-to-noise ratio as the dependent variable and the terrain humidity index as the independent variable, establish a daily spatial fitting function model, and use the daily spatial fitting function model to spatialize the signal-to-noise ratio of the 1 km resolution grids whose signal-to-noise ratio is null on that day to obtain the signal-to-noise ratio of all 1 km resolution grids on that day;

[0013] Step 5: The altitude, slope, aspect, longitude and latitude, vegetation index, clay content, corrected signal-to-noise ratio, and pre-processed daily data of soil moisture at the observation station corresponding to all 1 km resolution grids are used as a data set. The data set is sorted from front to back in chronological order. Starting from the front of the data set, the daily data corresponding to the 1 km resolution grids with the corrected signal-to-noise ratio and non-null soil moisture in a period of time are selected as the training set and validation set. All daily data corresponding to the remaining time are used as the test set. The test set includes the daily data corresponding to all 1 km resolution grids with null signal-to-noise ratio and non-null signal-to-noise ratio. The signal-to-noise ratio of the 1 km resolution grid with null signal-to-noise ratio is obtained by spatialization in step 4.

[0014] Step 6: Construct a GNSS-R soil moisture inversion model for mountainous areas, including a random forest model, an XGBoost model, a support vector machine model, and a stacking ensemble model. Altitude, slope, aspect, longitude and latitude, vegetation index, clay content, and corrected signal-to-noise ratio are all used as inputs to the random forest model, the XGBoost model, and the support vector machine model. The soil moisture prediction results output by the random forest model, the XGBoost model, and the support vector machine model are used as inputs to the stacking ensemble model. The stacking ensemble model assigns weights to the soil moisture prediction results output by the random forest model, the XGBoost model, and the support vector machine model, and outputs the final soil moisture prediction result. The training set and validation set are used to train and validate the GNSS-R soil moisture inversion model for mountainous areas, thereby obtaining a trained GNSS-R soil moisture inversion model for mountainous areas.

[0015] Step 7: Input the test set into the trained mountain GNSS-R soil moisture inversion model to obtain the daily soil moisture prediction data at a 1km resolution grid. The daily soil moisture prediction data of each month are averaged to obtain the monthly soil moisture prediction data at a 1km resolution grid.

[0016] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0017] This paper, for the first time, focuses on algorithm design for complex mountainous areas, proposing an inversion algorithm that integrates the Terrain Moisture Index (TMI) and GNSS-R data. The TMI reflects the influence of topography on the uneven distribution of soil moisture under the influence of gravity. Fusion of the TMI with CYGNSS data leverages topographic factors and remote sensing technology to further improve the accuracy of soil moisture distribution and the ability to predict spatiotemporal changes, thereby better revealing the dynamic changes in soil moisture in complex terrain, such as mountainous areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a technical flow chart of the present invention;

[0019] Figure 2 It is a technical flow chart of the specific processing involved in the present invention;

[0020] Figure 3 Schematic diagram of the distribution of observation sites involved in the embodiment of the present invention;

[0021] Figure 4 This is a schematic diagram of the impact factor distribution;

[0022] Figure 5 This is a schematic diagram of the relationship between CYGNSS SNR and soil moisture;

[0023] Figure 6 It is a schematic diagram of the distribution of topographic humidity index;

[0024] Figure 7 This is a flow chart of the method for spatializing CYGNSS SNR data using the terrain wetness index;

[0025] Figure 8 This is a schematic diagram of the model inversion results of soil moisture in the southern mountainous area of Anhui in January, April, July and October 2022 involved in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be interpreted as limiting the present invention.

[0027] like Figure 1 and Figure 2 As shown, the present invention proposes a GNSS-R soil moisture integrated inversion method in mountainous areas that integrates terrain moisture index, including the following steps:

[0028] Step S1, obtaining Cyclone Global Navigation Satellite System (CYGNSS) data, digital elevation model (DEM) data, medium-resolution imaging spectrometer observation data, and SoilGrids world soil information data in the target mountain area; and simultaneously obtaining soil moisture data at each observation station in the target mountain area;

[0029] CYGNSS satellite data comes from Level 1 V3.0 data provided by NASA; DEM data comes from the Shuttle Radar Topography Mission (SRTM); and SoilGrids world soil information data comes from SoilGrids data developed by the International Soil Reference and Information Centre (ISRIC).

[0030] The following will take the southern mountainous area of Anhui as an example, including the seven cities of Huangshan, Wuhu, Ma'anshan, Tongling, Xuancheng, Chizhou and Anqing, and select data from 2021 and 2022 to express the technical solution in this invention in detail and completely.

[0031] In step S2, four factors that have a profound impact on CYGNSS soil moisture inversion, namely vegetation, terrain, water bodies, and soil properties, are selected as parameters for model construction, as shown in Table 1 and Figure 4 As shown, vegetation factors include vegetation index (NDVI), terrain factors include altitude, slope, aspect and longitude and latitude, soil properties include clay content, and water factors include monthly water data; monthly water data is only used to eliminate water areas in the target area.

[0032] Table 1

[0033]

[0034] Preprocess the acquired data:

[0035] 2.1, Conduct quality control on measured soil moisture data. Screen soil moisture data from each observation station, remove missing values and outliers, and ensure data accuracy. The distribution of observation stations is as follows: Figure 3 shown.

[0036] 2.2. Quality control of the acquired CYGNSS data: Observation data with an incidence angle greater than 65° were eliminated to preliminarily ensure that the error caused by terrain height was within a reasonable range; data with a signal-to-noise ratio less than 2dB or greater than 14dB; a receiver gain Gr less than 0dB; and a reflected signal peak power greater than -147dB were eliminated.

[0037] 2.3. Calculate the vegetation index using the preprocessed MOD09Q1 dataset. The preprocessing steps include format conversion, projection conversion, and data resampling. The specific formula for calculating the vegetation index is as follows:

[0038] ,

[0039] Where, is the reflectivity of the red light band from 620nm to 670nm, It is the reflectivity of the near-infrared band of 841~875nm; NDVI is the vegetation index, and its positive value indicates the surface vegetation coverage, that is, the greater the coverage, the greater the NDVI value.

[0040] 2.4, deal with the noise and outliers in the DEM data, and use the "Slope" calculation tool in ArcGIS software to calculate the slope based on the DEM data. The formula is as follows:

[0041] ,

[0042] in, is the slope, is the elevation value on the terrain surface, and Indicates the change in spatial position in the horizontal direction.

[0043] In ArcGIS, we calculated aspect using the Aspect tool within Spatial Analyst. Aspect values are expressed as a range of 0 to 360 degrees, with each grid cell's primary orientation calculated clockwise, with true north at 0 degrees. We then used the Reclassify tool to reclassify the aspect values into eight directions.

[0044] DEM is operated through ArcGIS software to obtain longitude and latitude data, specifically: DEM extracts points, then uses points to extract the xy coordinates of each point, and then exports each xy to a raster.

[0045] 2.5, Global gridded soil information (SoilGrids) is used to obtain clay content. Since the penetration depth of L-band signals is usually from the surface to 5 cm, only the surface layer (0-5 cm) data in SoilGrids is used.

[0046] Step S3, calculating the signal-to-noise ratio (SNR) data using the CYGNSS data, and calculating the terrain wetness index using the elevation data;

[0047] 3.1. Using CYGNSS data to correct the signal-to-noise ratio (SNR) data. SNR can be used to characterize the strength of the reflected signal. When soil moisture increases, the soil dielectric constant increases, which in turn enhances signal reflection and causes a change in SNR. The relationship diagram is shown in the figure below. Figure 5 shown.

[0048] The calculation formula of the corrected signal-to-noise ratio is as follows:

[0049] ,

[0050] In the formula, SNR represents the corrected signal-to-noise ratio data; is the transmit power of the Global Navigation Satellite System (GNSS) signal; is the gain of the transmitting antenna; is the gain of the receiving antenna; is the carrier wavelength of the GNSS signal; is the distance between the sampling point and the GNSS transmitter; is the distance between the sampling point and the Cyclone Global Navigation Satellite System (CYGNSS) receiver; DDM_SNR is the raw signal-to-noise ratio data provided by CYGNSS.

[0051] 3.2, calculate the terrain moisture index. It can reflect the impact of terrain on the uneven distribution of soil moisture under the action of gravity. The calculation formula is as follows:

[0052] ,

[0053] Where, is the unit catchment area, is the slope of the grid cell. It is usually assumed that the effective contour length is the same as the length of the unit grid cell. The slope is the maximum slope direction in the downhill direction. The relevant definitions are as follows:

[0054] ,

[0055] ,

[0056] Where, represents the total upstream area entering the grid cell; is the effective contour length, that is, the length of the unit grid; Represents the height difference between adjacent grid cells; Represents the length between the centers of adjacent grid cells. The flow direction of each grid can be calculated by using the flow direction tool in ArcGIS software, and then calculated using the flow accumulation tool. Figure 6 shown.

[0057] 3.3. Temporally and spatially match all data. All data (altitude, slope, aspect, latitude and longitude, vegetation index, clay content, and terrain wetness index) were resampled to a 1 km resolution using the nearest neighbor method. The signal-to-noise ratio data for CYGNSS sampling points within the same grid on the same day (within 24 hours) were averaged to represent the signal-to-noise ratio data for the entire grid. Grids without sampling points had a null value for the signal-to-noise ratio. Soil moisture data collected at a station represented the soil moisture data for the grid where the station was located. Grids without a station had a null value for the soil moisture.

[0058] Step S4, establishing a spatial fitting function between the terrain wetness index and the daily corrected SNR data, and spatializing the corrected SNR data;

[0059] Using only the 1km grid where the actual sampling points have valid values, with the terrain humidity index as the independent variable and the pre-processed 1km resolution daily CYGNSS SNR data as the dependent variable, the most appropriate functional relationship is selected to establish a spatial fitting function model of daily updated changes:

[0060] ,

[0061] In this embodiment, linear fitting is preferred as the spatial fitting function model:

[0062] ,

[0063] The coefficient of determination needs to be calculated for the fitting results Value, right The value sets the threshold value. Greater than the threshold, it means the function relationship is valid. When the value is less than the threshold, the data of the day is not retained to ensure the validity of the relationship between the SNR data and the terrain humidity index data distribution. Using the established spatial fitting function model, the daily SNR data with reasonable fitting results can be spatialized using the known terrain humidity index data, and finally the 1km daily SNR data distribution is obtained. The detailed process of the spatialization part is as follows Figure 7 shown.

[0064] In step S5, the spatialized SNR data is combined with other data (altitude, slope, aspect, longitude and latitude, vegetation index, clay content, and preprocessed daily soil moisture data) as a dataset. The dataset is then divided based on the time series. The data corresponding to a portion of dates are used as training and validation sets, selecting only 1km-resolution grids where the signal-to-noise ratio is the actual corrected value and the soil moisture is not null during this period. The data corresponding to another portion of dates are used as a test set, selecting all data corresponding to the grids corresponding to the spatialized SNR. In this example, the 2021 data is randomly divided into training and validation sets at a ratio of 80% and 20% for optimal model construction. The 2022 data is used as the test set to test the model's effectiveness.

[0065] In step S6, the training set and the validation set are used to perform integrated learning on the three machine learning algorithms of support vector machine (SVM), XGBoost, and random forest (RF) through the Stacking integrated model to obtain the optimal integrated inversion model.

[0066] RF, XGBoost, and SVM models were selected as base learners for the Stacking ensemble method. RF can handle high-dimensional, nonlinear data, providing high prediction accuracy, strong noise immunity, and automatic feature selection capabilities. XGBoost excels at processing structured data and can optimize performance through gradient boosting, making it particularly suitable for nonlinear problems. SVM is particularly well-suited to high-dimensional data, using kernel techniques to find the optimal segmentation hyperplane in high-dimensional space.

[0067] The outputs of the RF, XGBoost, and SVM models are used as the input of the Stacking ensemble model, and the final result of the site soil moisture prediction is used as the output of the Stacking ensemble model. That is, the training set is used as the input item to be input into the three base models respectively, and then the output items of the three base models are used as the input of the Stacking ensemble model to train the model.

[0068] The stacking ensemble model sets weights for the outputs of the RF, XGBoost, and SVM models. It then verifies the predictions under different weight combinations using methods such as 10-fold cross-validation, and uses the best verified result as the optimal result. This example uses an even distribution of weights, i.e., 1 / 3, 1 / 3, 1 / 3.

[0069] The present invention adopts (Root mean square error), (correlation coefficient) is used to characterize the accuracy, and the calculation formula is as follows:

[0070] ,

[0071] ,

[0072] in, is the number of samples, For the The soil moisture observation value of the corresponding station of the sample, For the model The soil moisture value predicted by the sample, is the average value of all sample predictions, is the average of all sample actual values.

[0073] Set when Greater than 0.7 and When it is less than 0.1, the trained model is output. The final verification result obtained in this embodiment is: 、 .

[0074] In step S7, the trained model is inputted using the test set, and the simulated daily values are averaged monthly to obtain the final soil moisture 1km monthly scale result.

[0075] This example uses the 2022 CYGNSS data and auxiliary data as model input. After taking the monthly average of the simulated daily values, accuracy tests are performed on a monthly scale using the SMAP remote sensing soil moisture data and station data as the true values.

[0076] The 1km soil moisture data inverted by the model was resampled to 9km, which is consistent with the SMAP data. The accuracy was calculated to be , There is good consistency between the soil moisture data inverted by the model and the site. Correlation coefficient reached 0.93, for .

[0077] The model simulates the soil moisture distribution in January, April, July and October 2022 as follows: Figure 8 shown.

[0078] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the aforementioned GNSS-R soil moisture inversion method for mountainous areas integrating the terrain moisture index are implemented.

[0079] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the aforementioned GNSS-R soil moisture inversion method in mountainous areas that integrates terrain moisture index.

[0080] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0081] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0082] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0083] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0084] The above embodiments are only for illustrating the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A GNSS-R soil moisture inversion method in mountainous areas that integrates terrain moisture index is characterized by: The steps include: Step 1: Acquire remote sensing data from areas other than water bodies in the target mountainous area, including Cyclone Global Navigation Satellite System data, digital elevation model data, medium-resolution imaging spectrometer observation data, and SoilGrids world soil information data; simultaneously acquire soil moisture data from each observation station in the target mountainous area; and preprocess the acquired remote sensing data and soil moisture data. Step 2: Select altitude, slope, aspect, longitude and latitude, vegetation index, clay content, and signal-to-noise ratio as influencing factors for GNSS-R soil moisture inversion in mountainous areas. Obtain altitude, slope, aspect, and longitude and latitude data from preprocessed digital elevation model data. Calculate vegetation index using preprocessed medium-resolution imaging spectrometer observation data. Obtain clay content data from preprocessed SoilGrids world soil information data. Step 3: Correct the signal-to-noise ratio (SNR) of the pre-processed Cyclone GNSS data and calculate the terrain wetness index (TWI) using the pre-processed digital elevation model (DEM) data. The altitude, slope, aspect, latitude and longitude, vegetation index, clay content, corrected SNR, TWI, and pre-processed soil moisture data from the observation station are spatiotemporally matched to daily data at a 1 km resolution grid. Step 4: Obtain daily data on the signal-to-noise ratio and terrain humidity index corresponding to 1 km resolution grids whose signal-to-noise ratio is not null, use the signal-to-noise ratio as the dependent variable and the terrain humidity index as the independent variable, establish a daily spatial fitting function model, and use the daily spatial fitting function model to spatialize the signal-to-noise ratio of the 1 km resolution grids whose signal-to-noise ratio is null on that day to obtain the signal-to-noise ratio of all 1 km resolution grids on that day; Step 5: The altitude, slope, aspect, longitude and latitude, vegetation index, clay content, corrected signal-to-noise ratio, and pre-processed daily data of soil moisture at the observation station corresponding to all 1 km resolution grids are used as a data set. The data set is sorted from front to back in chronological order. Starting from the front of the data set, the daily data corresponding to the 1 km resolution grids with the corrected signal-to-noise ratio and non-null soil moisture in a period of time are selected as the training set and validation set. All daily data corresponding to the remaining time are used as the test set. The test set includes the daily data corresponding to all 1 km resolution grids with null signal-to-noise ratio and non-null signal-to-noise ratio. The signal-to-noise ratio of the 1 km resolution grid with null signal-to-noise ratio is obtained by spatialization in step 4. Step 6: Construct a GNSS-R soil moisture inversion model for mountainous areas, including a random forest model, an XGBoost model, a support vector machine model, and a stacking ensemble model. Altitude, slope, aspect, longitude and latitude, vegetation index, clay content, and corrected signal-to-noise ratio are all used as inputs to the random forest model, the XGBoost model, and the support vector machine model. The soil moisture prediction results output by the random forest model, the XGBoost model, and the support vector machine model are used as inputs to the stacking ensemble model. The stacking ensemble model assigns weights to the soil moisture prediction results output by the random forest model, the XGBoost model, and the support vector machine model, and outputs the final soil moisture prediction result. The training set and validation set are used to train and validate the GNSS-R soil moisture inversion model for mountainous areas, thereby obtaining a trained GNSS-R soil moisture inversion model for mountainous areas. Step 7: Input the test set into the trained mountain GNSS-R soil moisture inversion model to obtain the daily soil moisture prediction data at a 1km resolution grid. The daily soil moisture prediction data of each month are averaged to obtain the monthly soil moisture prediction data at a 1km resolution grid.

2. The GNSS-R soil moisture inversion method for mountainous areas integrated with terrain moisture index according to claim 1 is characterized in that: In step 1, the acquired remote sensing data and soil moisture data are preprocessed as follows: Preprocessing of Cyclone GNSS received data includes: sequentially eliminating received data with an incident angle greater than 65°, a signal-to-noise ratio less than 2dB or greater than 14dB, a receiver gain less than 0dB, and a reflected signal peak power greater than -147dB; Preprocessing the digital elevation model data includes: removing noise data from the digital elevation model data; Preprocessing of the observation data of the Moderate Resolution Imaging Spectrometer, including format conversion, projection conversion and data resampling; The soil moisture data were preprocessed, including removing missing values and outliers from the soil moisture data observed at each observation station.

3. The GNSS-R soil moisture inversion method for mountainous areas integrated with terrain moisture index according to claim 1 is characterized in that: In step 2, the altitude and longitude and latitude are obtained from the pre-processed digital elevation model data; The slope calculation tool in ArcGIS software was used to calculate the slope. The slope calculation formula is as follows: ， in, is the slope, is the elevation value on the terrain surface, and Indicates the change of spatial position in the horizontal direction; Use the slope calculation tool in ArcGIS software to calculate the slope direction in a clockwise direction, with due north being 0 degrees; The calculation formula of vegetation index is as follows: ， in, is the vegetation index, is the reflectivity in the red light band, is the reflectivity in the near-infrared band; From the preprocessed SoilGrids world soil information data, the surface clay content data corresponding to the L band is selected.

4. The GNSS-R soil moisture inversion method for mountainous areas integrated with terrain moisture index according to claim 1 is characterized in that: In step 3, the calculation formula of the corrected signal-to-noise ratio is as follows: ， in, is the corrected signal-to-noise ratio, is the transmission power of the GNSS signal, is the gain of the transmitting antenna, is the distance between the sampling point and the GNSS transmitter, is the distance between the sampling point and the Cyclone GNSS receiver, is the carrier wavelength of the GNSS signal, is the signal-to-noise ratio in the pre-processed Cyclone GNSS received data; The calculation formula of topographic wetness index is as follows: ， in, is the topographic wetness index, is the unit catchment area, is the slope of the grid cell, and ， ， in, is the total upstream area of the grid cell, which is the grid cell of the digital elevation model data; is the effective contour length, i.e. the length of the unit grid; is the height difference between adjacent grid cells; is the length between the centers of adjacent grid cells.

5. The GNSS-R soil moisture inversion method for mountainous areas integrated with terrain moisture index according to claim 1 is characterized in that: In step 3, the time-space matching is as follows: The altitude, slope, aspect, latitude and longitude, vegetation index, clay content, and terrain moisture index were resampled to daily data at a 1 km resolution grid using the nearest neighbor method; The signal-to-noise ratio data format is converted to a 1km resolution grid. That is, for each grid, the signal-to-noise ratios calculated from all sampling points within the grid within 24 hours are averaged to obtain the daily signal-to-noise ratio at the 1km resolution grid. The daily signal-to-noise ratio of the grid with no sampling points within 24 hours is recorded as a null value. The pre-processed soil moisture data is converted to a 1km resolution grid, that is, the pre-processed soil moisture of each observation station is used as the soil moisture of the grid where each observation station is located; The soil moisture of grids without observation stations is recorded as null value.

6. The GNSS-R soil moisture inversion method for mountainous areas integrated with terrain moisture index according to claim 1 is characterized in that: In step 4, after establishing the spatial fitting function model corresponding to each day, the determination coefficient is calculated for the spatial fitting function model. When the determination coefficient is greater than the preset third threshold, it indicates that the established spatial fitting function model is valid, and the signal-to-noise ratio fitted by the spatial fitting function model established on that day is retained; otherwise, the signal-to-noise ratio fitted by the spatial fitting function model established on that day is deleted.

7. The GNSS-R soil moisture inversion method for mountainous areas integrated with terrain moisture index according to claim 1 is characterized in that: In step 6, when training and validating the GNSS-R soil moisture inversion model for mountainous areas, the accuracy of the inversion model is characterized by the root mean square error and the correlation coefficient. When the root mean square error is less than a preset first threshold and the correlation coefficient is greater than a preset second threshold, the training of the GNSS-R soil moisture inversion model for mountainous areas is completed. The calculation formulas for the root mean square error and correlation coefficient are as follows: ， ， in, is the root mean square error, is the correlation coefficient, is the number of samples, For the inversion model The soil moisture value predicted by the sample, For the The soil moisture observation value of the corresponding station of the sample, is the average value of all sample observations, is the average of all sample predictions.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the computer program, the steps of the GNSS-R soil moisture inversion method for mountainous areas by integrating terrain moisture index are implemented as described in any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the GNSS-R soil moisture inversion method for mountainous areas by integrating terrain moisture index are implemented as described in any one of claims 1 to 7.

Images