Search

CN-122024929-A - Satellite-borne GNSS-R soil humidity inversion method and system considering earth surface influence factors

CN122024929ACN 122024929 ACN122024929 ACN 122024929ACN-122024929-A

Abstract

The invention relates to a satellite-borne GNSS-R soil humidity inversion method and system considering earth surface influence factors, wherein the method comprises the steps of obtaining CYGNSS L reflection data, SMAP soil moisture data and multisource static earth surface factor data, preprocessing the data, calculating the correlation between the preprocessed CYGNSS L reflection data and the SMAP soil moisture data to obtain space-time correlation, evaluating, analyzing and checking the multisource static earth surface factor data by adopting a Geodetector method, screening out optimal influence factor combinations by combining the space-time correlation, training a soil moisture inversion model according to the optimal influence factor combinations, and obtaining the trained soil moisture inversion model to invert the soil humidity. According to the invention, the interpretation power and the interaction effect of the surface factors are quantitatively screened, so that an inversion model with concise input, high precision and strong adaptability is constructed, and the high-resolution inversion of soil moisture is realized.

Inventors

  • JIN SHUANGGEN
  • ZHANG SHUO

Assignees

  • 河南理工大学

Dates

Publication Date
20260512
Application Date
20260129

Claims (8)

  1. 1. The satellite-borne GNSS-R soil humidity inversion method considering the earth surface influence factors is characterized by comprising the following steps of: Obtaining CYGNSS L1 reflection data, SMAP soil moisture data and multi-source static surface factor data, and carrying out data preprocessing; calculating the correlation between the CYGNSS L reflection data and the SMAP soil moisture data after pretreatment to obtain space-time correlation; evaluating, analyzing and checking the multisource static surface factor data by adopting a Geodetector method, and screening out an optimal influence factor combination by combining with the space-time correlation; and training a soil moisture inversion model according to the optimal influence factor combination, and obtaining the trained soil moisture inversion model to invert the soil humidity, wherein the soil moisture inversion model is constructed by adopting a XGBoost machine learning model, and the optimal influence factor combination is land utilization, reflectivity and vegetation biomass.
  2. 2. The method for performing the inversion of the earth's surface-influencing element-considered satellite-borne GNSS-R soil humidity according to claim 1, wherein the performing of the data preprocessing comprises: Performing quality screening on the CYGNSS L reflection data according to the signal-to-noise ratio, the antenna gain and the incident angle, and calculating the surface reflectivity according to the bistatic radar sectional area model; Unifying the multisource static surface factor data on the pixel scale according to unified projection, spatial resolution and geographic range; And carrying out long-term normalization processing on the earth surface reflectivity and SMAP soil moisture data based on the time sequence statistical characteristics, and introducing logarithmic transformation and deviation enhancement to the earth surface reflectivity.
  3. 3. The method of inversion of earth surface-influencing element-considered satellite-borne GNSS-R soil moisture according to claim 1, wherein calculating the correlation between the CYGNSS L reflection data and SMAP soil moisture data after preprocessing comprises: Calculating the pixel-by-pixel Pearson correlation coefficient of CYGNSS L reflection data and SMAP soil moisture data according to a month to obtain a correlation distribution map; And judging the effectiveness of the reflected signals in different areas and seasons according to the correlation distribution diagram, and acquiring the space-time correlation.
  4. 4. The method for performing satellite-borne GNSS-R soil moisture inversion with consideration of surface influencing factors according to claim 1, wherein evaluating, analyzing and verifying the multi-source static surface factor data using the Geodetector method comprises: carrying out single-factor interpretation effort quantitative evaluation on each factor in the multi-source static surface factor data; performing double-factor interaction effect analysis on factors in the multi-source static surface factor data, judging whether enhancement or nonlinear interaction exists among different factors, and obtaining a significance test result; and screening the optimal influence factor combination according to the interpretation effort and the significance test result.
  5. 5. A satellite-borne GNSS-R soil moisture inversion system taking into account earth's surface influencing elements for implementing the satellite-borne GNSS-R soil moisture inversion method taking into account earth's surface influencing elements as claimed in any of claims 1 to 4, comprising: The data acquisition module is used for acquiring CYGNSS L reflection data, SMAP soil moisture data and multi-source static surface factor data; the preprocessing module is used for preprocessing the acquired data; The factor screening module is used for calculating the correlation between the CYGNSS L reflection data after pretreatment and the SMAP soil moisture data to obtain space-time correlation, evaluating, analyzing and checking the multi-source static surface factor data by adopting a Geodetector method, and screening out an optimal influence factor combination by combining the space-time correlation; And the inversion model module is used for training a soil moisture inversion model according to the optimal influence factor combination, and obtaining the trained soil moisture inversion model to invert the soil humidity, wherein the soil moisture inversion model is constructed by adopting a XGBoost machine learning model, and the optimal influence factor combination is land utilization, reflectivity and vegetation biomass.
  6. 6. The earth-surface-influencing-element-considered satellite-borne GNSS-R soil moisture inversion system of claim 5 wherein performing data preprocessing comprises: Performing quality screening on the CYGNSS L reflection data according to the signal-to-noise ratio, the antenna gain and the incident angle, and calculating the surface reflectivity according to the bistatic radar sectional area model; Unifying the multisource static surface factor data on the pixel scale according to unified projection, spatial resolution and geographic range; And carrying out long-term normalization processing on the earth surface reflectivity and SMAP soil moisture data based on the time sequence statistical characteristics, and introducing logarithmic transformation and deviation enhancement to the earth surface reflectivity.
  7. 7. The earth-surface-influencing-element-considered satellite-borne GNSS-R soil moisture inversion system of claim 5 wherein calculating the correlation between the CYGNSS L reflection data and SMAP soil moisture data after preprocessing comprises: Calculating the pixel-by-pixel Pearson correlation coefficient of CYGNSS L reflection data and SMAP soil moisture data according to a month to obtain a correlation distribution map; And judging the effectiveness of the reflected signals in different areas and seasons according to the correlation distribution diagram, and acquiring the space-time correlation.
  8. 8. The earth-surface-influencing-element-considered satellite-borne GNSS-R soil moisture inversion system of claim 5 wherein evaluating, analyzing and verifying the multi-source static earth surface factor data using the Geodetector method comprises: carrying out single-factor interpretation effort quantitative evaluation on each factor in the multi-source static surface factor data; performing double-factor interaction effect analysis on factors in the multi-source static surface factor data, judging whether enhancement or nonlinear interaction exists among different factors, and obtaining a significance test result; and screening the optimal influence factor combination according to the interpretation effort and the significance test result.

Description

Satellite-borne GNSS-R soil humidity inversion method and system considering earth surface influence factors Technical Field The invention relates to the technical field of soil moisture inversion, in particular to a satellite-borne GNSS-R soil moisture inversion method and system considering earth surface influence factors. Background Soil moisture is a key parameter in land hydrologic cycle, and has important influence on agricultural production management, drought monitoring, ecological system dynamic evaluation and climate model simulation. The soil moisture obtaining method at the present stage mainly comprises three modes of ground actual measurement, optical remote sensing and microwave remote sensing. The ground actual measurement method relies on sampling points or a sensor network, and has the advantage of high precision, but is limited in space coverage, insufficient in arrangement density and other factors, so that continuous monitoring of soil moisture in an area or global scale is difficult to realize. The optical remote sensing method generally utilizes vegetation indexes, earth surface reflectivity and the like to indirectly calculate soil moisture, but is highly sensitive to cloud cover, solar altitude angle and earth surface vegetation coverage, has limited application in a high vegetation coverage area or a long-term cloud and fog area, and has weak inversion stability. The microwave active or passive remote sensing technology has the advantage of penetrating cloud layers and being observable day and night, wherein satellite data such as SMAP (surface mounted technology) are used as representatives, and the technology is widely used for global soil moisture monitoring. However, its spatial resolution is typically on the order of 36km, which is difficult to meet for medium and high resolution data requirements for agricultural management, hydrologic models, and fine earth surface studies. In addition, the repeated observation period is long, and the rapid dynamic change of the soil moisture is difficult to capture. The GNSS-R (global navigation satellite system reflection) technology developed in recent years can provide higher space-time resolution with less influence of weather conditions by utilizing the passive receiving characteristics of the navigation satellite signals after surface scattering. The CYGNSS satellite system consists of 8 microwave receiving satellites, can realize repeated daily observation, and has great potential in the field of soil moisture inversion. However, the existing CYGNSS-based soil moisture inversion method still has the following problems: (1) Usually, more external auxiliary data such as vegetation parameters, soil types, earth surface structures and the like are relied on, so that a model is complex and parameters are redundant; (2) Lack of systematic quantitative analysis of how each surface factor affects CYGNSS the relationship of reflected signals to soil moisture; (3) Whether interaction enhancement or nonlinear coupling effect exists between factors has not been fully studied, and the model interpretability is affected; (4) In areas with strong surface heterogeneity (such as hills, farmland and forest land staggered distribution areas), inversion accuracy is unstable and reliability is insufficient. Therefore, the prior art is not provided with a CYGNSS soil moisture inversion method which can effectively screen earth surface influence factors, has a simple structure and is high in precision. An inversion model based on key factors is urgently needed to be constructed, the complexity of the model is reduced, and meanwhile inversion precision and space adaptability are improved, so that the requirements of refined hydrology and agricultural monitoring are met. Disclosure of Invention The invention aims to provide a satellite-borne GNSS-R soil humidity inversion method and system considering earth surface influence factors, and aims to solve the problems of excessive auxiliary factors, high model complexity, ambiguous earth surface factor influence mechanism, unstable inversion precision in an earth surface heterogeneous region and the like in the existing earth moisture inversion process based on CYGNSS satellite reflection signals. In order to achieve the above object, the present invention provides the following solutions: A satellite-borne GNSS-R soil humidity inversion method considering earth surface influence factors comprises the following steps: Obtaining CYGNSS L1 reflection data, SMAP soil moisture data and multi-source static surface factor data, and carrying out data preprocessing; calculating the correlation between the CYGNSS L reflection data and the SMAP soil moisture data after pretreatment to obtain space-time correlation; evaluating, analyzing and checking the multisource static surface factor data by adopting a Geodetector method, and screening out an optimal influence factor combination by combining with the space-time correla