Method and system for retrieving vegetation canopy fuel moisture content based on geostationary meteorological satellite

Abstract

The application discloses a vegetation canopy fuel moisture content inversion method and system based on a stationary meteorological satellite, and relates to the technical field of remote sensing inversion. First, stationary meteorological satellite observation data is acquired, a vegetation radiation transfer model is constructed, and a cost function taking a root mean square error as a prototype is constructed. Then, gradient information of the cost function is calculated by using an automatic differentiation tool. Next, fuel moisture content is quickly inverted based on a two-stage numerical iteration process. Finally, a multi-iteration initial value selection strategy is combined to relieve a sick inversion problem of a weak sensitive parameter FMC. The method is simple to operate, only needs to input a remote sensing vegetation index, and can quickly invert the fuel moisture content. Moreover, dynamic changes of geometric parameters such as a solar zenith angle and an azimuth angle can be considered. The method is suitable for FMC inversion based on stationary meteorological satellite multi-temporal and multi-angle observation data, which is of great significance for near real-time forest fire risk management and control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote sensing inversion, and in particular to a vegetation canopy fuel moisture content inversion method and system based on a geostationary meteorological satellite. BACKGROUND

[0002] Wildfire is a complex natural disturbance process in the ecosystem, which not only destroys the ecological environment, releases greenhouse gases, and affects the global climate, but also threatens people's life and property safety. Wildfire risk assessment and early warning, as an important measure for forest and grassland management, is of great significance for effective prevention and control of wildfires.

[0003] Fuel moisture content (FMC) is defined as the ratio of the water content in the vegetation canopy to its dry weight, and is closely related to the occurrence and spread of wildfires. It is a key indicator factor for assessing wildfire risk. Therefore, accurate and timely acquisition of FMC is crucial for wildfire risk assessment and early warning. Satellite remote sensing technology can provide long-time series and high-resolution ground observations, providing an effective way for large-scale and near-real-time FMC monitoring. Currently, FMC inversion usually uses optical remote sensing images obtained by sun-synchronous orbit satellites, including Landsat series satellites, Sentinel-2 series satellites, and MODIS products. Although these optical remote sensing images have high spatial resolution, their time resolution is low, resulting in poor timeliness of the FMC obtained by inversion, which cannot meet the needs of near-real-time wildfire risk assessment and early warning. In contrast, the geostationary meteorological satellite Himawari-8 can observe the earth with a 10-minute revisit period, and can form a wealth of remote sensing observation data, which has great application potential in near-real-time FMC acquisition and wildfire risk assessment and early warning.

[0004] FMC inversion research can use empirical statistics and physical model-based methods. The former is heavily dependent on measured data and is difficult to apply on a large scale, while the latter considers the radiation transfer mechanism of vegetation and has high universality. However, there are few studies on FMC inversion based on stationary satellite data. How to invert FMC based on the multi-temporal and multi-angle observation data of geostationary meteorological satellites to achieve near-real-time FMC acquisition is a key to further improve the timeliness of wildfire risk assessment and early warning. SUMMARY

[0005] Therefore, the present application aims to provide a vegetation canopy fuel moisture content inversion method based on a geostationary meteorological satellite, which is used for near-real-time estimation of vegetation canopy fuel moisture content.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] The application provides a vegetation canopy fuel moisture content inversion method based on a stationary meteorological satellite, and comprises the following steps:

[0008] Step 1: obtaining stationary meteorological satellite observation data to generate surface reflectivity data for vegetation index calculation;

[0009] Step 2: constructing a vegetation radiation transfer model; constructing a cost function T_RMSE about normalized difference vegetation index NDVI, enhanced vegetation index EVI and normalized difference infrared index NDII;

[0010] Step 3: obtaining the partial derivative of the cost function T_RMSE about the free variable, which is used for calculating gradient information;

[0011] Step 4: using the reduced gradient algorithm to solve FMC in two stages, in the first stage, sensitive parameters such as moisture thickness EWT, leaf area index LAI and soil factor psoil are inverted, and in the second stage, weak sensitive parameters fuel moisture content FMC and chlorophyll a and b content C ab are inverted;

[0012] Step 5: all fuel moisture content FMC results solved by using the two-stage reduced gradient algorithm under different initial value combinations.

[0013] Further, the normalized difference vegetation index NDVI, the enhanced vegetation index EVI, the normalized difference infrared index NDII and the cost function T_RMSE thereof are calculated according to the following formula:

[0014]

[0015]

[0016]

[0017]

[0018] Wherein, ρ Blue , ρ Red , ρ NIR , ρ SWIR2 are spectral reflectivity of blue light, red light, near-infrared and short-wave infrared bands respectively; NDVI o , EVI o , NDII o are observation vegetation indexes extracted from satellite data; NDVI m , EVI m , NDII m are vegetation indexes simulated by the PROSAIL model, and are also about the free variable C ab, EWT, FMC, LAI and psoil, T_RMSE represents the cost function.

[0019] Further, the partial derivative of the free variable in step 3 is calculated by using the automatic differentiation tool TAPENADE.

[0020] Further, the FMC in step 4 is solved by using the homotopy gradient algorithm in two stages of iteration, and the specific steps are as follows:

[0021] Step 4.1: Set the initial values of the sensitive parameters EWT, LAI and psoil;

[0022] Step 4.2: Solve the sensitive parameters EWT, LAI and psoil by using the homotopy gradient algorithm, and when the cost function T_RMSE1 meets formula (5), stop the first stage of iteration, and output EWT, LAI and psoil, as well as FMC and C ab , which are used as the initial value input for the second stage of numerical iteration;

[0023]

[0024] Wherein, the cost function T_RMSE1 is based on formula (4), and the parameters FMC and C ab are fixed, and only EWT, LAI and psoil are taken as free variables; ∇T_RMSE1 is the gradient of T_RMSE1, which is calculated in step 3; ε1 is the allowable error limit;

[0025] Step 4.3: Introduce the ecological rules based on FMC and LAI and EWT, and dynamically update the FMC value range through formula (6);

[0026]

[0027] Wherein, g1(LAI) is the linear fitting equation of FMC with respect to LAI, and RMSE FMC-LAI is the root mean square error of the fitting equation; g2(EWT) is the linear fitting equation of FMC with respect to EWT, and RMSE FMC-EWT is the root mean square error of the fitting equation;

[0028] Step 4.4: Solve the weak sensitive parameters FMC and C ab by using the homotopy gradient algorithm; when the cost function T_RMSE1 does not meet formula (5), go to step 4.2; when the cost function T_RMSE2 meets formula (11), stop the second stage of iteration, and output the iteration results of FMC and C ab ;

[0029]

[0030] wherein the cost function T_RMSE2 is modeled after equation (4) with fixed parameters EWT, LAI and psoil, and only FMC and C ab as a free variable; ∇T_RMSE2 is the gradient of T_RMSE2, calculated from step 3; and ε2 is the allowed error limit.

[0031] Further, all the results of the combustible moisture content FMC solved by the two-stage reduced gradient algorithm in step 5 for different initial value combinations are further solved by the following steps:

[0032] According to the primary and secondary orders of the cost function values T_RMSE and the modulus |∇T_RMSE| of the gradients thereof, the m groups of the combustible moisture content FMC inversion results with the minimum cost function are first selected, and then the n groups of the combustible moisture content FMC inversion results with the minimum modulus of the gradients are selected from the m groups, and the mean value of the n groups of the combustible moisture content FMC is taken as the final inversion result.

[0033] Further, the value range of ε1 is 0.1-1; or the value range of ε2 is 10 -6 -10 -5 ; or the value of m is 15; or the value of n is 9.

[0034] Further, the satellite observation data in step 1 include the standard albedo product and the geometric parameter products of the solar zenith angle and azimuth angle, the satellite zenith angle and azimuth angle in the full-disk observation mode; and the cloud mask and the atmospheric correction of the data are completed.

[0035] Further, the vegetation radiation transfer model in step 2 is performed in the following manner:

[0036] The chlorophyll a and b content C ab , the equivalent water thickness EWT, the combustible moisture content FMC, the leaf area index LAI and the soil factor psoil are selected as the free variables of the PROSAIL vegetation radiation transfer model, and the other parameters of the model are constants.

[0037] The vegetation canopy combustible moisture content inversion system based on the stationary meteorological satellite provided by the application comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, and the processor implements the above method when executing the program.

[0038] The application has the following beneficial effects:

[0039] The application provides a combustible moisture content inversion method based on a stationary meteorological satellite, and relates to the technical field of remote sensing inversion. The application is aimed at Himawari-8 data of the stationary meteorological satellite, and is based on a PROSAIL vegetation radiation transmission model; first, a cost function taking a root mean square error as a prototype is constructed by using a vegetation index; then, gradient information of the cost function is calculated by using an automatic differentiation tool; next, combustible moisture content is quickly inverted based on a two-stage numerical iteration process; finally, a multi-iteration initial value selection strategy is combined to relieve the ill-posed inversion problem of a weak sensitive parameter FMC.

[0040] The method is simple to operate, can quickly invert the combustible moisture content by only inputting a remote sensing vegetation index, and can consider dynamic changes of geometric parameters such as a solar zenith angle and an azimuth angle, and is suitable for FMC inversion based on stationary meteorological satellite multi-temporal and multi-angle observation data, which is of great significance for near-real-time forest fire risk management and control.

[0041] The method is simple to operate, can quickly estimate the combustible moisture content based on a two-stage iteration inversion process by only inputting vegetation index and geometric parameter information observed by the stationary meteorological satellite. The method can consider dynamic changes of geometric parameters such as a solar zenith angle and an azimuth angle, and is more suitable for FMC inversion based on the stationary meteorological satellite, and is of great significance for near-real-time forest fire risk management and control.

[0042] Other advantages, objects, and features of the application will be set forth in part by the description that follows, and in part will become apparent to those skilled in the art upon examination of the following specification or can be learned by practice of the application. The objects and other advantages of the application can be realized and attained by the methods particularly pointed out in the written description and claims hereof. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to make the objects, technical solutions and advantages of the application clearer, the application provides the following drawings for description:

[0044] Figure 1 It is a whole method flowchart of the application.

[0045] Figure 2 It is a geographic location schematic diagram of the research plot in the specific embodiment of the application.

[0046] Figure 3 It is the result of inverting the combustible moisture content of China and Australia in the specific embodiment of the application. DETAILED DESCRIPTION

[0047] The application will be further described below in combination with the drawings and specific embodiments, so that those skilled in the art can better understand the application and implement it, but the embodiments are not used as limitations to the application.

[0048] Embodiment 1

[0049] As Figure 1 shown, the embodiment provides a vegetation canopy fuel moisture content inversion method based on geostationary meteorological satellite, comprising the following steps:

[0050] Step 1: Obtain geostationary meteorological satellite observation data, including standard albedo product in full disc observation mode and solar zenith angle and azimuth angle, satellite zenith angle and azimuth angle and other geometric parameter products; and complete cloud mask, atmospheric correction on the data, generate ground reflectance data for vegetation index calculation;

[0051] Step 2: Construct vegetation radiation transfer model and cost function T_RMSE, select chlorophyll a and b content C ab , equivalent water thickness EWT, fuel moisture content FMC, leaf area index LAI and soil factor psoil as free variables of PROSAIL vegetation radiation transfer model, other parameters of the model are constant; Construct the cost function T_RMSE about normalized difference vegetation index NDVI (formula 1), enhanced vegetation index EVI (formula 2), normalized difference infrared index NDII (formula 3), the calculation method is shown in formula (4).

[0052]

[0053]

[0054]

[0055]

[0056] Wherein, ρ Blue , ρ Red , ρ NIR , ρ SWIR2 are the spectral reflectance of blue light, red light, near infrared and short wave infrared band respectively; NDVI o , EVI o , NDII o are the observed vegetation indexes extracted from satellite data; NDVI m , EVI m , NDII m are the vegetation indexes simulated by PROSAIL model, which are also functions of free variables C ab , EWT, FMC, LAI and psoil, T_RMSE represents the cost function when

[0057] Step 3: Obtain the partial derivatives of the cost function T_RMSE with respect to the above 5 free variables using the automatic differentiation tool TAPENADE, which is used to calculate the gradient information;

[0058] Step 4: Solve the FMC in two stages using the gradient algorithm, in the first stage, invert the sensitive parameters EWT, LAI and psoil, in the second stage, invert the weak sensitive parameters FMC and C ab ;

[0059] Specifically, it includes 4 parts:

[0060] Step 4.1: Initial value setting of sensitive parameters EWT, LAI and psoil;

[0061] The initial value range of EWT, LAI and psoil is determined by the measured data, and each variable is selected from the minimum value, mean value and maximum value, so there are 3x3x3 = 27 initial value combinations in total;

[0062] Step 4.2: Solve the sensitive parameters EWT, LAI and psoil using the gradient algorithm, when the cost function T_RMSE1 meets formula (5), stop the first stage iteration, and output EWT, LAI and psoil, as well as FMC and C ab , which are used as initial value input for the second stage numerical iteration;

[0063]

[0064] Wherein, the cost function T_RMSE1 is based on formula (4), and the parameters FMC and C ab are fixed, and only EWT, LAI and psoil are used as free variables; is the gradient of T_RMSE1, which is calculated in step 3; ε1 is the allowed error limit;

[0065] The value range of ε1 is 0.1-1;

[0066] Step 4.3: Introduce the ecological rule based on FMC and LAI and EWT, and dynamically update the FMC value range through formula (6);

[0067]

[0068] Wherein, g1(LAI) is a linear fitting equation of FMC with respect to LAI, RMSE FMC-LAI is the root mean square error of the fitting equation; g2(EWT) is a linear fitting equation of FMC with respect to EWT, RMSE FMC-EWT is the root mean square error of the fitting equation;

[0069] Based on measured grassland canopy combustible moisture content data in China:

[0070]

[0071] Based on measured data of forest canopy combustible moisture content in China:

[0072]

[0073] Based on measured grassland canopy combustible moisture content data in Australia:

[0074]

[0075] Based on measured data on forest canopy combustible moisture content in Australia:

[0076]

[0077] in, The fitted combustible moisture content is LAI, the measured leaf area index is LAI, and the measured isohydrate thickness is EWT.

[0078] Step 4.4: Iteratively solve for the weakly sensitive parameters FMC and C using the reduced gradient algorithm. ab When the cost function T_RMSE1 does not satisfy formula (5), proceed to step 4.2; when the cost function T_RMSE2 satisfies formula (11), stop the second stage iteration and output FMC and C. ab The iteration results;

[0079]

[0080] The cost function T_RMSE2 is based on formula (4), with fixed parameters EWT, LAI, and psoil, and only FMC and C are included. ab As a free variable; The gradient of T_RMSE2 is calculated in step 3; ε2 is the allowable error limit.

[0081] The value range of ε2 is 10. -6 -10 -5 .

[0082] Step 5: For all FMC results obtained iteratively using the two-stage reduced gradient algorithm under different initial value combinations, calculate the FMC results according to the magnitude of their corresponding cost function values ​​(T_RMSE) and gradients. Perform primary and secondary sorting. First, select the m groups of FMC inversion results with the smallest cost function. Then, select the n groups of FMC inversion results with the smallest gradient magnitude from them. Take the average of these n groups of FMC as the final inversion result.

[0083] The value of m is 15; the value of n is 9.

[0084] Embodiment 2

[0085] As Figure 1 shown below, the vegetation canopy fuel moisture content inversion method based on geostationary meteorological satellite provided by the present application is further described in combination with specific embodiments and the accompanying drawings of the specification, including the following steps:

[0086] Step 1: Data preparation

[0087] The measured Chinese grassland canopy fuel moisture content (FMC) data was obtained by field investigation in the Qinghai Lake Basin grassland in China from July 28 to August 2, 2015; the measured Chinese forest canopy fuel moisture content data was obtained by field investigation in the southwest region of Sichuan Province, China, from December 5 to 10, 2020 and from March 16 to 20, 2021. The measured Australian forest and grassland canopy fuel moisture content data was obtained by field investigation in Canberra, Australia, from October 1, 2015 to April 20, 2016. As Figure 2 shown, the geographical position distribution of the measured data used in this embodiment is shown. The standard albedo product and the geometric parameter products such as solar zenith angle, solar azimuth angle, satellite zenith angle and satellite azimuth angle of the Himawari-8 satellite full-disk observation mode used in this embodiment. The data is pre-processed, including cloud mask and atmospheric correction, to generate the surface reflectance product.

[0088] Step 2: Construction of cost function

[0089] The PROSAIL vegetation radiation transfer model is parameterized, as shown in Table 1;

[0090] Table 1 PROSAIL model input parameters (G: grassland, F: forest)

[0091]

[0092] The chlorophyll a and b content C ab (μg / cm 2 ), equivalent water thickness EWT (g / cm 2 ), dry matter weight DMC (g / cm 2 ), leaf area index LAI and soil factor psoil are selected as the free variables of the PROSAIL vegetation radiation transfer model, and the other parameters of the model are constant; wherein the fuel moisture content The cost function T_RMSE of normalized difference vegetation index NDVI (formula 1), enhanced vegetation index EVI (formula 2), and normalized difference infrared index NDII (formula 3) is constructed, and the calculation method is shown in formula (4).

[0093] Step 3: Partial derivative acquisition of cost function

[0094] The automatic differentiation tool TAPENADE (download address: https: / / team.inria.fr / ecuador / en / tapenade , version: 3.6) is used to acquire the partial derivatives of the cost function T_RMSE with respect to C ab , EWT, FMC, LAI and psoil, which are used to calculate the gradient information;

[0095] Step 4: Numerical iterative solution of FMC

[0096] The FMC is iteratively solved in two stages by using the reduced gradient algorithm. In the first stage, the sensitive parameters EWT, LAI and psoil are inverted, and in the second stage, the weak sensitive parameters FMC and C ab are inverted.

[0097] Specifically, it includes four parts:

[0098] Step 4.1: Initial value setting of sensitive parameters EWT, LAI and psoil;

[0099] Under a single set of sensitive parameter initial value combination, the iteration point is easy to fall into local optimal value, which may lead to the uncertainty of FMC inversion, so it is necessary to set multiple initial values to repeat the numerical optimization process, which can effectively avoid the local optimal problem. The EWT, LAI and psoil initial value selection range is determined through the measured data, and each variable is selected from the minimum value, mean value and maximum value, so there are 3x3x3=27 initial value combinations in total.

[0100] Step 4.2: Numerical iterative solution of sensitive parameters EWT, LAI and psoil;

[0101] The sensitive parameters EWT, LAI and psoil are iteratively solved by using the reduced gradient algorithm. When the cost function T_RMSE1 meets formula (5), the first stage iteration is stopped, and EWT, LAI and psoil, as well as FMC and C ab , are output as the initial value input for the second stage numerical iteration.

[0102] Step 4.3: Introduction of ecological rules;

[0103] Based on the measured combustible moisture content data in the field, an empirical equation was fitted between the measured FMC and the measured LAI and EWT. Combined with formulas (6) to (10), an ecological rule was constructed, as shown in (12) to (15), and the range of FMC values ​​was dynamically updated accordingly.

[0104] Inversion of measured grassland canopy combustible moisture content in China:

[0105]

[0106] Inversion of measured forest canopy combustible moisture content in China:

[0107] |FMC-56.08*LAI-61.34||≤14.09% (13)

[0108] Inversion of grassland canopy combustible moisture content based on measured data in Australia:

[0109] |FMC-108.5*LAI-4.52||≤29.95% (14)

[0110] Inversion of measured forest canopy combustible moisture content in Australia:

[0111] |FMC-9.25*LAI-98.34||≤11.67% (15)

[0112] Step 4.4: Numerical iterative solution of the weakly sensitive parameters FMC and C ab ;

[0113] Iteratively solve the weakly sensitive parameters FMC and C using the reduced gradient algorithm. ab When the cost function T_RMSE1 does not satisfy formula (5), proceed to step 4.2; when the cost function T_RMSE2 satisfies formula (11), stop the second stage iteration and output FMC and C. ab The iteration results;

[0114] Using gridded search, the interval distribution of the partial derivatives of the cost function with respect to the variables is statistically analyzed, as shown in Table 2. The error limit ε2 should be less than 10. -5 To improve inversion accuracy and computational efficiency, ε2 is preferably 10. -6 ;

[0115] Table 2. Range of partial derivatives of the cost function with respect to the input parameters

[0116]

[0117] Step 5: Moisture content inversion of combustibles

[0118] All FMC results solved by using two-stage reduced gradient algorithm to iteratively solve under different initial value combinations, according to the cost function value (T_RMSE) and the modulus of its gradient The main and secondary sorting are carried out, first, the m groups of FMC inversion results with the minimum cost function are selected, then, the n groups of FMC inversion results with the minimum modulus of the gradient are selected from the m groups of FMC inversion results, and the mean value of the n groups of FMC is taken as the final inversion result; in order to alleviate the ill-posed inversion problem, the grid search method is used for parameter adjustment, that is, m=6, 9, …, 27; n=3, 6, …, 24 (m>n) are set respectively, the values of the parameters m and n producing the best inversion result are determined through repeated experiments, and it is known that the value of m is preferably 18 and the value of n is preferably 9;

[0119] According to the above inversion method, the combustible water content is solved. Figure 3 As shown in the figure, it is the inversion result of the combustible water content in the embodiment of the present application.

[0120] The above-mentioned embodiments are only preferred embodiments for fully illustrating the present application, and the protection scope of the present application is not limited thereto. The equivalent substitutions or transformations made by the person skilled in the art on the basis of the present application are all within the protection scope of the present application. The protection scope of the present application is subject to the claims.

Claims

1. A method for retrieving vegetation canopy fuel moisture content from geostationary meteorological satellite data, characterized in that: The method comprises the following steps: Step 1: obtaining static meteorological satellite observation data to generate surface reflectance data for vegetation index calculation; Step 2: constructing a vegetation radiation transfer model; constructing a cost function T_RMSE about normalized difference vegetation index NDVI, enhanced vegetation index EVI, and normalized difference infrared index NDII; Step 3: obtaining partial derivatives of the cost function T_RMSE about free variables for calculating gradient information; Step 4: FMC is solved in two stages using a reduced gradient algorithm, in the first stage the sensitive parameters equivalent water thickness EWT, leaf area index LAI and soil factor psoil are inverted, in the second stage the weak sensitive parameters fuel moisture content FMC and chlorophyll a and b content C are inverted ab ; Step 5: all fuel moisture content FMC results obtained by using a two-stage reduced gradient algorithm to iteratively solve under different initial value combinations; The all fuel moisture content FMC results obtained by using a two-stage reduced gradient algorithm to iteratively solve under different initial value combinations in the step 5 are specifically as follows: According to the main and secondary orders of the corresponding cost function values T_RMSE and the modulus of the gradient |▽T_RMSE|, the m groups of fuel moisture content FMC inversion results with the minimum cost function are first selected, and then the n groups of fuel moisture content FMC inversion results with the minimum modulus of the gradient are selected from the m groups, and the mean value of the n groups of fuel moisture content FMC is taken as the final inversion result.

2. The method of claim 1, wherein the method is based on static meteorological satellite-based vegetation canopy fuel moisture content retrieval. The normalized difference vegetation index NDVI, the enhanced vegetation index EVI, the normalized difference infrared index NDII, and the cost function T_RMSE thereof are calculated according to the following formula: (1) (2) (3) (4) where, , , , are the spectral reflectances in the blue, red, near-infrared and shortwave infrared bands, respectively; , , are the observed vegetation indices extracted from satellite data; , , are the vegetation indices simulated by the PROSAIL model, which are functions of the free variables C ab , EWT, FMC, LAI and psoil, and T_RMSE represents the cost function.

3. The method of claim 1, wherein: The partial derivatives of the free variables in the step 3 are calculated by using an automatic differentiation tool TAPENADE.

4. The method for inverting the combustible material moisture content of vegetation canopy based on geostationary meteorological satellites as described in claim 2, characterized in that: The FMC is iteratively solved in two stages by using a reduced gradient algorithm in the step 4, and the specific steps are as follows: Step 4.1: setting initial values of sensitive parameters EWT, LAI, and psoil; Step 4.2: Iterative solution of sensitive parameters EWT, LAI and psoil using the method of reduced gradient, when the cost function T_RMSE1 meets equation (5), stop the first stage iteration, and output EWT, LAI and psoil, as well as FMC and C ab , initial value input for the second stage numerical iteration; Meanwhile, go to step 4.4, otherwise execute step 4.3; (5) wherein the cost function T RMSE1 is modeled on equation (4) with the fixed parameters FMC and C ab with only EWT, LAI and psoil as free variables; ∇T RMSE1 is the gradient of T RMSE1, computed from step 3; ε 1 is the allowed error limit; Step 4.3: introducing an ecological rule based on FMC and LAI and EWT, and dynamically updating the FMC value range through formula (6); (6) where g1(LAI) is the linear fit equation for FMC vs. LAI, RMSE FMC-LAI is the root mean square error of the fit equation; g2(EWT) is the linear fit equation for FMC vs. EWT, RMSE FMC-EWT is the root mean square error of the fit equation; When the cost function T_RMSE1 does not satisfy formula (5), repeat this step until formula (5) is satisfied; Step 4.4: Iteratively solve the weak sensitive parameters FMC and C using the Hermite gradient algorithm ab ; stop the second stage iteration when the cost function T_RMSE2 satisfies equation (11), and output the iteration results of FMC and C ab ; (11) where the cost function T RMSE2 is modeled after equation (4) with fixed parameters EWT, LAI, and psoil, and only FMC and C ab as a free variable; ∇T RMSE2 is the gradient of T RMSE2, computed from step 3; ε 2 is the allowed error limit.

5. The method for inverting the combustible material moisture content of vegetation canopy based on geostationary meteorological satellites as described in claim 4, characterized in that: said e1 has a value in the interval of 0.1-1 ; or said e2 has a value in the interval of 10 -6 -10 -5 ; or said m has a value of 15; or said n has a value of 9.

6. The method of claim 1, wherein: The satellite observation data in the step 1 comprises standard albedo products, solar zenith angle and azimuth angle, satellite zenith angle and azimuth angle, and geometric parameter products in a full-disk observation mode; and cloud masking and atmospheric correction of the data are completed.

7. The method of claim 1, wherein: The vegetation radiation transfer model in the step 2 is constructed in the following manner: Chlorophyll a and b content C ab Moisture content of fuel FMC, leaf area index LAI and soil factor psoil as free variables of the PROSAIL vegetation radiative transfer model, the other parameters of the model are constants.

8. A system for retrieving vegetation canopy fuel moisture content from static meteorological satellite data, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the method of any one of claims 1 to 7 when executing the program.

Images