Method and device for inversely calculating methane concentration based on FP (Fabry-Perot) load

Abstract

The invention belongs to the technical field of methane concentration detection, and discloses a methane concentration inverse calculation method and device based on FP (Fabry-Perot) load, and the method comprises the steps: carrying out the transmission modulation of a solar reflection and scattering spectrum of a ground pixel through a fixed cavity Fabry-Perot interferometer; and the change information of the CH4 absorption intensity of the atmosphere in the observation area is inverted according to the change of the collected modulation signal, so that the emission source of CH4 is identified under the uniform background. Through the optimized spectrum restoration algorithm and the Levenberg-Marquardt iterative algorithm, the inversion precision is remarkably improved, the inversion error is reduced, the reliability and accuracy of the inversion result are ensured, in addition, the technology has more advantages in application in real-time monitoring and large-scale data processing due to improvement of the inversion efficiency, and the method is suitable for popularization and application. And finally, by combining an SCIATRAN forward modeling model, external data and an efficient algorithm, the problems of low efficiency and high error in a traditional method are solved, so that the monitoring of the methane concentration is more accurate and reliable.

Description

Technical Field

[0001] The present invention belongs to the technical field of methane concentration detection, and specifically relates to a methane concentration inversion method and device based on FP load. Background Art

[0002] Methane is a potent greenhouse gas, and its warming effect in the atmosphere is much greater than that of carbon dioxide. By monitoring methane concentration, we can promptly discover and control methane emission sources, and reduce its pollution and damage to the environment. Methane emissions are related to a variety of pollution sources, such as industrial emissions, agricultural activities, landfills, etc. Monitoring methane emissions helps identify these pollution sources and take corresponding control measures to improve air quality.

[0003] Satellite monitoring can provide real-time and accurate information on the total amount of methane emissions, the distribution of emission sources, and the intensity of emissions. This information is an important scientific basis for formulating effective emission reduction measures and helps national and local governments to formulate and adjust environmental protection policies in a targeted manner.

[0004] At present, the existing methane concentration measurement scheme mainly relies on satellite remote sensing technology, and inverts its concentration by analyzing the absorption spectrum of methane in the atmosphere. Existing schemes include full physical algorithm, PPDF algorithm, Levenberg-Marquardt iterative algorithm, etc. However, in the inversion of methane concentration, the existing satellite Fabry-Perot (FP) interferometer payload has an overlap in the absorption spectrum of methane in the atmosphere with other gases, as well as interference from factors such as surface reflection and atmospheric scattering, making the inversion process complicated and prone to errors. These errors not only affect the accuracy of methane concentration, but also limit its reliability in environmental monitoring and emission reduction strategy formulation. At the same time, the existing methane concentration inversion methods usually require a lot of computing resources and time, especially when processing large-scale data sets, which limits the application of inversion methods in real-time monitoring and rapid response, so they need to be improved and optimized. Summary of the invention

[0005] The object of the present invention is to provide a method and device for inverse calculation of methane concentration based on FP load to solve the problems raised in the above background technology.

[0006] In order to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a methane concentration inversion method and device based on FP load, which uses a fixed cavity Fabry-Pérot interferometer to transmittance modulate the solar reflection scattering spectrum of ground pixels, and inverts the atmospheric CH4 absorption intensity change information in the observation area according to the changes in the collected modulation signal, thereby identifying the emission source of CH4 under a uniform background, and the specific steps are as follows:

[0007] Step 1: Based on the obtained FP interferogram that has been corrected by system geometry, the spectral data of each pixel is inverted using the interferometric data inversion spectral algorithm;

[0008] Step 2: Using external data including initial methane concentration data, surface albedo data, atmospheric temperature, pressure, cloud and aerosol data, the methane spectrum curve is simulated based on the SCIATRAN forward model, and the optimal estimation iterative algorithm is used to invert the methane concentration of each pixel;

[0009] Step 3: Based on the methane concentration map, the methane plume is extracted using Gaussian filtering and threshold segmentation algorithm;

[0010] Step 4: Based on the plume extraction results, the integrated mass enhancement model is used to calculate the methane leakage emissions.

[0011] Preferably, the method restores the spectral intensity distribution of the incident light from the intensity distribution of the interference pattern by using the Levenberg-Marquardt algorithm, and when the light is incident vertically on the Fabry-Pérot interferometer, the intensity distribution I(x) of the output light satisfies the integral equation:

[0012]

[0013] Preferably, the numerical solution process of the integral equation is:

[0014] A1, discretization, divide σ in the interval [σmin,σmax] into N small intervals, and the width of each small interval is δσ;

[0015] A2, matrix form, convert the discrete equation into matrix form to obtain I = KB;

[0016] A3, solve the matrix equation. Solve the matrix equation I = K·B by numerical method to obtain B.

[0017] Preferably, the integral equation belongs to Fredholm's first-class integral equation, and the error is large. It is necessary to use a stability formula to determine the values ​​of various parameters in the equation, and the stability formula is:

[0018]

[0019] n=2P c

[0020]

[0021] Preferably, the integral equation is to optimize spectral restoration, and an error function E(B)=||I-KB|| 2To perform iterative optimization, this function is used to quantify the difference between I(x) and the result generated by the integral equation kernel function K and the function to be solved B(σ). When E(B) is small enough to fall within an acceptable range, it means that the error between the solved B(σ) and the true B0(σ) is acceptable, and the iteration will be stopped at this time.

[0022] Preferably, the iterative steps are:

[0023] B1, first define the error vector r and parameter vector B1(σ): r = I-KB;

[0024] B2, calculate the Jacobian matrix:

[0025] B3, construct Hessian matrix and gradient: H = J T J and g = J T r;

[0026] B4, at this time the parameter update amount is The new parameter B(σ) is B 2 (σ)=B 1 (σ)+ΔB(σ);

[0027] B5, calculate the error between B2(σ) and the true B(σ) again. If the error is large, continue to calculate the error between B3(σ) and the true B(σ) until the error is within an acceptable range.

[0028] Preferably, a hyperspectral satellite methane inversion system is established based on the inversion method, and the hyperspectral satellite methane inversion system includes an input module, a calculation module and an output module;

[0029] The calculation module includes a forward model calculation unit and an inverse solution calculation unit, and the output module includes the generation output of the profiles and column totals of O2 and CH4, the output of the average kernel function and the simulated spectrum, and the calculation of the average column concentration XCH4.

[0030] Preferably, the methane concentration inversion algorithm of the hyperspectral satellite methane inversion system will first apply the forward model F to x, optimize the state vector x relative to the measured value y, and finally construct an objective function and select an optimization strategy to gradually approach the true solution in an iterative form.

[0031] Preferably, the specific steps of the methane concentration inversion algorithm are:

[0032] C1, first find x by minimizing the cost function, the cost function is:

[0033] J(x)=[yF(x)] T S ε-1 [yF(x)]+(xx a ) T S a -1 (xx a );

[0034] C2, then use the Gauss-Newton iterative algorithm for nonlinear optimization, and the iterative algorithm is:

[0035]

[0036] C3, the inversion is completed and the final output is "profiles_nd.dat" and "profiles_vmr.dat" files, "ret_column.dat" file, "pro_retr_pre.dat" and "pro_retr_rms.dat" files, "wf_fit.dat" file, "precision.dat" file and "av_test.dat" file.

[0037] Preferably, the Fabry-Pérot interferometer comprises two parallel and highly reflective semi-transparent mirrors, and a parallel flat cavity is formed between the two. Light is reflected multiple times between the two mirrors, and each reflection produces partial transmitted light, eventually forming equi-inclined interference fringes, which appear as a series of concentric rings.

[0038] The beneficial effects of the present invention are as follows:

[0039] The present invention firstly provides a higher spatial resolution through the FP interferometer, which enables the methane emission source to be more accurately identified in a complex environment compared with the existing remote sensing technology. Secondly, the present invention significantly improves the inversion accuracy and reduces the inversion error through the optimized spectral restoration algorithm and the Levenberg-Marquardt (LM) iterative algorithm, thereby ensuring the reliability and accuracy of the inversion result. In addition, the improvement of the inversion efficiency also makes the application of the technology in real-time monitoring and large-scale data processing more advantageous, and the methane emission source can be discovered in time and effective environmental protection measures can be taken. Finally, by combining the SCIATRAN forward model, external data and an efficient algorithm, the problems of inefficiency and high error in the traditional method are solved, making the monitoring of methane concentration more accurate and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of the methane concentration inversion method of the present invention;

[0041] Figure 2 This is a diagram showing the functional modules of the hyperspectral satellite methane inversion system of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] like Figures 1 to 2 As shown, an embodiment of the present invention provides a methane concentration inversion method and device based on FP load. The method uses a fixed cavity Fabry-Pérot (FP) interferometer to transmittance modulate the solar reflection scattering spectrum of ground pixels, and inverts the CH4 absorption intensity change information of the atmosphere in the observation area according to the collected modulation signal changes, thereby identifying the CH4 emission source under a uniform background. The specific steps are as follows:

[0044] Step 1: Based on the obtained FP interferogram that has been corrected by system geometry, the spectral data of each pixel is inverted using the interferometric data inversion spectral algorithm;

[0045] Step 2: Using external data including initial methane concentration data, surface albedo data, atmospheric temperature, pressure, cloud and aerosol data, the methane spectrum curve is simulated based on the SCIATRAN forward model, and the optimal estimation iterative algorithm is used to invert the methane concentration of each pixel;

[0046] Step 3: Based on the methane concentration map, the methane plume is extracted using Gaussian filtering and threshold segmentation algorithm;

[0047] Step 4: Based on the plume extraction results, the integrated mass enhancement model is used to calculate the methane leakage emissions.

[0048] The method uses the Levenberg-Marquardt (LM) algorithm to restore the spectral intensity distribution of the incident light from the intensity distribution of the interference pattern, and when the light is vertically incident on the Fabry-Pérot (FP) interferometer, the intensity distribution I(x) of the output light satisfies the integral equation:

[0049]

[0050] Where σ is the beam, the unit is cm-1, B(σ) is the spectral intensity distribution, σ∈[σmin, σmax], is the spectral range under study, x is twice the distance between the two reflecting surfaces of the Fabry-Pérot interferometer, R is the reflectivity of the two reflecting surfaces, and I(x) is the intensity distribution of the output light of the Fabry-Pérot interferometer, that is, the intensity distribution of the interference pattern.

[0051] The numerical solution process of the integral equation is as follows:

[0052] A1, discretization, divide σ in the interval [σmin,σmax] into N small intervals, and the width of each small interval is δσ;

[0053] The corresponding discrete integral equation can be expressed as:

[0054]

[0055] A2, matrix form, convert the discrete equation into matrix form to obtain I = KB;

[0056] Where I is the experimentally measured intensity distribution vector [I(x1), I(x2), ..., I(xM)]T, and K is an M×N-order coefficient matrix composed of an integral kernel, which is:

[0057]

[0058] B is the spectral intensity distribution vector to be determined [B(σ1), B(σ2), ..., B(σN)]T. According to the given integral kernel, the elements of the matrix A are calculated:

[0059]

[0060] A3, solve the matrix equation, and solve the matrix equation I = K·B by numerical method to obtain B;

[0061] Among them, the integral equation belongs to Fredholm's first-class integral equation, and the error is large. It is necessary to use the stability formula to value each parameter in the equation, and the stability formula is:

[0062]

[0063] n=2P c

[0064]

[0065] Where Δσ=(σmax-σmin) is the beam range, δσ is the spectrum sampling interval when calculating the restored spectrum, Δx=(xmax-xmin) is the variation range of twice the distance between the two reflection surfaces of the Fabry-Pérot interferometer, n is the number of sampling points and the number of calculation points, R is the reflectivity of the two reflection surfaces of the Fabry-Pérot interferometer, Pc is the equivalent beam number, and σmax can be obtained according to the above formula min and xmax min .

[0066] The discretized algebraic equation of this integral equation is always ill-conditioned, and the solution of the equation is unstable. That is, if there is a large disturbance in B, it will only cause a small change in I after the action of K. However, if there is a small disturbance in I, a large error will be introduced in B. Therefore, from the perspective of the stability of solving the integral equation with respect to each parameter of the integral equation, it is necessary to use the stability formula to determine the value.

[0067] The integral function is to optimize the spectral restoration, and the error function E(B)=||I-KB|| 2 To perform iterative optimization, this function is used to quantify the difference between I(x) and the result generated by the integral equation kernel function K and the function to be solved B(σ). When E(B) is small enough to fall within an acceptable range, it means that the error between the solved B(σ) and the true B0(σ) is acceptable, and the iteration will be stopped at this time.

[0068] Wherein, the iterative steps are:

[0069] B1, first define the error vector r and parameter vector B1(σ): r = I-KB;

[0070] B2, calculate the Jacobian matrix:

[0071] B3, construct Hessian matrix and gradient: H = J T J and g = J T r;

[0072] Where H is the constructed Hessian matrix and JT is the transposed matrix of the Jacobian matrix. If the infinite norm of the gradient ||g||∞ is lower than the threshold, the calculation converges, otherwise the parameter B(σ) is updated and recalculated.

[0073] B4, at this time the parameter update amount is The new parameter B(σ) is B 2 (σ)=B 1 (σ)+ΔB(σ);

[0074] B5, calculate the error between B2(σ) and the true B(σ) again. If the error is large, continue to calculate the error between B3(σ) and the true B(σ) until the error is within an acceptable range.

[0075] When solving the integral equation, the main goal is to find the optimal B(σ) so that the error function E(B) is minimized, that is, by continuously iteratively optimizing B(σ), we can find an optimal function B(σ) to approximate the actual data B0(σ). Before the start of the iteration, we must first give an initial B1(σ), calculate the error E(B) under B1(σ), and the Jacobian matrix J. The Jacobian matrix is ​​the derivative matrix of the error function with respect to the initial parameters B1(σ). Each element Jij of the matrix is ​​the partial derivative of the error function E(B) with respect to the jth parameter B1j(σ).

[0076] Among them, a hyperspectral satellite methane inversion system is established based on the inversion method, which is suitable for processing satellite and other satellite detector data, and the hyperspectral satellite methane inversion system includes an input module, a calculation module and an output module;

[0077] The calculation module includes a forward model calculation unit and an inverse solution calculation unit, and the output module includes the generation output of the profiles and column totals of O2 and CH4, the output of the average kernel function and the simulated spectrum, and the calculation of the average column concentration XCH4.

[0078] The forward model calculation unit can perform accurate atmospheric radiation transfer simulation calculations based on the SCIATRAN radiation transfer model, taking into account solar radiation and instrument model functions, and using the line-by-line integration method to accurately simulate the spectrum of greenhouse gas CH4 sensed by hyperspectral satellite remote sensing;

[0079] The inversion solution calculation unit calls the simulated spectrum generated by the forward modeling module, and uses the optimization estimation algorithm to compare it with the spectrum actually observed by the instrument. The inverted CO2 concentration profile is used as the convergence detection condition. The input parameters of the forward model are continuously adjusted through iteration to make the simulated prediction results and the actual observation results of the instrument as consistent as possible, and these input parameters are regarded as actual observation values ​​to obtain the best atmospheric O2 and CH4 profiles and absolute column totals.

[0080] The output module contains the reading of the measured spectrum, prior information such as solar radiation parameters, atmospheric state parameters, surface parameters, satellite attitude parameters, instrument parameters and gas absorption characteristics.

[0081] Among them, the methane concentration inversion algorithm of the hyperspectral satellite methane inversion system will first apply the forward model F to x, optimize the state vector x relative to the measured value y, and finally construct the objective function and select the optimization strategy to gradually approach the true solution in an iterative form.

[0082] The specific steps of the methane concentration inversion algorithm are as follows:

[0083] C1, first find x by minimizing the cost function, the cost function is:

[0084] J(x)=[yF(x)] T S ε -1 [yF(x)]+(xx a ) T S a -1 (xx a );

[0085] The cost function represents the different minimum costs generated by optimizing the difference between the measured and simulated spectra and the difference between the state vector and the prior information, where S a is the prior error variance matrix, x a Prior state vector, S ε is the error covariance matrix;

[0086] C2, then use the Gauss-Newton iterative algorithm for nonlinear optimization, and the iterative algorithm is:

[0087]

[0088] In order to get closer to the real state, if the inverted atmospheric composition or physical quantity has reliable prior information, that is, the difference between the prior value and the inverted true value is small, the Gauss-Newton iterative algorithm is usually used, where: i+1 ,x i Respectively represent the inversion state vectors of the i+1th and ith times; K i is the weight function matrix at the i-th iteration. This iterative method minimizes the cost function. For each iteration step, the forward model is used to recalculate the weight function matrix and simulate the spectrum. In other words, at each iteration step, the solution to the inversion problem can be regarded as the deviation from the prior value.

[0089] C3, the inversion is completed and the final output is "profiles_nd.dat" and "profiles_vmr.dat" files, "ret_column.dat" file, "pro_retr_pre.dat" and "pro_retr_rms.dat" files, "wf_fit.dat" file, "precision.dat" file and "av_test.dat" file.

[0090] Stored in the files "profiles_nd.dat" and "profiles_vmr.dat", they are number density (molec / cm3) and column concentration mixing ratio (ppV), respectively. These two files have the same format: the first four columns of each file are a header row with auxiliary information; the first header row is the number of atmospheric species in the inversion; the second row is the names of these atmospheric species; the third row is each inverted atmospheric species defined with "T" or "F", if it is "T" then a vertical profile inversion is performed, if it is "F" then a priori profile scaling is performed (the inverted vertical profile is obtained by multiplying the a priori profile by the scaling factor); the fourth row is the vertical column concentration of each gas, which is obtained by integrating their vertical profiles over the entire height range. The header row is followed by a block of vertical profile data, the first column of which is the height in "km", the next three columns (vertical profiles) are for each atmospheric species, the order of the species being the same as the second data row. The first vertical profile represents the a priori information, the second and third profiles are the results after the second to last and last inversion iteration steps, respectively;

[0091] "ret_column.dat" outputs the total column concentration of all inverted species (i.e., the height integral of the vertical profile). The first three lines of the file are exactly the same inverted species information as "profiles_nd.dat" and "profiles_vmr.dat". The fourth line is the a priori total column concentration of the inverted species (i.e., the height integral of the prior profile), but the fifth line is the total column concentration after inversion (the same as the fourth line of "profiles_nd.dat" and "profiles_vmr.dat"). The column number represents the gas species;

[0092] The measured data, simulated data and fitting residuals before and after the last inversion iteration step are stored in the "pro_retr_pre.dat" and "pro_retr_rms.dat" files respectively. The first number in the first line of these two files is the spectral point used in the inversion process, the second number is the number of observed spectral points read in, and the third number is the total number of all sight lines; the second line is a blank line; the third line is all the wavelength grid points entered in the forward model; this data is saved in three columns; after the lines are separated, the next data block has 5 columns of data, and the first and second columns of the data block represent the number of sight lines and wavelengths of the corresponding data points, respectively. The third data column is the fitting residual. The fourth data column is the measured spectrum of the preprocessing process, that is, all required transformations and preprocessing corrections have been completed; the fifth data column is the simulated spectrum, which includes the effects caused by parameter changes in the current iterative inversion step. Similar to the measured spectrum, all required transformations and all corrections in the preprocessing stage have been completed;

[0093] The "wf_fit.dat" file includes the effects of individual atmospheric species in all simulated spectra. Compared to the format of the "pro_retr_*.dat" files, the first header line of this file has one more value (the fourth one), which gives the total number of atmospheric species in the inversion. Starting from the third line is the second data block, that is, all wavelength grid points input in the forward model, which is the same format as "pro_retr_*.dat"; the following data block includes two numbered data columns with the same meaning as the "pro_retr_*.dat" file. The next two spectral data columns represent a description of an inverted atmospheric species: the first column is the contribution of the initial atmospheric value to the total simulated signal; the second column is the contribution of the change in the inverted value to the simulated spectrum; therefore, the sum of the two columns of data gives the contribution of the total inverted profile to the simulated spectrum. The order of the inverted atmospheric species is listed in the second line of "profiles_*.dat";

[0094] The first column of the file "precision.dat" represents the sequence number of the altitude layer, which is arranged in the order of the inverted gases; the second and third columns are the diagonal elements of the covariance matrix of the inverted solution and the diagonal elements of the prior covariance matrix, respectively; the output order is as follows: first output the first covariance value of the first inverted gas type of all inverted altitudes, then the second, and so on; the altitude grid of the inverted gas is listed in the first column of "profiles_*.dat"; the order of the inverted atmospheric types is listed in the second row of "profiles_*.dat";

[0095] The file "av_test.dat" is the averaging kernel matrix, and the data is stored in the same way as the file "precision.dat" in the column direction.

[0096] The Fabry-Pérot (FP) interferometer includes two parallel and highly reflective semi-transparent mirrors (or mirrors), and a parallel flat cavity is formed between the two. Light is reflected multiple times between the two mirrors, and each reflection will produce partial transmitted light, eventually forming equi-inclined interference fringes, which appear as a series of concentric rings.

[0097] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0098] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A methane concentration inversion method based on FP load, characterized by: This method uses a fixed-cavity Fabry-Pérot interferometer to transmittance modulate the solar reflection scattering spectrum of ground pixels, and inverts the information of atmospheric CH4 absorption intensity changes in the observation area based on the changes in the collected modulation signals, thereby identifying the emission source of CH4 under a uniform background. The specific steps are as follows: Step 1: Based on the obtained FP interferogram that has been corrected by system geometry, the spectral data of each pixel is inverted using the interferometric data inversion spectral algorithm; Step 2: Using external data including initial methane concentration data, surface albedo data, atmospheric temperature, pressure, cloud and aerosol data, the methane spectrum curve is simulated based on the SCIATRAN forward model, and the optimal estimation iterative algorithm is used to invert the methane concentration of each pixel; Step 3: Based on the methane concentration map, the methane plume is extracted using Gaussian filtering and threshold segmentation algorithm; Step 4: Based on the plume extraction results, the integrated mass enhancement model is used to calculate the methane leakage emissions.

2. The methane concentration inversion method based on FP load according to claim 1 is characterized in that: This method uses the Levenberg-Marquardt algorithm to restore the spectral intensity distribution of the incident light from the intensity distribution of the interference pattern. When the light is incident vertically on the Fabry-Pérot interferometer, the intensity distribution I(x) of the output light satisfies the integral equation:

3. The methane concentration inversion method based on FP load according to claim 2 is characterized in that: The numerical solution process of the integral equation is: A1, discretization, divide σ in the interval [σmin,σmax] into N small intervals, and the width of each small interval is δσ; A2, matrix form, convert the discrete equation into matrix form to obtain I = KB; A3, solve the matrix equation. Solve the matrix equation I = K·B by numerical method to obtain B.

4. The methane concentration inversion method based on FP load according to claim 2 is characterized in that: The integral equation belongs to Fredholm's first-class integral equation, and the error is large. It is necessary to use the stability formula to determine the values ​​of each parameter in the equation, and the stability formula is: n=2P c 5. The methane concentration inversion method based on FP load according to claim 2 is characterized in that: The integral equation is to optimize the spectral restoration, and the error function E(B)=||I-KB|| 2 To perform iterative optimization, this function is used to quantify the difference between I(x) and the result generated by the integral equation kernel function K and the function to be solved B(σ). When E(B) is small enough to be within an acceptable range, it means that the error between the solved B(σ) and the true B0(σ) is acceptable, and the iteration will be stopped at this time.

6. The methane concentration inversion method based on FP load according to claim 5 is characterized in that: The iterative steps are: B1, first define the error vector r and parameter vector B1(σ): r = I-KB; B2, calculate the Jacobian matrix: B3, construct Hessian matrix and gradient: H = J T J and g = J T r; B4, at this time the parameter update amount is The new parameter B(σ) is B2(σ)=B1(σ)+ΔB(σ); B5, calculate the error between B2(σ) and the true B(σ) again. If the error is large, continue to calculate the error between B3(σ) and the true B(σ) until the error is within an acceptable range.

7. The methane concentration inversion method based on FP load according to claim 1 is characterized by: A hyperspectral satellite methane inversion system is established based on the inversion method, and the hyperspectral satellite methane inversion system includes an input module, a calculation module and an output module; The calculation module includes a forward model calculation unit and an inverse solution calculation unit, and the output module includes the generation output of the profiles and column totals of O2 and CH4, the output of the average kernel function and the simulated spectrum, and the calculation of the average column concentration XCH4.

8. The methane concentration inversion method based on FP load according to claim 7 is characterized by: The methane concentration inversion algorithm of the hyperspectral satellite methane inversion system first applies the forward model F to x, optimizes the state vector x relative to the measured value y, and finally constructs the objective function and selects the optimization strategy to gradually approach the true solution in an iterative form.

9. The methane concentration inversion method based on FP load according to claim 1 is characterized in that: The specific steps of the methane concentration inversion algorithm are: C1, first find x by minimizing the cost function, the cost function is: J(x)=[y-F(x)] T S ε -1 [y-F(x)]+(x-x a ) T S a -1 (x-x a ); C2, then use the Gauss-Newton iterative algorithm for nonlinear optimization, and the iterative algorithm is: C3, after the inversion is completed, the final output files are "profiles_nd.dat" and "profiles_vmr.dat", "ret_column.dat", "pro_retr_pre.dat" and "pro_retr_rms.dat", "wf_fit.dat", "precision.dat" and "av_test.dat".

10. The methane concentration inversion device based on FP load according to claim 1, characterized in that: The Fabry-Pérot interferometer includes two parallel and highly reflective semi-transparent mirrors, and a parallel plate cavity is formed between the two. Light is reflected multiple times between the two mirrors, and each reflection will produce partial transmitted light, eventually forming equal-inclination interference fringes, which appear as a series of concentric rings.

Images