A method and system for retrieving sea surface wind field based on two-dimensional spectrum observation data of sea waves

Abstract

This invention discloses a method and system for inverting sea surface wind fields based on two-dimensional wave spectrum observation data. The method includes: obtaining a two-dimensional spectrum of wind wave components and several two-dimensional spectra of swell components based on the observed two-dimensional wave spectrum; calculating the parameter data of the two-dimensional spectrum of wind wave components and the parameter data of each two-dimensional spectrum of swell components; constructing basic elements for wind speed inversion based on the parameter data of the two-dimensional spectrum of wind wave components; obtaining an initial wind speed inversion model based on the basic elements of wind speed inversion; correcting the model using the modulation effect of swell components to obtain a final wind speed inversion model; establishing an initial wind direction inversion model based on the parameter data of the two-dimensional spectrum of wind wave components; correcting the initial wind direction inversion model considering the asymmetry of the directional distribution of the two-dimensional spectrum of wind wave components and the high-frequency modulation effect of swell components on wind wave components to obtain a final wind direction inversion model; and calculating the wind speed and wind direction corresponding to the two-dimensional wave spectrum based on the final wind speed inversion model and the final wind direction inversion model.

Description

Technical Field

[0001] This invention relates to the field of marine engineering technology, and in particular to a method and system for inverting sea surface wind field based on two-dimensional wave spectrum observation data. Background Technology

[0002] Information on wind fields and waves at sea surface is of great significance to various human activities, production, and life at sea, and is also essential observational data for many marine scientific studies.

[0003] Currently, many field observation devices or remote sensing satellites can only observe one parameter of either the wind field or the ocean waves. However, the growth of ocean waves is the result of the combined control of wind speed, wind time, and wind region. Due to the presence of swells from the outer ocean, it is impossible to accurately infer ocean wave information based solely on local wind field data.

[0004] However, inferring wind field information from wave spectra is feasible to some extent. Under uniform and stable wind forcing, high-frequency wave spectra will enter an "equilibrium" state where energy input and output are equal. Existing technology has established models for inverting wind speed and direction using one-dimensional wave spectra observed from buoy data and the average direction of energy in the equilibrium domain. However, this model assumes a constant wave span and ignores issues such as swell modulation, inconsistency between the main wave direction and the average direction, and differences in the actual speed of waves at different frequencies. The resulting root mean square error of the wind speed is too large to be practical.

[0005] Building upon this, existing technology discloses a deep neural network model that uses the Fourier coefficients of the first five wave spectra as input and wind speed and direction as output, achieving an accuracy sufficient for practical application. However, training this model requires a large amount of simultaneous observation of the Fourier coefficients of the first five wave spectra and wind speed / direction, and acquiring such simultaneous observation data can sometimes be costly. Furthermore, due to the immutable input format and "black box" nature of neural networks, neural network models trained on observation data from one type of equipment are often not applicable to other equipment.

[0006] Therefore, a general-purpose and practical algorithm that is independent of the observation equipment and its data format is needed to accurately invert the wind field on the sea surface using ocean wave observation information. Summary of the Invention

[0007] The purpose of this invention is to provide a method and system for retrieving sea surface wind fields based on two-dimensional wave spectrum observation data. By utilizing the observation of two-dimensional wave spectrum, reliable retrieval of sea surface wind field data can be achieved, making the observation of various types of two-dimensional wave spectrum more practical.

[0008] To achieve the above objectives, this invention provides a method for inverting sea surface wind fields based on two-dimensional wave spectral observation data, comprising:

[0009] The observed two-dimensional wave spectrum was sequentially processed by spectral smoothing and spectral segmentation to obtain two-dimensional spectra of wind wave components and several two-dimensional spectra of swell components, respectively.

[0010] Calculate the parameter data of the two-dimensional spectrum of wind wave component and the parameter data of the two-dimensional spectrum of each swell component separately;

[0011] Based on the parameter data of the two-dimensional spectrum of wind and wave components, basic elements for wind speed inversion are constructed, and an initial wind speed inversion model is obtained based on the basic elements for wind speed inversion. Based on the parameter data of the two-dimensional spectrum of each surge component, considering the modulation effect of the surge component on the high frequency of wind and waves, the initial wind speed inversion model is modified to obtain the final wind speed inversion model.

[0012] An initial wind direction inversion model is established based on the parameter data of the two-dimensional spectrum of wind and wave components. Considering the asymmetry of the directional distribution of the two-dimensional spectrum of wind and wave components and the high-frequency modulation effect of the swell components on the wind and wave components, the initial wind direction inversion model is corrected to obtain the final wind direction inversion model.

[0013] The coefficients of the final wind speed inversion model and the final wind direction inversion model are determined respectively. Based on the final wind speed inversion model and the final wind direction inversion model, the wind speed and wind direction corresponding to the two-dimensional wave spectrum are calculated.

[0014] Furthermore, the parameter data of the two-dimensional spectrum of the wind and wave components includes the energy at different frequencies, the average direction at different frequencies, the main peak direction at different frequencies, the wave spread at different frequencies, and the spectral peak frequency and spectral peak direction of the wind and wave components. Calculating the parameter data of the two-dimensional spectrum of the wind and wave components includes the following steps:

[0015] Let the two-dimensional spectrum of wind and wave components be S(f) i ,θ j The number of grids for frequency and direction in the two-dimensional spectrum of wind and wave components are N, respectively. f and N θ And the two-dimensional spectrum of wind and wave components S(f) i ,θ j It can be decomposed into the product of energy at different frequencies and distribution functions in different frequency directions: in S represents the distribution function of different frequency directions of the wind and wave components. f (f i () represents the different frequency energies of the wind and wave components;

[0016] Based on the two-dimensional spectrum of wind and wave components S(f) i ,θ j The frequency and direction corresponding to the maximum value are used to obtain the spectral peak frequency f of the wind and wave components. p Spectral peak direction θ of wind and wave components p ;

[0017] The energy of different frequencies of the wind and wave components is calculated using the following formula:

[0018]

[0019] In the formula, S f (f i S(f) represents the energy of different frequencies of the wind and wave components. i ,θ j ) represents the two-dimensional spectrum of wind and wave components, f i θ represents the frequency corresponding to the i-th frequency grid. j Δθ represents the direction corresponding to the j-th direction grid, Δθ represents the angular resolution of the wave direction spectrum, and j represents the grid number of the direction grid;

[0020] The distribution function of wind wave components in different frequency directions is calculated using the following formula:

[0021]

[0022] In the formula, Δf represents the frequency distribution function of the wave component at different frequencies, i represents the frequency resolution of the wave direction spectrum, and i represents the frequency grid number.

[0023] The main peak directions θ of different frequencies of wind and wave components iM Let be the direction corresponding to the maximum value of the distribution function of the wind and wave components at different frequencies. The average direction of the wind and wave components at different frequencies is calculated based on the distribution function of the wind and wave components at different frequencies:

[0024]

[0025] In the formula, θ io The average direction of different frequencies of the wave components is represented by ai and bi, which are intermediate variables in the calculation of wave span and direction. The calculation formula is as follows:

[0026] The wave spans of different frequencies of wind wave components are calculated using the following formula:

[0027]

[0028] In the formula, σ(f i ) represents the wave spread of different frequencies of the wind and wave components.

[0029] Furthermore, the parameter data of the two-dimensional spectrum of the surge component includes the significant wave height, peak frequency, and peak direction of the surge component. Calculating the parameter data of the two-dimensional spectrum of each surge component includes the following steps:

[0030] Assuming that the two-dimensional wave spectrum yields N swell components through spectral segmentation, then let the two-dimensional spectrum of the nth swell component be F. n (f i ,θ j The number of grids for frequency and direction in the two-dimensional spectrum of swell components are N, respectively. f and N θ Where 1 ≤ n ≤ N;

[0031] Calculate the effective wave height of the two-dimensional spectrum of the nth surge component based on the number of grids for frequency and direction.

[0032] Based on the two-dimensional spectrum F of the nth surge component n (f i ,θ j The frequency and direction of the surge spectrum peak are obtained by determining the frequency and direction of the maximum value of the peak value.

[0033] The effective wave height of the nth swell component is calculated using the following formula:

[0034]

[0035] In the formula, F represents the effective wave height of the nth swell component. n (f i ,θ j ) represents the two-dimensional spectrum of the nth surge component, f i θ represents the frequency corresponding to the i-th frequency grid, where i represents the grid number; j Δθ represents the direction corresponding to the j-th direction grid, where j represents the grid number; Δθ represents the angular resolution of the wave direction spectrum, and Δf represents the frequency resolution of the wave direction spectrum.

[0036] Furthermore, basic elements for wind speed inversion are constructed based on the parameter data of the two-dimensional spectrum of wind and wave components. An initial wind speed inversion model is obtained based on these basic elements, including:

[0037] Selecting different frequency energies S of the wind and wave components in the equilibrium domain f (f i Based on the proportional relationship between energy and frequency to the power of -4 in the equilibrium domain, and considering the wave spread of different frequencies of wind and wave components, the basic elements for wind speed inversion are constructed.

[0038] Based on the equilibrium domain theory, after approximating the basic elements of wind speed inversion as a constant, the basic elements of wind speed inversion are subjected to quality control to obtain the quality-controlled basic elements of wind speed inversion.

[0039] Calculate the arithmetic mean of the basic elements of wind speed inversion after quality control within the equilibrium domain;

[0040] In the equilibrium domain, the wind speed is approximated as a linear function of the arithmetic mean of the basic elements of the wind speed inversion, thus obtaining the initial model for wind speed inversion.

[0041] The basic elements for wind speed inversion are calculated using the following formula:

[0042] W(f i )=σ(f i )S f (f i )f i 4 ,

[0043] In the formula, W(f) i ) represents the basic element for wind speed inversion at different frequencies, σ(f i f represents the wave span of different frequencies of the wind wave component. i This represents the frequency corresponding to the i-th frequency grid.

[0044] The arithmetic mean of the basic elements of the wind speed inversion after mass control within the equilibrium domain is calculated using the following formula:

[0045]

[0046] In the formula, The expression represents the arithmetic mean of the basic elements of the wind speed inversion, where m represents the value in the equilibrium domain 1.3f. p <f i <3f p After intermediate quality control, it meets the condition f i The number of; f i f represents the frequency corresponding to the i-th frequency grid. p Indicates the spectral peak frequencies of the wind and wave components;

[0047] The initial model for wind speed inversion is:

[0048]

[0049] In the formula, U ori This represents the wind speed retrieved from the initial model for wind speed inversion, where α and η are both undetermined coefficients.

[0050] Furthermore, based on the parameter data of the two-dimensional spectral data of swell, considering the high-frequency modulation effect of swell components on wind wave components, the initial wind speed inversion model is modified to obtain the final wind speed inversion model, including:

[0051] Considering the high-frequency modulation effect of swell components on wind wave components, the influence of swell components on wind speed inversion is approximated as a linear superposition of the influences of various swell components.

[0052] The corrections made by each wave component to the initial model for wind speed inversion are considered as functions of the effective wave height, wave peak frequency, and wave peak direction of the wave component.

[0053] Based on the relationship between the influence of the swell component on the high-frequency spectrum of the wind wave component and the energy of the swell, as well as the influence of whether the wind wave component and the swell component are aligned on the modulation effect, the wind speed inversion correction amount corresponding to each swell component is obtained.

[0054] The initial wind speed inversion model is corrected based on the correction amount of the wind speed inversion to obtain the final wind speed inversion model.

[0055] The correction amount for wind speed inversion is calculated using the following formula:

[0056]

[0057] In the formula, U Cor This represents the correction amount for wind speed inversion. This represents the effective wave height of the nth swell component. This represents the spectral peak frequency of the nth surge component. Let θp represent the spectral peak direction of the nth swell component, N represent the total number of swell components, and θp represent the spectral peak direction of the wind wave component. β, γ, κ, ξ and ψ both represent undetermined coefficients;

[0058] Adding the original wind speed inversion model and the correction amount of the wind speed inversion, we obtain the final wind speed inversion model as follows:

[0059]

[0060] In the formula, U represents the wind speed in the final wind speed inversion model, and α and η are both undetermined coefficients. This represents the arithmetic mean of the basic elements of wind speed inversion after quality control within the equilibrium domain.

[0061] Furthermore, an initial wind direction inversion model is established based on the parameter data of the two-dimensional spectrum of wind and wave components. Considering the asymmetry of the directional distribution of the two-dimensional spectrum of wind and wave components and the high-frequency modulation effect of swell components on wind and wave components, the initial wind direction inversion model is modified to obtain the final wind direction inversion model, including:

[0062] Assuming there is no swell component and the distribution functions of different frequencies of the wind and wave components are symmetrical, the energy weighted average of the average directions of different frequencies in the equilibrium domain of the wind and wave components is regarded as the wind direction, and an initial wind direction inversion model is obtained.

[0063] Considering the skewness of the two-dimensional spectral distribution of wind and wave components, the main peak directions of different frequencies of wind and wave components are used as the first correction factor for the wind direction model. The initial wind direction inversion model is then corrected using the first correction factor to obtain the corrected wind direction inversion model.

[0064] Considering the high-frequency modulation effect of swell components on wind wave components, the influence of swell components on wind direction inversion is approximated as a linear superposition of the influences of various swell components.

[0065] The second correction of each surge component to the wind direction model is regarded as a function of the effective wave height, surge spectrum peak frequency and surge spectrum peak direction of the surge component; the second correction is used to perform a second correction on the corrected wind direction inversion model to obtain the final wind direction inversion model.

[0066] Furthermore, the initial model for wind direction inversion is as follows:

[0067]

[0068] In the formula, θ W _ori represents the wind direction retrieved from the initial model for wind direction inversion, f i f represents the frequency corresponding to the i-th frequency grid. p S represents the spectral peak frequency of the wind and wave components; f (f i ) represents the energy of different frequencies of the wind and wave components, P represents the undetermined coefficient, and θ io Indicates the average direction of different frequencies of wind and wave components;

[0069] The corrected wind direction inversion model is as follows:

[0070]

[0071] In the formula, θ W θ represents the wind direction retrieved by the corrected wind direction inversion model. iM The directions of the main peaks of different frequencies of the wind and wave components are indicated by r and q, where r and q are both undetermined coefficients.

[0072] The final wind direction inversion model is as follows:

[0073]

[0074] In the formula, θ W_Final θ represents the wind direction retrieved by the final wind direction inversion model. W This indicates the wind direction retrieved by the corrected wind direction inversion model. This represents the effective wave height of the nth swell component. This represents the spectral peak frequency of the nth surge component. The spectral peak direction of the nth surge component is indicated by s, where n represents the nth surge component, N represents the total number of surge components, and s and t are undetermined coefficients.

[0075] Based on the same inventive concept, this invention also provides a system for inverting sea surface wind fields based on two-dimensional wave spectral observation data, comprising:

[0076] The processing unit is used to perform spectral smoothing and spectral segmentation on the observed two-dimensional wave spectrum in sequence to obtain two-dimensional spectra of wind wave components and several two-dimensional spectra of swell components, respectively.

[0077] The calculation unit is used to calculate the parameter data of the two-dimensional spectrum of wind wave component and the parameter data of the two-dimensional spectrum of each swell component, respectively.

[0078] The wind speed inversion unit is used to construct basic elements for wind speed inversion based on the parameter data of the two-dimensional spectrum of wind and wave components, and obtain an initial wind speed inversion model based on the basic elements of wind speed inversion. Based on the parameter data of the two-dimensional spectrum of each surge component, considering the high-frequency modulation effect of the surge component on the wind and wave component, the initial wind speed inversion model is corrected to obtain the final wind speed inversion model.

[0079] The wind direction inversion unit is used to establish an initial wind direction inversion model based on the parameter data of the two-dimensional spectrum of wind and wave components. The initial wind direction inversion model is modified by considering the asymmetry of the directional distribution of the two-dimensional spectrum of wind and wave components and the high-frequency modulation effect of the swell component on the wind and wave components, so as to obtain the final wind direction inversion model.

[0080] The integration unit is used to determine the coefficients of the final wind speed inversion model and the final wind direction inversion model, respectively, and calculate the wind speed and wind direction corresponding to the two-dimensional wave spectrum based on the final wind speed inversion model and the final wind direction inversion model.

[0081] Based on the same inventive concept, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the aforementioned method for inverting sea surface wind field based on two-dimensional wave spectrum observation data.

[0082] Based on the same inventive concept, embodiments of the present invention also provide a computer storage medium storing computer-executable instructions, which, when executed, implement the aforementioned method for inverting sea surface wind field based on two-dimensional wave spectrum observation data.

[0083] The technical effects and advantages of this invention are as follows: 1. Before inverting the two-dimensional spectrum, this invention first segments the two-dimensional spectrum of ocean waves, considering the direct response of the wind wave component to wind speed and the modulation of the swell component to the response of the wind wave component separately; when constructing the basic elements for wind speed inversion, the contribution of different frequency ocean wave spans is considered, and based on the weak variation (stability) characteristics of the basic elements in the equilibrium domain, the data quality is controlled to improve the accuracy of wind speed inversion; when establishing the wind direction inversion model, the factor that the weight of high-energy frequencies is greater than that of low-energy frequencies is considered, and an energy-related weighting factor is introduced. At the same time, by considering the asymmetry of the direction distribution (i.e., the average direction and the main peak direction are inconsistent) and the modulation of the high-frequency ocean wave two-dimensional spectrum by the swell, a second-order correction of the wind direction is made to improve the accuracy of wind direction inversion; it also considers the modulation effect of the swell component on the wind wave component's response to the wind field under different angles between the wind wave component and the swell component, and introduces a swell correction term for wind speed and wind direction inversion, making the inversion results more accurate;

[0084] 2. This invention enables reliable inversion of sea surface wind field data using the observation of two-dimensional wave spectra, making the observation of various types of two-dimensional wave spectra more practical. It can provide more data references for marine meteorological forecasting, marine disaster prevention and mitigation, marine management and other fields, increase the sources of sea surface wind field data, and thus reduce the average acquisition cost of sea surface wind field data.

[0085] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0086] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0087] Figure 1 This is a flowchart illustrating the steps of a method for retrieving sea surface wind field based on two-dimensional wave spectrum observation data according to an embodiment of the present invention.

[0088] Figure 2 This is a schematic diagram of the structure of a system for inverting sea surface wind field based on two-dimensional wave spectrum observation data according to an embodiment of the present invention;

[0089] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0090] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0091] To address the shortcomings of existing technologies, this invention discloses a method for inverting sea surface wind fields based on two-dimensional wave spectrum observation data, such as... Figure 1 As shown, it includes the following steps:

[0092] Step S1: Perform spectral smoothing and spectral segmentation on the observed two-dimensional wave spectrum to obtain two-dimensional spectra of wind wave component and swell component respectively;

[0093] Specifically, the convolution kernel method or other similar methods are used to perform spectral smoothing on the two-dimensional wave spectrum, and the watershed algorithm or other similar methods are used to perform spectral segmentation on the two-dimensional wave spectrum.

[0094] Step S2: Calculate the parameter data of the two-dimensional spectrum of wind wave components and the parameter data of the two-dimensional spectrum of swell components, respectively; including:

[0095] The parameter data of the two-dimensional spectrum of wind and wave components include energy at different frequencies, average direction at different frequencies, main peak direction at different frequencies, wave span at different frequencies, and spectral peak frequency and direction of the wind and wave components. The calculation of the parameter data of the two-dimensional spectrum of wind and wave components includes the following steps:

[0096] Let the two-dimensional spectrum of wind and wave components be S(f) i ,θ j The number of grids for frequency and direction in the two-dimensional spectrum of wind and wave components are N, respectively. f and N θ And the two-dimensional spectral matrix S(f) of wind and wave components i ,θ j It can be decomposed into the product of energy at different frequencies and distribution functions in different frequency directions: in S represents the distribution function of different frequency directions of the wind and wave components. f (f i () represents the different frequency energies of the wind and wave components;

[0097] Based on the two-dimensional spectral matrix of wind and wave components, S(f i ,θ j The frequency and direction corresponding to the maximum value of ) are used to obtain the spectral peak frequency f of the wind wave component. p Spectral peak direction θ of wind and wave components p .

[0098] The energy of different frequencies of the wind and wave components is calculated using the following formula:

[0099]

[0100] In the formula, S f (f i S(f) represents the energy of different frequencies of the wind and wave components. i ,θ j ) represents the two-dimensional spectrum of wind and wave components, f i θ represents the frequency corresponding to the i-th frequency grid. j Δθ represents the direction corresponding to the j-th direction grid, Δθ represents the angular resolution of the wave direction spectrum, and j represents the grid number of the direction;

[0101] The distribution function of wind wave components in different frequency directions is calculated using the following formula:

[0102]

[0103] In the formula, Let f represent the distribution function of different frequency directions of the wave component, Δf represent the frequency resolution of the wave direction spectrum, i represent the frequency grid number, and j represent the direction grid number.

[0104] The main peak direction θ of the wind and wave components iM Let be the direction corresponding to the maximum value of the distribution function of the wind and wave components at different frequencies. The average direction of the wind and wave components at different frequencies is calculated based on the distribution function of the wind and wave components at different frequencies:

[0105]

[0106] In the formula, θ io The average direction of different frequencies of the wave components is represented by ai and bi, which are intermediate variables in the calculation of wave span and direction. The calculation formula is as follows:

[0107] The wave spans of different frequencies of wind wave components are calculated using the following formula:

[0108]

[0109] In the formula, σ(f i ) represents the wave spread of different frequencies of the wind and wave components.

[0110] The parameter data of the two-dimensional spectrum of the surge component includes the significant wave height, peak frequency, and peak direction of the surge component. Calculating the parameter data of the two-dimensional spectrum of each surge component includes the following steps:

[0111] Assuming that the two-dimensional wave spectrum yields N swell components through spectral segmentation, then let the two-dimensional spectrum of the nth swell component be F. n (f i ,θ j The number of grids for frequency and direction in the two-dimensional spectrum of swell components are N, respectively. f and N θ Where 1 ≤ n ≤ N;

[0112] The effective wave height H of the two-dimensional spectrum of the nth surge component is calculated based on the number of grids for frequency and direction. Swell ;

[0113] Based on the two-dimensional spectral matrix F of the nth surge component n (f i ,θ i The frequency and direction of the surge spectrum peak are obtained by determining the maximum value of the peak value.

[0114] The effective wave height of the nth swell component is calculated using the following formula:

[0115]

[0116] In the formula, F represents the effective wave height of the nth swell component. n (f i ,θ j ) represents the two-dimensional spectrum of the nth surge component, f i θ represents the frequency corresponding to the i-th frequency grid, where i represents the grid number; j Δθ represents the direction corresponding to the j-th direction grid, where j represents the grid number; Δθ represents the angular resolution of the wave direction spectrum, and Δf represents the frequency resolution of the wave direction spectrum.

[0117] Step S3: Construct basic elements for wind speed inversion based on the parameter data of the two-dimensional spectrum of wind and wave components, and obtain the initial wind speed inversion model based on the basic elements; based on the parameter data of the two-dimensional spectrum of swell components, considering the high-frequency modulation effect of swell components on wind and wave components, modify the initial wind speed inversion model to obtain the final wind speed inversion model; including:

[0118] Select the equilibrium region (1.3f) p <f i <3f p The different frequency energies S of the wind and wave components in ) f (f i ), based on f -4 (i.e., the relationship between energy and frequency is proportional to the -4th power in the equilibrium domain) and considering the wave spread of different frequencies of wind and wave components, we construct the basic elements for wind speed inversion.

[0119] The basic elements for wind speed inversion are calculated using the following formula:

[0120] W(f i )=σ(f i )S f (f i )f i 4 ,

[0121] In the formula, W(f) i ) represents the basic element for wind speed inversion, σ(f i S represents the wave span of different frequencies of the wind wave components. f (f i f represents the different frequency energies of the wind and wave components. i This represents the frequency corresponding to the i-th frequency grid.

[0122] According to the equilibrium domain theory, the basic element W(f) for wind speed inversion is... i After approximating the wind speed to a constant, the basic element W(f) for wind speed inversion is removed. i In the equilibrium domain, the difference between the value and the mean is greater than three times the standard deviation. The basic elements of wind speed inversion are quality controlled to obtain the basic elements of wind speed inversion after quality control.

[0123] The arithmetic mean of the basic elements for wind speed inversion after quality control within the equilibrium domain is:

[0124]

[0125] In the formula, The equilibrium domain is represented by (1.3f). p <f i <3f p The arithmetic mean of the basic elements of wind speed inversion after quality control within the equilibrium domain, where m represents the condition W(f) meets the following criteria after quality control within the equilibrium domain. i The number of ); f i f represents the frequency corresponding to the i-th frequency grid. p Indicates the spectral peak frequencies of the wind and wave components;

[0126] Under the equilibrium domain assumption, wind speed is approximated as a linear function of the arithmetic mean of the basic elements of wind speed inversion, resulting in the following initial model for wind speed inversion:

[0127] U ori =αW+η,

[0128] In the formula, U ori This represents the wind speed retrieved from the initial model for wind speed inversion, where α and η are both undetermined coefficients.

[0129] Considering the high-frequency modulation effect of swell components on wind wave components, the influence of swell components on wind speed inversion is approximated as a linear superposition of the influences of various swell components; the correction amount of each swell component to the initial model of wind speed inversion is regarded as a function of the effective wave height, peak frequency of the swell spectrum, and peak direction of the swell spectrum.

[0130] For each surge component, the degree of influence of the surge component on the high-frequency spectrum of the wind-wave component is proportional to the energy of the surge. Furthermore, the modulation effects of the wind-wave component and the surge component are opposite when their directions are the same or opposite. Therefore, the correction amount for wind speed inversion is expressed as:

[0131]

[0132] In the formula, U Cor This represents the correction amount for wind speed inversion. This represents the effective wave height of the nth swell component. This represents the spectral peak frequency of the nth surge component. θp represents the spectral peak direction of the nth swell component, N represents the total number of swell components, and θp represents the spectral peak direction of the wind wave component; β, γ, κ, ξ and ψ both represent undetermined coefficients. Since the components of the surge are generally considered to be independent of each other and their order of arrangement is not specifically related, the values ​​of the undetermined coefficients are consistent for all surge components.

[0133] Adding the original wind speed inversion model and the correction amount of the wind speed inversion, we obtain the final wind speed inversion model as follows:

[0134]

[0135] In the formula, U represents the wind speed in the final wind speed inversion model, and α and η are both undetermined coefficients. This represents the arithmetic mean of the basic elements of wind speed inversion after quality control within the equilibrium domain.

[0136] Step S4: Based on the parameter data of the two-dimensional spectrum of wind and wave components, establish an initial model for wind direction inversion. Considering the asymmetry of the directional distribution of the two-dimensional spectrum of wind and waves, and the high-frequency modulation effect of the swell component on the wind and wave components, modify the initial model for wind direction inversion to obtain the final wind direction inversion model; including:

[0137] Assuming there is no swell component and the distribution functions of the wind and wave components at different frequencies are symmetrical, the energy weighted average of the average directions of different frequencies in the equilibrium domain of the wind and wave components is taken as the wind direction. The initial model for wind direction inversion obtained by calculating the following formula is:

[0138]

[0139] In the formula, θ W_ori represents the wind direction retrieved from the initial model for wind direction inversion, f i f represents the frequency corresponding to the i-th frequency grid. p S represents the spectral peak frequency of the wind and wave components; f (f i ) represents the energy of different frequencies of the wind and wave components, P represents the undetermined coefficient, and θ io This indicates the average direction of different frequencies of the wind and wave components.

[0140] To prevent phase entanglement during the weighted averaging of the average directions of different frequencies in the equilibrium domain of wind and wave components, the average directions of different frequencies of wind direction and wind and wave components, the main peak direction, and the swell components are all decomposed into meridional and zonal components and then the arctangent is calculated.

[0141] Considering the skewness of the directional distribution of wind and wave components, the directions of the main peaks of different frequencies of wind and wave components are used as the first correction factor for the wind direction model. The initial wind direction inversion model is then corrected using the first correction factor, and the corrected wind direction inversion model is obtained by calculating the following formula:

[0142]

[0143] In the formula, θ W f represents the wind direction retrieved by the corrected wind direction inversion model. i f represents the frequency corresponding to the i-th frequency grid; p S represents the spectral peak frequency of the wind and wave components; f (f i ) represents the different frequency energies of the wind and wave components, θ iM The directions of the main peaks of different frequencies of the wind and wave components are indicated by q, P, and r, which are all undetermined coefficients.

[0144] Further considering the high-frequency modulation effect of swell components on wind wave components, the influence of swell components on wind direction inversion is approximated as a linear superposition of the influences of various swell components.

[0145] The second correction factor for each surge component to the wind direction model is considered as a function of the effective wave height, peak frequency, and peak direction of the surge component. The corrected wind direction inversion model is then further corrected using this second correction factor, and the final wind direction inversion model is obtained by calculating the following formula:

[0146]

[0147] In the formula, θ W_Final θ represents the wind direction retrieved by the final wind direction inversion model. W This indicates the wind direction retrieved by the corrected wind direction inversion model. This represents the effective wave height of the nth swell component. This represents the spectral peak frequency of the nth surge component. The spectral peak direction of the nth surge component is indicated by s, where n represents the nth surge component, N represents the total number of surge components, and s and t are undetermined coefficients.

[0148] Step S5: Determine the coefficients of the final wind speed inversion model and the final wind direction inversion model respectively. Based on the final wind speed inversion model and the final wind direction inversion model, calculate the wind speed and wind direction corresponding to the two-dimensional wave spectrum; including:

[0149] The coefficients in the final wind speed inversion model are determined using dozens to hundreds of independent sets of two-dimensional wave spectra and wind speed synchronous observation data; the coefficients in the final wind direction inversion model are determined using dozens to hundreds of independent sets of two-dimensional wave spectra and wind direction synchronous observation data.

[0150] In the process of determining the undetermined coefficients of the final wind speed inversion model or the final wind direction inversion model, the precision and signal-to-noise ratio of the two-dimensional wave spectrum measured by different devices differ. In order to achieve the best inversion effect, the undetermined coefficients of the final wind speed inversion model or the final wind direction inversion model are determined separately using the two-dimensional wave spectrum measured by two different observation devices.

[0151] If simultaneous observation data of the two-dimensional wave spectrum, wind speed, and wind direction can be obtained, the following loss function can be minimized to jointly determine the undetermined coefficients in the final wind speed inversion model and the final wind direction inversion model. The loss function is:

[0152]

[0153] In the formula, LOSS represents the loss function, θ W_Final This represents the final wind direction inversion model, where M represents the total number of samples, the subscript obs indicates the actual observed wind speed or direction, and θ... W_Final,i' θ represents the wind direction retrieved by the final wind direction inversion model in the i'th sample. W_obs,i' U represents the actual observed wind direction in the i'-th sample. i' U represents the wind speed retrieved by the final wind speed inversion model in the i'th sample. obs,i' This represents the actual wind speed observed in the i'th sample.

[0154] Once all the aforementioned undetermined coefficients are determined, under the condition of no sea surface wind speed and direction observation, the final wind speed inversion model U and the final wind direction inversion model θ are used. W_Final The wind speed and direction can be calculated from the two-dimensional spectrum of the ocean waves.

[0155] Based on the same method, a system for inverting sea surface wind fields using two-dimensional wave spectral observation data was finally disclosed, such as... Figure 2As shown, it includes: a processing unit, a calculation unit, a wind speed inversion unit, a wind direction inversion unit, and an integration unit.

[0156] The processing unit is used to perform spectral smoothing and spectral segmentation on the observed two-dimensional wave spectrum in sequence to obtain two-dimensional spectra of wind wave components and several two-dimensional spectra of swell components, respectively.

[0157] The calculation unit is used to calculate the parameter data of the two-dimensional spectrum of wind wave component and the parameter data of the two-dimensional spectrum of each swell component, respectively.

[0158] The wind speed inversion unit is used to construct basic elements for wind speed inversion based on the parameter data of the two-dimensional spectrum of wind and wave components, and obtain an initial wind speed inversion model based on the basic elements of wind speed inversion. Based on the parameter data of the two-dimensional spectrum of each surge component, considering the high-frequency modulation effect of the surge component on the wind and wave component, the initial wind speed inversion model is corrected to obtain the final wind speed inversion model.

[0159] The wind direction inversion unit is used to establish an initial wind direction inversion model based on the parameter data of the two-dimensional spectrum of wind and wave components. The initial wind direction inversion model is modified according to the asymmetry of the directional distribution of the two-dimensional spectrum of wind and wave components and the high-frequency modulation effect of the swell components on the wind and wave components, so as to obtain the final wind direction inversion model.

[0160] The integration unit is used to determine the coefficients of the final wind speed inversion model and the final wind direction inversion model, respectively, and calculate the wind speed and wind direction corresponding to the two-dimensional wave spectrum based on the final wind speed inversion model and the final wind direction inversion model.

[0161] Regarding the system in the above embodiments, the specific manner in which each unit module performs operations has been described in detail in the embodiments related to the method, and will not be elaborated here.

[0162] Based on the same inventive concept, embodiments of the present invention also provide an electronic device, the structure of which is as follows: Figure 3 As shown, it includes: a memory and a processor, wherein the processor is used to read and execute the computer program stored in the memory to implement the aforementioned method for inverting sea surface wind field based on two-dimensional wave spectrum observation data.

[0163] Based on the same inventive concept, embodiments of the present invention also provide a computer storage medium storing computer-executable instructions, which, when executed, implement the aforementioned method for inverting sea surface wind field based on two-dimensional wave spectrum observation data.

[0164] This invention considers the high-frequency modulation effect of swell components on wind wave components by adding a correction term, multiplying by the wave span and f. 4The parameter construction takes into account the influence of wind speed on wave spread, and the stability of the parameter is used for data quality control, achieving accurate modeling of wind speed inversion. At the same time, this invention selects the optimal wind direction estimation frequency band by high-order weighting of energy in the wind direction sensitive area. By considering the asymmetry of direction distribution (the average direction and the main peak direction are inconsistent) and the modulation of the high-frequency wave two-dimensional spectrum by the swell component, a second-order correction of wind direction is made, which can reduce the error of wind speed and wind direction inversion by nearly 40% (wind speed reaches 1.3m / s, wind speed greater than 7m / s wind direction within 15°), achieving the data accuracy required for operational observation.

[0165] Furthermore, the physical meaning of the model in this invention is clearer, and there are fewer undetermined coefficients; the number of training data elements required for calibration is smaller, and the transferability is relatively better, avoiding the overfitting or underfitting problems of neural networks in existing technologies. At the same time, as an analytical empirical formula, it can be easily applied to various low-cost hardware platforms with insufficient computing and storage performance.

[0166] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for inverting sea surface wind field based on two-dimensional wave spectrum observation data, characterized in that, include: The observed two-dimensional wave spectrum was sequentially processed by spectral smoothing and spectral segmentation to obtain two-dimensional spectra of wind wave components and several two-dimensional spectra of swell components, respectively. Calculate the parameter data of the two-dimensional spectrum of wind wave component and the parameter data of the two-dimensional spectrum of each swell component separately; Based on the parameter data of the two-dimensional spectrum of wind and wave components, basic elements for wind speed inversion are constructed, and an initial wind speed inversion model is obtained based on the basic elements for wind speed inversion. Based on the parameter data of the two-dimensional spectrum of each surge component, considering the modulation effect of the surge component on the high frequency of wind and waves, the initial wind speed inversion model is modified to obtain the final wind speed inversion model. An initial wind direction inversion model is established based on parameter data of the two-dimensional spectrum of wind and wave components. Considering the asymmetry of the directional distribution of the two-dimensional spectrum of wind and wave components and the high-frequency modulation effect of swell components on wind and wave components, the initial wind direction inversion model is modified to obtain the final wind direction inversion model, which more specifically includes: Assuming there is no swell component and the distribution functions of different frequencies of the wind and wave components are symmetrical, the energy weighted average of the average directions of different frequencies in the equilibrium domain of the wind and wave components is regarded as the wind direction, and an initial wind direction inversion model is obtained. Considering the skewness of the two-dimensional spectral distribution of wind and wave components, the main peak directions of different frequencies of wind and wave components are used as the first correction factor for the wind direction model. The initial wind direction inversion model is then corrected using the first correction factor to obtain the corrected wind direction inversion model. Considering the high-frequency modulation effect of swell components on wind wave components, the influence of swell components on wind direction inversion is approximated as a linear superposition of the influences of various swell components. The second correction of each swell component to the wind direction model is regarded as a function of the effective wave height, swell spectrum peak frequency and swell spectrum peak direction of the swell component; the second correction is used to perform a second correction on the corrected wind direction inversion model to obtain the final wind direction inversion model. The coefficients of the final wind speed inversion model and the final wind direction inversion model are determined respectively. Based on the final wind speed inversion model and the final wind direction inversion model, the wind speed and wind direction corresponding to the two-dimensional wave spectrum are calculated.

2. The method for inverting sea surface wind field based on two-dimensional wave spectrum observation data according to claim 1, characterized in that, The parameter data of the two-dimensional spectrum of the wind and wave components include the energy at different frequencies, the average direction at different frequencies, the main peak direction at different frequencies, the wave spread at different frequencies, and the spectral peak frequency and spectral peak direction of the wind and wave components. The calculation of the parameter data of the two-dimensional spectrum of the wind and wave components includes the following steps: The two-dimensional spectrum of wind and wave components is set as follows: The number of grids for frequency and direction in the two-dimensional spectrum of wind and wave components are N, respectively. f and N θ And the two-dimensional spectrum of wind and wave components Decomposed into the product of energy at different frequencies and distribution functions in different frequency directions: ,in This represents the distribution function of different frequency directions of the wind and wave components. Represents the different frequencies of energy in the components of wind and waves; Based on the two-dimensional spectrum of wind and wave components The frequency and direction corresponding to the maximum value are used to obtain the spectral peak frequencies of the wind and wave components. f p Spectral peak direction of wind and wave components θ p ; The energy of different frequencies of the wind and wave components is calculated using the following formula: ， In the formula, This represents the different frequencies of energy in the components of wind and waves. This represents a two-dimensional spectrum of wind and wave components. This represents the frequency corresponding to the i-th frequency grid. This represents the direction corresponding to the j-th direction grid. This indicates the angular resolution of the wave direction spectrum. Indicates the grid number representing the direction; The distribution function of wind and wave components in different frequency directions is calculated using the following formula: ， In the formula, This represents the distribution function of different frequency directions of the wind and wave components. This indicates the frequency resolution of the ocean wave directional spectrum. Indicates the frequency grid number; Main peak directions of different frequencies of wind and wave components θ iM Let be the direction corresponding to the maximum value of the directional distribution function of the wind and wave components at different frequencies. The average direction of the wind and wave components at different frequencies is calculated based on the directional distribution function of the wind and wave components at different frequencies: ， In the formula, This indicates the average direction of different frequencies of the wind and wave components. and The intermediate variables representing the wave span and wave direction calculations are expressed by the following formula: , ; The wave spans of different frequencies of wind wave components are calculated using the following formula: ， In the formula, This represents the wave spread of different frequencies of the wind and wave components.

3. The method for inverting sea surface wind field based on two-dimensional wave spectrum observation data according to claim 1, characterized in that, The parameter data of the two-dimensional spectrum of the surge component includes the effective wave height, peak frequency, and peak direction of the surge component. Calculating the parameter data of the two-dimensional spectrum of each surge component includes the following steps: Assuming that the two-dimensional wave spectrum yields N swell components through spectral segmentation, then the two-dimensional spectrum of the nth swell component is defined as follows: The number of grids for frequency and direction in the two-dimensional spectrum of swell components are N, respectively. f and N θ Where 1 ≤ n ≤ N; The effective wave height of the two-dimensional spectrum of the nth surge component is calculated based on the number of grids for frequency and direction. ; Based on the two-dimensional spectrum of the nth surge component The frequency and direction of the surge spectrum peak are obtained by determining the frequency and direction of the maximum value of the peak. The effective wave height of the nth swell component is calculated using the following formula: ， In the formula, This represents the effective wave height of the nth swell component. This represents the two-dimensional spectrum of the nth surge component. This represents the frequency corresponding to the i-th frequency grid. Indicates the frequency grid number; This represents the direction corresponding to the j-th direction grid. Indicates the grid number representing the direction; This indicates the angular resolution of the wave direction spectrum. This indicates the frequency resolution of the ocean wave directional spectrum.

4. A method for inverting sea surface wind field based on two-dimensional wave spectrum observation data according to claim 1, comprising constructing basic elements for wind speed inversion based on parameter data of the two-dimensional spectrum of wind and wave components, and obtaining an initial wind speed inversion model based on the basic elements for wind speed inversion, including: Selecting different frequency energies of wind and wave components in the equilibrium domain Based on the proportional relationship between energy and frequency to the power of -4 in the equilibrium domain, and considering the wave spread of different frequencies of wind and wave components, the basic elements for wind speed inversion are constructed. Based on the equilibrium domain theory, after approximating the basic elements of wind speed inversion as a constant, the basic elements of wind speed inversion are subjected to quality control to obtain the quality-controlled basic elements of wind speed inversion. Calculate the arithmetic mean of the basic elements of wind speed inversion after quality control within the equilibrium domain; In the equilibrium domain, the wind speed is approximated as a linear function of the arithmetic mean of the basic elements of the wind speed inversion, thus obtaining the initial model for wind speed inversion. The basic elements for wind speed inversion are calculated using the following formula: ， In the formula, The basic elements representing wind speed inversion at different frequencies. Represents the wave spans of different frequencies of wind and wave components. i This represents the frequency corresponding to the i-th frequency grid. The arithmetic mean of the basic elements of the wind speed inversion after mass control within the equilibrium domain is calculated using the following formula: ， In the formula, This represents the arithmetic mean of the basic elements used in the wind speed inversion. In the equilibrium domain After intermediate quality control, it meets the requirements. The number of; This represents the frequency corresponding to the i-th frequency grid. Indicates the spectral peak frequencies of the wind and wave components; The initial model for wind speed inversion is: ， In the formula, This represents the wind speed retrieved from the initial model for wind speed inversion. and All of these represent coefficients to be determined.

5. A method for inverting sea surface wind field based on two-dimensional wave spectrum observation data according to claim 1 or 4, comprising: based on the parameter data of the two-dimensional wave spectrum, considering the high-frequency modulation effect of wave components on wind wave components, modifying the initial wind speed inversion model to obtain the final wind speed inversion model, including: Considering the high-frequency modulation effect of swell components on wind wave components, the influence of swell components on wind speed inversion is approximated as a linear superposition of the influences of various swell components. The corrections made by each wave component to the initial model for wind speed inversion are considered as functions of the effective wave height, wave peak frequency, and wave peak direction of the wave component. Based on the relationship between the influence of the swell component on the high-frequency spectrum of the wind wave component and the energy of the swell, as well as the influence of whether the wind wave component and the swell component are aligned on the modulation effect, the wind speed inversion correction amount corresponding to each swell component is obtained. The initial wind speed inversion model is corrected based on the correction amount of the wind speed inversion to obtain the final wind speed inversion model. The correction amount for wind speed inversion is calculated using the following formula: ， In the formula, This represents the correction amount for wind speed inversion. Indicates the first The effective wave height of each swell component Indicates the first The spectral peak frequencies of the surge component, Indicates the first The direction of the spectral peaks of the surge component. This indicates the total number of swell components. Indicates the direction of the spectral peaks of the wind and wave components. All represent undetermined coefficients; Adding the original wind speed inversion model and the correction amount of the wind speed inversion, we obtain the final wind speed inversion model as follows: ， In the formula, This represents the wind speed in the final wind speed inversion model. and All represent undetermined coefficients. This represents the arithmetic mean of the basic elements of wind speed inversion after quality control within the equilibrium domain.

6. The method for inverting sea surface wind field based on two-dimensional wave spectrum observation data according to claim 1, The initial model for wind direction inversion is: ， In the formula, This indicates the wind direction retrieved from the initial model for wind direction inversion. This represents the frequency corresponding to the i-th frequency grid. Indicates the spectral peak frequencies of the wind and wave components; This represents the different frequencies of energy in the components of wind and waves. P Denotes undetermined coefficients. Indicates the average direction of different frequencies of wind and wave components; The corrected wind direction inversion model is as follows: ， In the formula, This indicates the wind direction retrieved by the corrected wind direction inversion model. Indicates the direction of the main peak of different frequencies of wind and wave components, r and All represent undetermined coefficients; The final wind direction inversion model is as follows: In the formula, This indicates the wind direction retrieved by the final wind direction inversion model. This indicates the wind direction retrieved by the corrected wind direction inversion model. Indicates the first The effective wave height of each swell component Indicates the first The spectral peak frequencies of the surge component, Indicates the first The spectral peak direction of the surge component, Indicates the first One surging component, This indicates the total number of swell components. and This represents an undetermined coefficient.

7. A system for retrieving sea surface wind field based on two-dimensional wave spectrum observation data, comprising: The processing unit is used to perform spectral smoothing and spectral segmentation on the observed two-dimensional wave spectrum in sequence to obtain two-dimensional spectra of wind wave components and several two-dimensional spectra of swell components, respectively. The calculation unit is used to calculate the parameter data of the two-dimensional spectrum of wind wave component and the parameter data of the two-dimensional spectrum of each swell component, respectively. The wind speed inversion unit is used to construct basic elements for wind speed inversion based on the parameter data of the two-dimensional spectrum of wind and wave components, and obtain an initial wind speed inversion model based on the basic elements of wind speed inversion. Based on the parameter data of the two-dimensional spectrum of each surge component, considering the high-frequency modulation effect of the surge component on the wind and wave component, the initial wind speed inversion model is corrected to obtain the final wind speed inversion model. The wind direction inversion unit is used to establish an initial wind direction inversion model based on parameter data of the two-dimensional spectrum of wind and wave components. Taking into account the asymmetry of the directional distribution of the two-dimensional spectrum of wind and wave components and the high-frequency modulation effect of swell components on wind and wave components, the initial wind direction inversion model is modified to obtain a final wind direction inversion model, which more specifically includes: Assuming there is no swell component and the distribution functions of different frequencies of the wind and wave components are symmetrical, the energy weighted average of the average directions of different frequencies in the equilibrium domain of the wind and wave components is regarded as the wind direction, and an initial wind direction inversion model is obtained. Considering the skewness of the two-dimensional spectral distribution of wind and wave components, the main peak directions of different frequencies of wind and wave components are used as the first correction factor for the wind direction model. The initial wind direction inversion model is then corrected using the first correction factor to obtain the corrected wind direction inversion model. Considering the high-frequency modulation effect of swell components on wind wave components, the influence of swell components on wind direction inversion is approximated as a linear superposition of the influences of various swell components. The second correction of each swell component to the wind direction model is regarded as a function of the effective wave height, swell spectrum peak frequency and swell spectrum peak direction of the swell component; the second correction is used to perform a second correction on the corrected wind direction inversion model to obtain the final wind direction inversion model. The integration unit is used to determine the coefficients of the final wind speed inversion model and the final wind direction inversion model, respectively, and calculate the wind speed and wind direction corresponding to the two-dimensional wave spectrum based on the final wind speed inversion model and the final wind direction inversion model.

8. An electronic device, characterized in that, include: Memory, processor; The processor is used to read and execute the computer program stored in the memory to implement the method for inverting sea surface wind field based on two-dimensional wave spectrum observation data as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed, implement the method for inverting sea surface wind field based on two-dimensional wave spectrum observation data as described in any one of claims 1-6.

Images