Sea surface wind speed inversion method based on wind wave average period

Abstract

The invention belongs to the technical field of microwave radar image processing, and relates to a sea surface wind speed inversion method based on a wind wave average period, which comprises the following steps: constructing a sea surface wind speed calculation model taking the wind wave average period as an input parameter; secondly, acquiring a sea surface microwave remote sensing image, and performing preprocessing operations such as coordinate transformation and noise suppression on the sea surface microwave remote sensing image; then, performing inversion from the preprocessed image by adopting a spectrum analysis technology to obtain a wind wave average period; and finally, the obtained wind wave average period is substituted into the preset wind speed calculation model, and the sea surface wind speed of the target area is obtained through calculation. According to the method, the wind wave average period is used as a key intermediate variable, deep coupling of image features and wind wave physical parameters is achieved, and the defects of an existing direct inversion method in the aspects of precision and stability are effectively overcome; finally, a sea surface wind speed acquisition scheme which is clear in physical significance, excellent in inversion precision and convenient for business application and popularization is provided.

Description

Technical Field

[0001] This invention belongs to the field of microwave radar image processing technology, specifically relating to a method for inverting sea surface wind speed based on the average period of wind and waves, which can be used to invert sea surface wind speed. Background Technology

[0002] Sea surface wind speed, as one of the key parameters of the marine environment, plays an irreplaceable role in many fields such as marine transportation safety, offshore oil and gas development, meteorological disaster early warning, ocean circulation simulation, and global climate research. Accurate and efficient acquisition of sea surface wind speed data, especially large-scale, high spatiotemporal resolution wind speed information, is of great practical significance for ensuring the safety of maritime operations, improving meteorological disaster prevention and control capabilities, and promoting the development of marine scientific research.

[0003] Traditional sea surface wind speed observations primarily rely on platforms such as ocean buoys, survey vessels, and coastal meteorological stations. While these methods can acquire high-precision measured data within fixed locations or along shipping routes, they suffer from inherent limitations such as limited observation range, uneven spatial coverage, and high costs, making them unsuitable for large-scale, real-time observations. This limitation is particularly pronounced in remote sea areas such as the open ocean and polar regions. With the rapid development of remote sensing technology, microwave remote sensing, with its all-weather, all-day operation and wide-area coverage, has become the mainstream technology for sea surface wind speed retrieval. Among these, image data acquired by microwave remote sensing equipment such as synthetic aperture radar, shore-based sea observation radar, and shipborne navigation radar are widely used in sea surface wind speed retrieval research due to their high spatial resolution.

[0004] Currently, sea surface wind speed inversion methods based on microwave remote sensing images are mainly divided into two categories: direct inversion methods and indirect inversion methods. Among them, the direct inversion method is the most widely used technical approach. Its core principle is to use the normalized radar backscattering coefficient (NRCS) obtained from radar observations, combined with a preset geophysical model function (GMF), to directly establish the mapping relationship between NRCS and sea surface wind speed, and then solve for the wind speed. Commonly used GMF models include the CMOD series and MS1A, but these methods have significant limitations: on the one hand, NRCS is not only affected by sea surface wind speed, but is also easily interfered with by various factors such as the observation incident angle, wind direction, and sea state type (mixed wind waves and swells), which leads to a significant decrease in the applicability of the model under complex sea conditions, and the inversion accuracy is difficult to guarantee; on the other hand, the direct inversion method is sensitive to image noise (such as the inherent speckle noise of microwave remote sensing images). Even after simple noise suppression processing, it may still introduce large inversion errors, especially in areas with complex sea conditions such as nearshore and polar regions, where the errors are more obvious.

[0005] Indirect inversion methods primarily construct a correlation model between indirect parameters such as wave texture and spectral features from remote sensing images and sea surface wind speed. Existing indirect inversion methods often rely on single image features or empirical fitting models, leading to issues such as unclear physical meaning and insufficient stability. Some methods require additional prior information such as wind direction and sea temperature, increasing operational complexity and data acquisition costs. Other methods employ purely empirical fitting models, lacking support from wave physical parameters, resulting in poor model generalization ability and significant fluctuations in inversion results across different sea areas and sea states. Still other methods suffer from inadequate image preprocessing; for example, they fail to perform accurate coordinate transformation for the polar coordinate characteristics of shore-based and shipborne radar images, or employ limited noise suppression methods, resulting in insufficient accuracy in subsequent wave feature parameter extraction and consequently affecting wind speed inversion results.

[0006] Furthermore, existing inversion methods generally suffer from poor adaptability to mixed wind and wave scenarios. In real marine environments, wind and waves often coexist on the sea surface, and their spectral characteristics overlap, making it difficult for direct inversion methods to accurately distinguish effective signals, and indirect inversion methods to accurately extract core features related to wind speed. At the same time, most existing methods do not fully utilize the close physical relationship between sea surface wind speed and the average period of wind and waves. According to wave theory, the average period of wind and waves, as a key physical parameter reflecting the energy propagation characteristics of ocean waves, has a stable statistical relationship with sea surface wind speed and is less affected by external interference factors. However, existing technologies have failed to effectively integrate this physical relationship into the inversion model, making it difficult to improve inversion accuracy and stability.

[0007] In summary, existing sea surface wind speed inversion methods suffer from limitations such as limited accuracy, insufficient stability, reliance on prior information, unclear physical meaning, and difficulty in operational application, making it difficult to meet the practical needs of high-precision, large-scale sea surface wind speed observation under complex sea conditions. Therefore, developing a sea surface wind speed inversion method with clear physical meaning, high inversion accuracy, strong anti-interference capabilities, and ease of operational application has become an urgent technical problem to be solved in this field. Summary of the Invention

[0008] The purpose of this invention is to fully explore and utilize the close relationship between sea surface wind speed and the average period of wind and waves, establish a functional model between sea surface wind speed and the average period of wind and waves, and provide a sea surface wind speed inversion method based on the average period of wind and waves. This addresses the problems of existing sea surface wind speed inversion methods, such as limited accuracy, insufficient stability, reliance on prior information, and unclear physical meaning, which make it difficult to meet the actual needs of high-precision and large-scale sea surface wind speed observation under complex sea conditions. This invention simplifies the inversion of sea surface wind speed based on microwave remote sensing images.

[0009] This invention is achieved through the following technical solution:

[0010] A method for inverting sea surface wind speed based on the average period of wind and waves includes the following steps:

[0011] S1. Construct a sea surface wind speed inversion model:

[0012] Establish a sea surface wind speed inversion model with the average period of wind and waves as the core input parameter;

[0013] S2. Acquire microwave remote sensing images of the sea area to be measured:

[0014] Microwave remote sensing images of the sea surface of the area to be measured are acquired through satellite remote sensing platforms, airborne remote sensing platforms, as well as shore-based sea observation radars and shipborne navigation radars.

[0015] S3. Image preprocessing operations:

[0016] For the acquired shore-based sea observation radar images and shipborne navigation radar images, coordinate transformation operations are performed to convert the polar coordinates of image pixels into spatial rectangular coordinates; median filtering technology is used to suppress speckle noise in the sea surface microwave remote sensing images;

[0017] S4. Inversion to obtain the average period of wind and waves:

[0018] Perform a Fourier transform on the preprocessed remote sensing image to obtain a two-dimensional image spectrum of the sea surface. Extract the wave number corresponding to the wind and waves based on this two-dimensional spectrum, and then solve the average period of the wind and waves based on the dispersion relation. The expression is as follows:

[0019] ;

[0020] Where k is the wave number corresponding to the wind and waves in the image spectrum, and g is the gravitational acceleration, typically taken as 9.81m. 2 / s, where d is the seawater depth; when the seawater depth is greater than half the wavelength of the waves, the expression for calculating the average period of wind and waves based on the dispersion relation can be simplified as follows:

[0021] ;

[0022] S5, Inverted sea surface wind speed:

[0023] The average period T of wind and waves obtained from step S4 m Substituting the values ​​into the sea surface wind speed inversion model preset in step S1, the final sea surface wind speed U is obtained through calculation. 10 .

[0024] Further, in step S1, the statistical characteristics of the wave spectrum in wave theory are inverted based on the model, and the expression is:

[0025] ;

[0026] Among them, U10 T represents sea surface wind speed. m Let f be the average period of wind and waves, and f be a function constructed based on existing publicly available data, including the European Centre for Medium-Range Weather Forecasts and the National Data Buoy Centre, which matches the average period of wind and waves with sea surface wind speed.

[0027] First, a power function was selected as the model form. Then, sea surface wind speed and mean wave period observation data from the European Centre for Medium-Range Weather Forecasts (ECMWF) in 2020 were used for fitting. Finally, a power function relationship model between sea surface wind speed and mean wave period was established, with the specific expression as follows:

[0028] ;

[0029] Where a and b are the model fitting parameters, and a = 2.38 and b = 0.87.

[0030] Further, step S3, converting the polar coordinates of the image pixels into spatial rectangular coordinates, specifically includes:

[0031] For shore-based sea observation radar images and shipborne navigation radar images, the size of the polar coordinate image is set to H×W, where H represents the image height and W represents the image width; the polar angle φ ranges from [0, 2π), corresponding one-to-one with the pixels of the image height H, where the polar angle corresponding to the i-th row of pixels is:

[0032] ;

[0033] The range of the polar radius r is [0, R]. max ], R max The maximum radius corresponds one-to-one with the pixels of the image width W, where the radius corresponding to the j-th column pixel is:

[0034] ;

[0035] After converting the polar coordinates of each pixel to rectangular coordinates, the corresponding rectangular coordinates (x, y, y) of each pixel are... ij ,y ij )for:

[0036] .

[0037] Further, step S4, retrieving the average period of wind and waves, specifically includes:

[0038] Performing a two-dimensional Fourier transform on the preprocessed sea surface microwave remote sensing image, which has been converted to spatial rectangular coordinates, can obtain the energy distribution characteristics of the sea surface radar image in the two-dimensional wavenumber domain. Its continuous form expression is as follows:

[0039] ;

[0040] Among them, L x L is the length of the image used in the x-direction (range direction). y Let I(x,y) be the length of the image in the y-direction (or azimuth direction), and k be the gray value at position (x,y) in the image. x Let k be the wave number in the x-direction. y Let be the wave number in the y-direction;

[0041] Using the discrete Fourier transform form, the specific expression is as follows:

[0042] ;

[0043] For γ(k) after discrete Fourier transform x ,k y Subsequent processing is performed to obtain the true two-dimensional wave image spectrum I(k). x ,k y ):

[0044] ;

[0045] If two distinct spectral peaks exist in a two-dimensional image spectrum, the peak with the larger wavenumber usually corresponds to wind and waves. Based on this, the two-dimensional wavenumber (k) corresponding to the wind and wave spectral peak is obtained. x ,k y Then, the average period of wind and waves is retrieved using this two-dimensional wavenumber inversion method. The specific calculation method is as follows:

[0046] ;

[0047] If there is only one obvious spectral peak in the two-dimensional image spectrum, the spectral peak may correspond to large-scale swells or small-scale wind waves; if the wave number corresponding to the spectral peak is greater than 0.1 rad / m, then the spectral peak is determined to correspond to wind waves, and the average period of wind waves is calculated based on the wave number corresponding to the spectral peak.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] This invention uses the average period of wind and waves as a key intermediate variable to achieve deep coupling between image features and wind and wave physical parameters, effectively solving the shortcomings of existing direct inversion methods in terms of accuracy and stability. Ultimately, it provides a sea surface wind speed acquisition scheme with clear physical meaning, excellent inversion accuracy, and easy application in business. Attached Figure Description

[0050] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 A flowchart of a sea surface wind speed inversion method based on the average period of wind and waves provided by the present invention;

[0052] Figure 2 A graph showing the relationship between sea surface wind speed and average wave period provided by this invention;

[0053] Figure 3 This invention uses a scatter plot of sea surface wind speed derived from the 2021 mean wind and wave period provided by the European Medium-Range Weather Forecasts (EMFR) and a reference sea surface wind speed.

[0054] Figure 4 This is a scatter plot showing the relationship between the sea surface wind speed derived from the 2022 wind and wave mean period provided by the European Medium-Range Weather Forecast and the reference sea surface wind speed, which is used in this invention.

[0055] Figure 5 The correlation coefficient between the sea surface wind speed derived from the average wind and wave period at some stations of the National Data Buoy Center in 2020 and the reference sea surface wind speed is used in this invention.

[0056] Figure 6 This invention uses the sea surface wind speed derived from the average wind and wave period at some stations of the National Data Buoy Center in 2020 to determine the absolute deviation between the sea surface wind speed and the reference sea surface wind speed.

[0057] Figure 7 This invention uses the average deviation between the sea surface wind speed derived from the average wind and wave period at some stations of the National Data Buoy Center in 2020 and the reference sea surface wind speed. Detailed Implementation

[0058] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0059] This invention provides a method for inverting sea surface wind speed based on the average period of wind and waves. The specific steps are as follows: First, a sea surface wind speed calculation model is constructed using the average period of wind and waves as input parameters; second, microwave remote sensing images of the sea surface are acquired, including synthetic aperture radar images from aerial platforms, shore-based sea observation radar images, and shipborne navigation radar images, and preprocessing operations such as coordinate transformation and noise suppression are performed on these images; subsequently, the average period of wind and waves is inverted from the preprocessed images using spectral analysis techniques; finally, the obtained average period of wind and waves is substituted into the aforementioned preset wind speed calculation model to calculate the sea surface wind speed in the target area. Figure 1 As shown, the specific steps include:

[0060] S1. Construct a sea surface wind speed inversion model:

[0061] A sea surface wind speed inversion model is established with the average period of wind and waves as the core input parameter. This model is based on the statistical characteristics of the wave spectrum in wave theory, and its core expression is:

[0062] ;

[0063] Among them, U 10 T represents sea surface wind speed. m Let f be the average period of wind and waves, and let f be the function relationship constructed based on existing publicly available data, including the European Centre for Medium-Range Weather Forecasts and the National Data Buoy Centre, which matches the average period of wind and waves with sea surface wind speed.

[0064] Establish a functional model between sea surface wind speed and the average period of wind waves, specifically including:

[0065] A power function was chosen as the model form. Sea surface wind speed and mean wave period observation data from the European Centre for Medium-Range Weather Forecasts (ECMWF) in 2020 were used for fitting. Finally, a power function relationship model between sea surface wind speed and mean wave period was established, with the specific expression as follows:

[0066] ;

[0067] Where a and b are model fitting parameters, and a = 2.38, b = 0.87. The functional model between sea surface wind speed and the average period of wind and waves is as follows: Figure 2 As shown by the black line in the image.

[0068] S2. Acquire microwave remote sensing images of the sea area to be measured:

[0069] Microwave remote sensing images of the sea surface of the area to be measured are acquired through satellite remote sensing platforms, airborne remote sensing platforms, as well as shore-based sea observation radars and shipborne navigation radars.

[0070] S3. Image preprocessing operations:

[0071] Image preprocessing is fundamental to ensuring the accuracy of wind and wave mean period inversion, and includes two main steps: coordinate transformation and image noise suppression. For the acquired shore-based sea observation radar images and shipborne navigation radar images, coordinate transformation is performed to convert the polar coordinates of image pixels to spatial rectangular coordinates, facilitating subsequent noise suppression and extraction of wind and wave mean period parameters. Median filtering is used to suppress speckle noise in the microwave remote sensing images of the sea surface, improving the recognizability of wind and wave texture features, thereby ensuring the accuracy of wind and wave mean period extraction.

[0072] Converting image pixel polar coordinates to Cartesian coordinates specifically includes:

[0073] For shore-based sea observation radar images and shipborne navigation radar images, the size of the polar coordinate image is set to H×W (H represents the image height, W represents the image width); the polar angle φ ranges from [0, 2π), corresponding one-to-one with the pixels of the image height H, where the polar angle corresponding to the pixel in the i-th row is:

[0074] ;

[0075] The range of the polar radius r is [0, R]. max ](R max (where is the maximum polar radius), and each pixel corresponds one-to-one with the image width W, where the polar radius corresponding to the j-th column pixel is:

[0076] ;

[0077] After converting the polar coordinates of each pixel to rectangular coordinates, the corresponding rectangular coordinates (x, y, y) of each pixel are... ij ,y ij )for:

[0078] .

[0079] S4. Inversion to obtain the average period of wind and waves:

[0080] Perform a Fourier transform on the preprocessed remote sensing image to obtain a two-dimensional image spectrum of the sea surface. Extract the wave number corresponding to the wind and waves based on this two-dimensional spectrum, and then solve the average period of the wind and waves based on the dispersion relation. The expression is as follows:

[0081] ;

[0082] Where k is the wave number corresponding to the wind and waves in the image spectrum, and g is the gravitational acceleration, typically taken as 9.81m. 2 / s, where d is the seawater depth; when the seawater depth is greater than half the wavelength of the waves, the expression for calculating the average period of wind and waves based on the dispersion relation can be simplified as follows:

[0083] ;

[0084] The inverted mean period of wind and waves specifically includes:

[0085] Performing a two-dimensional Fourier transform on the preprocessed sea surface microwave remote sensing image, which has been converted to spatial rectangular coordinates, can obtain the energy distribution characteristics of the sea surface radar image in the two-dimensional wavenumber domain. Its continuous form expression is as follows:

[0086] ;

[0087] Among them, L x L is the length of the image used in the x-direction (range direction). y Let I(x,y) be the length of the image in the y-direction (azimuth direction), and k be the gray value at position (x,y) in the image. x Let k be the wave number in the x-direction. y Let be the wave number in the y-direction; since real radar images are discrete data, the continuous Fourier transform is no longer applicable. Therefore, the discrete Fourier transform form must be used, and the specific expression is:

[0088] ;

[0089] For γ(k) after discrete Fourier transform x ,k y After further processing, the true two-dimensional wave image spectrum I(k) can be obtained. x ,k y ).

[0090] ;

[0091] If two distinct spectral peaks exist in a two-dimensional image spectrum, the peak with the larger wavenumber usually corresponds to wind and waves. Based on this, the two-dimensional wavenumber (k) corresponding to the wind and wave spectral peak is obtained. x ,k y Then, the average period of wind and waves is retrieved using this two-dimensional wavenumber inversion method. The specific calculation method is as follows:

[0092] ;

[0093] If there is only one obvious spectral peak in the two-dimensional image spectrum, the spectral peak may correspond to large-scale swells or small-scale wind waves; if the wave number corresponding to the spectral peak is large (usually set to be greater than 0.1 rad / m), then the spectral peak is determined to correspond to wind waves, and then the average period of wind waves is calculated based on the wave number corresponding to the spectral peak.

[0094] S5, Inverted sea surface wind speed:

[0095] The average period T of wind and waves obtained from step S4 mSubstituting the values ​​into the sea surface wind speed inversion model preset in step S1, the final sea surface wind speed U is obtained through calculation. 10 .

[0096] To demonstrate the effectiveness of this invention, the average wind and wave periods for 2021 and 2022 provided by the European Medium-Range Weather Forecast (ECMWF) were used as the model input to retrieve sea surface wind speeds, and the accuracy of the retrieval results was evaluated. The sea surface wind speed retrieval results are as follows: Figure 3 and Figure 4 As shown, the present invention utilizes the average period of wind and waves provided by the European Medium-Range Weather Forecast to invert the accuracy of sea surface wind speed. The inversion accuracy is shown in Table 1.

[0097] In addition, this invention also used the average wind and wave period at some stations of the National Data Buoy Center in 2020 to invert sea surface wind speed, and compared it with the reference wind speed provided by the National Data Buoy Center. The inversion accuracy is as follows: Figure 5 , Figure 6 and Figure 7 As shown in Table 1 and Figures 3-7 It can be seen that the deviation between the sea surface wind speed inverted using the model established by this invention and the reference wind speed is small, indicating high accuracy in sea surface wind speed inversion and demonstrating the good inventive effect of this invention. This invention can directly obtain sea surface wind speed from the average period of wind and waves; the entire inversion process is relatively simple and easy to implement, thus helping to improve the efficiency of sea surface wind speed inversion.

[0098] Table 1

[0099] years Correlation coefficient Root mean square error / meter 2021 0.96 0.82 2022 0.97 0.83

[0100] It will be understood by those skilled in the art that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for retrieving sea surface wind speed based on the average period of wind waves, characterized in that, The method comprises the following steps: S1, constructing a sea surface wind speed inversion model: S2, collecting microwave remote sensing images of the sea area to be measured: Through satellite remote sensing platforms, aerial remote sensing platforms, and shore-based marine observation radars, shipborne navigation radars and other equipment, microwave remote sensing images of the sea area to be measured are obtained; S3, image preprocessing operation: For the obtained shore-based marine observation radar images and shipborne navigation radar images, coordinate conversion operation is performed to convert the polar coordinates of the image pixels into spatial rectangular coordinates; the median filtering technique is used to suppress the coherent speckle noise in the sea surface microwave remote sensing images; S4, inversion of the average period of wind waves: The Fourier transform is performed on the preprocessed remote sensing images to obtain a two-dimensional image spectrum of the sea surface, the wave number corresponding to the wind waves is extracted according to the two-dimensional spectrum, and the average period of the wind waves is solved according to the dispersion relationship, which is expressed as: S5, inversion of the sea surface wind speed: ； where k is the wave number corresponding to the wind wave in the image spectrum, g is the gravity acceleration, usually taking the value of 9.81 m / s, d is the depth of seawater; when the depth of seawater is greater than half of the wavelength of the sea wave, the expression for calculating the average period of the wind wave based on the dispersion relation can be simplified, and the specific expression is as follows: 2 ​ ； Step S1, the inversion of the model is based on the statistical characteristic relationship of the sea wave spectrum in the sea wave theory, which is expressed as: The average period T of the wind wave obtained by the inversion in step S4 is substituted into the preset sea surface wind speed inversion model in step S1, and the final sea surface wind speed U is obtained by calculation. m , into the preset sea surface wind speed inversion model in step S1, and the final sea surface wind speed U is obtained by calculation. 10 .

2. The method of retrieving sea surface wind speed based on average period of wind wave according to claim 1, characterized in that, First, a power function is selected as the model form, and the observation data of the sea surface wind speed and the average period of the wind waves in 2020 of the European Centre for Medium-Range Weather Forecasts are used for fitting, and finally the power function relationship model of the sea surface wind speed and the average period of the wind waves is established, which is specifically expressed as: ； wherein, U 10 represents the sea surface wind speed, T m is the average period of wind wave, and f is a function relationship based on the existing public data, including the matching data of the average period of wind wave and the sea surface wind speed provided by the European Centre for Medium-Range Weather Forecasts and the National Data Buoy Center. Wherein, a, b are model fitting parameters, and a=2.38, b=0.

87. ； Step S3, the polar coordinates of the image pixels are converted into spatial rectangular coordinates, which specifically includes:

3. The method of retrieving sea surface wind speed from wind wave significant period according to claim 1, characterized in that, For the shore-based marine observation radar images and the shipborne navigation radar images, the size of the polar coordinate image is set as HxW, H represents the image height, and W represents the image width; the value range of the polar angle is [0, 2π), which is one-to-one corresponding to the pixel of the image height H, wherein the polar angle corresponding to the i-th row of pixels is: Step S4, inversion of the average period of the wind waves, specifically including: ； The polar radius r is in the range [0, R max ], R max is the maximum polar radius, and corresponds to the pixels of the image width W, wherein the polar radius corresponding to the jth column of pixels is: ； After converting the polar coordinates of each pixel into rectangular coordinates, the rectangular coordinates (x ij ,y ij ) corresponding to each pixel are: 。 4. The method of retrieving sea surface wind speed from wind wave significant period according to claim 1, characterized in that, The two-dimensional Fourier transform is performed on the sea surface microwave remote sensing images converted into spatial rectangular coordinates after preprocessing, and the energy distribution characteristics of the sea surface radar image in the two-dimensional wave number domain can be obtained, which is expressed in a continuous form as: The discrete Fourier transform form is used, and the specific expression is: ； where L x is the length of the used image in the x direction (range direction), L y is the length of the used image in the y direction (azimuth direction), I(x, y) is the gray value at position (x, y) in the image, k x is the wave number in the x direction, k y is the wave number in the y direction; If there is only one obvious spectrum peak in the two-dimensional image spectrum, the spectrum peak may correspond to large-scale swell or small-scale wind waves; if the wave number corresponding to the spectrum peak is greater than 0.1 rad / m, it is determined that the spectrum peak corresponds to wind waves, and then the average period of the wind waves is calculated according to the wave number corresponding to the spectrum peak. ； The γ(k x ,k y ) after the discrete Fourier transform is processed subsequently to obtain the real two-dimensional sea wave image spectrum I(k x ,k y ): ； If there are two obvious spectrum peaks in the two-dimensional image spectrum, the spectrum peak with larger wave number usually corresponds to the wind wave. Based on this, the two-dimensional wave number (k x ,k y ) corresponding to the wind wave spectrum peak is obtained, and then the average period of the wind wave is inverted using the two-dimensional wave number. The specific calculation method is as follows: ； ​

Images