Dynamic rough sea surface coherent scattering three-scale hybrid modeling and simulation method

Abstract

The invention relates to the field of electromagnetic scattering modeling and simulation, in particular to a dynamic rough sea surface coherent scattering three-scale hybrid modeling and simulation method, which comprises the following steps of: firstly, inputting sea conditions and radar parameters, generating a dynamic sea surface physical field, and dividing scatterers with large, medium and small scales; the method comprises the following steps: carrying out deterministic geometric modeling on pixel-level large-scale components, carrying out Monte Carlo sampling on wavelength-level mesoscale components, carrying out statistical modeling on sub-wavelength-level small-scale components, and then calculating a propagation function by utilizing a multi-layer steepest descent smooth coherence Green function method; according to the method, high sea condition simulation errors and calculation redundancy are effectively reduced, the sea surface coherent information retention degree and simulation reliability are improved, and the method has remarkable technical advantages.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic scattering modeling and simulation, specifically a three-scale hybrid modeling and simulation method for coherent scattering from a dynamic rough sea surface. Background Technology

[0002] Sea surface electromagnetic scattering modeling plays a fundamental role in marine remote sensing parameter inversion, sea surface / maritime target detection, and imaging radar (e.g., SAR) processing chains. Under typical sea conditions, the sea surface is not a single-scale random undulation, but often exhibits a multi-scale structure ranging from subwavelength fine coarseness to wavelength-level undulations and pixel-level large-scale waveforms. Simultaneously, driven by wind fields and gravity waves, the sea surface exhibits significant temporal evolution characteristics, causing the scattering phase to change over time, leading to phase center migration. This combination of "multi-scale coupling + large undulations + dynamics" means that sea surface scattering is influenced not only by geometry but also by propagation paths, coherent superposition, and motion modulation, placing higher demands on high-precision scattering modeling.

[0003] Existing research offers a rich array of modeling methods for electromagnetic scattering from rough sea surfaces. Common approaches include approximate deterministic methods such as geometrical optics (GO) and physical optics (PO), as well as numerical solutions based on integral equations. Studies on modeling scattering from complex three-dimensional rough sea surfaces reveal that when sea surface morphology is complex and scales are wide, a single approximation model often struggles to comprehensively account for scattering mechanisms at different scales within a unified framework. On the one hand, large-scale undulations significantly impact local incident angles and visibility; on the other hand, small-scale roughness (especially components related to Bragg scattering) makes crucial contributions to scattering intensity and polarization response. Using only a single-scale representation or lacking coherent and consistent connections between different scales can easily lead to difficulties in maintaining multi-scale coherence, thus affecting the reliability and comparability of the overall scattered field.

[0004] Besides the multi-scale problem, errors introduced by dynamics are also a prominent shortcoming of existing technologies. Many modeling and simulation works in engineering applications still habitually employ static or quasi-static sea surface approximations, i.e., generating the sea surface morphology and performing scattering calculations at a single moment, then stitching the time series together in some way. However, existing literature points out that dynamic sea surface scene modeling based on dual-scale sea spectra emphasizes the importance of sea surface evolution over time. Static approximations struggle to accurately describe the continuous change of the scattering phase over time, especially failing to reflect the impact of phase center migration on imaging processing (e.g., SAR focusing). In applications requiring phase information, coherent superposition information, or Doppler modulation information, insufficient dynamic modeling often leads to accumulated phase errors or decreased coherence, causing a mismatch between scattering simulation results and actual observations in the imaging / detection link.

[0005] Meanwhile, the trade-off between efficiency and accuracy in propagation computation and Green's function solving has long existed in the pursuit of high-precision numerical solutions. While integral equation methods or Green's function-based solution frameworks possess strong theoretical completeness, they often face enormous computational burdens under conditions of large-scale sea surfaces, dense discrete surfaces, and highly undulating rough surfaces. Therefore, fast algorithms or approximation acceleration strategies are often introduced in engineering. Related research involves fast solutions using different Green's functions (e.g., fast multipole subclass approaches) to improve efficiency. However, under complex rough surfaces, especially with large undulations, approximation may introduce non-negligible errors, making it difficult to obtain high-precision numerical solutions or requiring significant sacrifices between accuracy and speed. Therefore, current technologies still struggle to simultaneously achieve coherence preservation, dynamic consistency, and high-precision propagation solutions under the combined conditions of large scale, strong roughness, and strong dynamics.

[0006] In recent years, some scholars have proposed multi-scale or hybrid schemes to alleviate the limitations of single models under complex sea conditions and to introduce statistical or hybrid modeling ideas into specific tasks (such as detection). However, from the perspective of sea surface electromagnetic scattering simulation, existing multi-scale / hybrid schemes may still have two shortcomings: First, the coherence and phase consistency between scales after multi-scale decomposition are not adequately handled, resulting in distortion of the overall coherent scattering field even though multi-scale components are introduced; second, the introduction of more complex numerical calculations to maintain a certain level of accuracy may lead to insufficient computational efficiency, making it difficult to apply in large-scale scenarios or when time-series dynamic simulations are required. In summary, for the multi-scale coupling, large undulations, and significant dynamic characteristics of rough sea surfaces under typical sea conditions, existing technologies still suffer from shortcomings such as difficulty in maintaining multi-scale coherence, difficulty in accurately characterizing dynamic phase evolution, and difficulty in balancing accuracy and efficiency in fast propagation / Green's function solution under large undulation conditions. There is an urgent need for a modeling and simulation method that can take into account multi-scale, dynamic, and high-precision propagation calculations within a unified framework. Summary of the Invention

[0007] The purpose of this invention is to provide a three-scale hybrid modeling and simulation method for coherent scattering of dynamic rough sea surface, so as to solve the technical problems mentioned in the background art.

[0008] Based on the above ideas, the present invention provides the following technical solution: A three-scale hybrid modeling and simulation method for coherent scattering from a dynamic rough sea surface includes the following steps: S1: Generate a dynamic sea surface physical field based on the input sea surface parameters and radar parameters; S2: Divide the generated sea surface into three components: pixel-level large scale, wavelength-level medium scale, and subwavelength-level small scale, and generate a grid; S3: Calculate the propagation function using the multi-layer steepest descent smooth coherent Green's function method; S4: Perform motion modeling and scattering modeling for each coherent scatterer; S5: Coherently superimpose the scattering contributions at each scale to obtain the overall sea surface scattering field; The sea surface parameters include the simulated sea surface geometry, number of grid nodes, frictional wind speed, wind direction, peak wave period, and wave direction spectrum. The radar parameters include carrier frequency, bandwidth, pulse repetition frequency, polarization, radar incident angle, and radar azimuth angle.

[0009] By defining the simulation process as dynamic sea surface physical field generation, three-scale partitioning and meshing, propagation function calculation, scatterer motion and scattering modeling, and coherent superposition output, a complete end-to-end modeling chain is formed for the coherent scattering problem of dynamic rough sea surface. This chain can simultaneously consider the time-varying nature of the sea surface, scale differences, and coherent superposition effects within a unified framework, thereby improving the systematicness and consistency of the overall scattering field simulation and reducing the accumulation of errors and incomparability of results caused by the fragmentation of different sub-modules.

[0010] Preferably, the dynamic sea surface physical field in S1 is composed of a discrete quadrilateral grid, including a sea surface height field, a local slope field, and an orbital velocity field.

[0011] By establishing the dynamic sea surface physical field as a discrete quadrilateral grid and explicitly including the sea surface height field, local slope field, and orbital velocity field, the geometric undulations, local orientation, and motion state of the sea surface can be uniformly described in the same discrete representation. This facilitates direct calls to the grid nodes for subsequent propagation calculations and scattering modeling, improving the usability of the modeling data structure and the efficiency of computational organization.

[0012] Preferably, the sea surface height field is numerically simulated using the linear wave superposition method, where the sea surface height is obtained by superimposing multiple plane advancing wave components at any spatial location and time. The local slope field is calculated based on the partial derivative of the sea surface height field with respect to the planar spatial coordinates. The local slope field includes a slope component along the first spatial coordinate direction and a slope component along the second spatial coordinate direction. The orbital velocity field is obtained by the gradient of the velocity potential function in the spatial direction; the velocity potential function is related to gravitational acceleration, angular frequency, depth attenuation term and phase term.

[0013] By using linear wave superposition to generate the sea surface height field, using the spatial partial derivative of the height field to obtain the slope field, and using the gradient of the velocity potential function to obtain the orbital velocity field, the inherent consistency among the three physical quantities of height, slope, and velocity is achieved, avoiding the physical inconsistency caused by setting the three separately. At the same time, this generation method is conducive to continuously updating the sea surface state in the spatiotemporal domain, thereby more realistically reflecting the influence of dynamic rough sea surface on coherent scattering.

[0014] Preferably, in step S2, the condition for dividing the generated sea surface into subwavelength-level small-scale components is that the sea surface feature scale is smaller than the radar carrier wavelength; the condition for dividing the generated sea surface into wavelength-level medium-scale components is that the sea surface feature scale is between the radar carrier wavelength and the radar system range resolution; and the condition for dividing the generated sea surface into pixel-level large-scale components is that the sea surface feature scale is larger than the radar system range resolution. The pixel-level large-scale components are modeled using deterministic geometric modeling, the wavelength-level mesoscale components are modeled using Monte Carlo sampling, and the subwavelength-level small-scale components are modeled using statistical modeling.

[0015] By using the relationship between sea surface feature scale and radar carrier wavelength and range resolution as the three-scale division condition, and employing deterministic geometric modeling, Monte Carlo sampling, and statistical modeling respectively, sea surface features at different scales can be matched with more suitable description methods, achieving scale-adaptive hierarchical modeling. This not only preserves the deterministic influence of large-scale geometric morphology but also effectively represents the coarse characteristics at small and medium scales, thus achieving a more reasonable balance between accuracy and computational cost.

[0016] Preferably, the grid size in S2 is seven times the radar carrier wavelength.

[0017] By setting the grid size to seven times the radar carrier wavelength, a clear correlation is established between the grid discrete scale and the electromagnetic wave scale. This helps to maintain coverage of key scattering scales while ensuring gridding efficiency, avoiding unnecessary computational burden caused by overly dense grids or feature loss caused by overly sparse grids, thereby improving the feasibility and stability of the overall simulation.

[0018] Preferably, the propagation function in step S3 is obtained by accumulating multiple approximate saddle points obtained from the steepest descent of multiple layers; each accumulated term includes an amplitude weight determined by the approximate saddle point, a complex exponential phase term determined by the phase function, and a Hamming window function.

[0019] By employing the multi-layer steepest descent smooth coherent Green's function method and representing the propagation function as an accumulation of multiple approximate saddle points, the propagation calculation can be organized in a saddle point approximation and accumulation manner, which is beneficial for obtaining more stable propagation function calculation results under complex propagation conditions. At the same time, the introduction of the Hamming window function helps to smooth the accumulation process and suppress adverse fluctuations, thereby improving the stability and usability of the numerical results of the propagation function.

[0020] Preferably, the amplitude weight is determined by the Hessian matrix of the phase function at an approximate saddle point; In each iteration, the wave vector of the next iteration is updated in the opposite direction of the gradient of the phase function by a preset step size.

[0021] By associating the amplitude weights with the Hessian matrix of the phase function at the approximate saddle point and obtaining the wave vector through iterative updates along the opposite direction of the phase function gradient, a unified organization of saddle point solving and weight evaluation in propagation calculation is achieved. This approach helps to give the saddle point search process a clear iterative direction and update mechanism, thereby improving the controllability of the approximate saddle point solution and enhancing the convergence and consistency of propagation function calculation at the numerical solution level.

[0022] Preferably, the motion modeling in S4 includes: For each coherent scatterer, a velocity vector field is first generated from the large-scale spectrum, and then the phase velocity of the corresponding component of the Bragg scattering part is superimposed. The direction of the phase velocity is consistent with the direction of wave propagation.

[0023] By first generating a velocity vector field from a large-scale spectrum in motion modeling, and then superimposing the phase velocity of the corresponding component of the Bragg scattering part, and limiting the direction of the phase velocity to be consistent with the direction of wave propagation, the motion state of the scatterer simultaneously includes the phase propagation characteristics of the large-scale flow background and Bragg-related phase propagation characteristics. This can more comprehensively reflect the influence of dynamic sea surface motion on the scattering phase and Doppler characteristics, thereby improving the physical rationality of introducing motion factors.

[0024] The preferred scattering modeling in S4 includes: Calculate the large-scale scattering contribution and the small-scale Bragg scattering contribution for each coherent scatterer; The magnitude of the contribution of large-scale scattering is characterized by exponential decay. Its decay relationship is related to the radar wave number, the standard deviation of sea surface height, and the radar incident angle, and is also related to the amplitude of the large-scale scattering reflection coefficient of the sea surface and the radar azimuth beamwidth. The magnitude of the small-scale Bragg scattering contribution is related to the radar wavenumber, the square of the polarization scattering matrix amplitude, and the spectral value of the zonal sea surface spectrum at the Bragg wavenumber, and also includes the spectral term contributions corresponding to both positive and negative Bragg wavenumbers. The polarization scattering matrix is ​​obtained by multiplying the incident and scattered polarization basis transformation matrices and the diagonal scattering coefficient matrix.

[0025] By calculating the large-scale scattering contribution and the small-scale Bragg scattering contribution separately, and characterizing the large-scale scattering amplitude in the form of exponential decay, while using the positive and negative Bragg wavenumber terms of the directional sea surface spectrum to jointly characterize the small-scale contribution, the scattering modeling simultaneously covers the different mechanisms by which macroscopic geometric undulations and microscopic rough textures affect scattering, avoiding modeling bias caused by a single mechanism. At the same time, the polarization relationship between incident and scattering is uniformly expressed by the polarization scattering matrix through basis transformation and diagonal scattering coefficient matrix, which helps to enhance the consistency and scalability of polarization channel modeling, thereby improving the adaptability of scattering results to different polarization modes.

[0026] Preferably, the overall sea surface scattering field of S5 is obtained by coherently superimposing the complex scattering amplitudes of large-scale scatterers, medium-scale scatterers, and small-scale scatterers.

[0027] By constructing the overall sea surface scattering field as a coherent superposition of the complex scattering amplitudes of large-scale, medium-scale, and small-scale scatterers, the scattering contributions of different scales are uniformly superimposed at the phase level, thereby reflecting the influence of coherent effects on the overall scattering field and avoiding the loss of coherent information caused by power or incoherent superposition alone. At the same time, this coherent superposition form facilitates a comprehensive comparison of the relative effects of contributions from different scales on the same output quantity.

[0028] The technical solution of the present invention may include the following beneficial effects: By integrating dynamic sea surface physical field generation, three-scale division, propagation function calculation, scatterer motion / scattering modeling, and coherent superposition output, the time-varying nature of the sea surface, scale differences, and coherent superposition are incorporated into the same simulation chain. This avoids physical inconsistencies and error accumulation caused by the fragmentation of different modules, making the overall scattering field simulation more systematic and consistent, and the output more comparable and reproducible.

[0029] By dividing the model into three scales based on the criterion of the sea surface feature scale relative to the radar wavelength / range resolution, and employing deterministic geometric modeling, Monte Carlo sampling, and statistical modeling for different scales, along with wavelength-related grid settings, a scale-adaptive hierarchical modeling is achieved. This effectively characterizes the coarseness characteristics at small and medium scales while preserving the influence of large-scale geometric morphology, thus achieving a more reasonable balance between simulation accuracy and computational complexity, and improving the feasibility and computational efficiency of the method.

[0030] By employing multi-layered steepest descent smooth coherent Green's function propagation calculation, Hessian matrix-related amplitude weighting and gradient iterative solution mechanism, and scattering modeling and polarization matrix expression based on large-scale contribution + Bragg small-scale contribution, and finally performing complex amplitude coherent superposition, the propagation solution is made more stable, the scattering mechanism is more comprehensively covered, and the influence of polarization and positive and negative Bragg spectral terms is consistently incorporated. This allows for a more accurate characterization of the phase and amplitude characteristics of coherent scattering from a dynamic rough sea surface, improving the physical rationality and credibility of the simulation results. Attached Figure Description

[0031] Figure 1 This is a flowchart of a three-scale hybrid modeling and simulation method for coherent scattering of a dynamic rough sea surface according to the present invention.

[0032] Figure 2 This invention relates to a dynamic sea surface height field based on a three-scale hybrid modeling and simulation method for coherent scattering of a dynamic rough sea surface.

[0033] Figure 3This invention relates to a dynamic sea surface orbital velocity vector field for a three-scale hybrid modeling and simulation method of coherent scattering on a dynamic rough sea surface.

[0034] Figure 4 This is a schematic diagram of a multi-scale hybrid coherent modeling method for a three-scale hybrid modeling and simulation method for coherent scattering of dynamic rough sea surface according to the present invention. Detailed Implementation Example 1

[0035] like Figure 1 This embodiment describes a three-scale hybrid modeling and simulation method for coherent scattering from a dynamic rough sea surface, including the following steps: S1: Generate a dynamic sea surface physical field based on the input sea surface parameters and radar parameters; S2: Divide the generated sea surface into three components: pixel-level large scale, wavelength-level medium scale, and subwavelength-level small scale, and generate a grid; S3: Calculate the propagation function using the smooth coherent Green's function method with the steepest descent across multiple layers. ; S4: Perform motion modeling and scattering modeling for each coherent scatterer; S5: Coherently superimpose the scattering contributions at each scale to obtain the overall sea surface scattered field. .

[0036] The sea surface parameters include the simulated sea surface geometry, number of grid nodes, frictional wind speed, wind direction, peak wave period, and wave direction spectrum. ; The radar parameters include carrier frequency, bandwidth, pulse repetition frequency, polarization, radar incident angle, and radar azimuth angle.

[0037] Specifically, the dynamic sea surface physical field in S1 is composed of a discrete quadrilateral grid, including the sea surface height field. Local slope field Orbital velocity field .

[0038] Specifically, the sea surface height field Numerical simulation is performed using the linear wave superposition method, and the calculation formula is as follows: (1) in: It is the amplitude of a plane-progressing wave. These represent the wave number, angular frequency, propagation direction, and initial phase of a plane-progressing wave. For the direction spectrum of ocean waves, These represent the number of grid nodes in the length and width directions, respectively.

[0039] Specifically, the local slope field It is calculated based on the partial derivative of the sea surface height field with respect to spatial coordinates, where: (2) (3) Specifically, the orbital velocity field It is composed of the velocity potential function The velocity potential function is obtained by differentiating in the spatial direction. The expression is: (4) Orbital velocity field It can be represented as: (5) in (6) like Figure 2 and Figure 3 In one embodiment, the sea surface height field The local slope field is calculated using formula (1). The orbital velocity field is obtained from formulas (2) and (3). It is calculated using formulas (4), (5) and (6).

[0040] Specifically, such as Figure 4 The condition in S2 for dividing the generated sea surface into subwavelength-level small-scale components is the characteristic scale of the sea surface. Smaller than radar carrier wavelength The condition for classifying components into wavelength-level mesoscale components is the characteristic scale of the sea surface. Due to radar carrier wavelength Range resolution of radar system The condition for dividing the data into pixel-level large-scale components is the sea surface feature scale. Greater than the range resolution of radar systems .

[0041] Specifically, the pixel-level large-scale components are modeled using deterministic geometric modeling, the wavelength-level mesoscale components are modeled using Monte Carlo sampling, and the subwavelength-level small-scale components are modeled using statistical modeling.

[0042] Specifically, the size of the grid in S2 for: (7) in This is the radar carrier wavelength.

[0043] Specifically, the propagation function in step S3 for: (8) in, To find the approximate saddle point obtained during the m-th steepest descent, For amplitude, For phase function, This is the Hamming window function.

[0044] Preferably, the amplitude term The expression is: (9) in Let be the Hessian matrix of the phase function.

[0045] Specifically, the approximate saddle point obtained during the m-th steepest descent... The steepest descent iteration is employed, and its iterative formula is as follows: (10) This is the step size for the m-th iteration, with a value of 0.1. Initial value. iteration step size The value is 0.1, and the number of layers m is 10.

[0046] Specifically, the motion modeling in S4 includes, for each coherent scatterer, first generating a velocity vector field from the large-scale spectrum, and then superimposing the phase velocity of the corresponding component of the Bragg scattering part. : (11) The direction of velocity is the same as the direction of wave propagation.

[0047] Preferably, the scattering modeling in S4 includes calculating the large-scale scattering contribution and the small-scale Bragg scattering contribution for each coherent scatterer, wherein the large-scale scattering contribution is obtained by: (12) The contribution of small-scale Bragg scattering is: (13) in For radar wave number, For sea level height standard deviation, For radar incident angle, The amplitude of the large-scale scattering and reflection coefficient of the sea surface, For radar azimuth beamwidth, For directional sea surface spectrum, Here is the polarization scattering matrix.

[0048] Specifically, the directional sea surface spectrum for: (14) in, and These are the wave numbers in the x and y directions, respectively.

[0049] Preferably, the polarization scattering matrix for: (15) in, and These are the horizontal and vertical polarization unit vectors of the incident wave, respectively. and These are the horizontal and vertical polarization unit vectors of the scattered wave, respectively. The scattering coefficients correspond to the incident horizontal polarization and the scattering horizontal polarization. , This refers to the scattering coefficient corresponding to the incident vertical polarization to the scattering vertical polarization. .

[0050] Preferably, the overall sea surface scattering field of S5 for: (16) in, These represent the total number of large-scale, medium-scale, and small-scale scatterers, respectively. , , The corresponding scatterers are located at positions The amplitude of complex scattering at that location.

[0051] This application uses the dynamic sea surface physical field as a common foundation, and clarifies that the sea surface height field, local slope field and orbital velocity field are generated on the same discrete quadrilateral grid and maintain physical consistency. This ensures that geometric undulations (height / slope) and time-varying motion (orbital velocity) are no longer modeled separately, and provides a unified state variable support for dynamic effects such as the migration of the phase center over time, which can be propagated and superimposed from the source.

[0052] The three-scale partitioning is not based on empirical block division, but rather on the comparison between the sea surface feature scale and the radar carrier wavelength and system range resolution. Small scale (smaller than wavelength), medium scale (between wavelength and range resolution), and large scale (larger than range resolution) are assigned to different modeling strategies, and grid scale settings related to carrier wavelength are given (grid size is bound to carrier wavelength). This ensures that discrete resolution, electromagnetic scale, and system resolution form consistent constraints at the parameter level, thereby ensuring that subsequent propagation and scattering calculations will not suffer from numerical dissipation of coherent information due to grid / scale mismatch, nor will excessive refinement bring unnecessary computational burden.

[0053] Building upon this foundation, the solution for the propagation function introduces a multi-layered, steepest descent smooth coherent Green's function method. This method expresses the propagation function as a term-by-term summation of multiple approximate saddle points obtained from the steepest descent. In each term, an amplitude weight determined by the Hessian matrix of the phase function and a Hamming window function are introduced simultaneously. Furthermore, the approximate saddle points are obtained through iterative updates with a given step size (step size 0.1, number of layers m = 10). This enables the propagation process to more stably preserve the detailed structure of phase and amplitude under conditions of large undulations and rough surfaces, thereby reliably transmitting phase changes caused by height / slope / velocity in the dynamic sea surface physical field to the scatterer level.

[0054] Subsequently, at the scatterer level, a further coupling of motion scattering is formed: in motion modeling, a velocity vector field is first generated from the large-scale spectrum for each coherent scatterer, and then the phase velocity of the corresponding component of the Bragg scattering part is superimposed with the direction consistent with the wave propagation direction, so that the influence of dynamic motion on the scattering phase / Doppler can be naturally connected with the Bragg mechanism; in scattering modeling, the large-scale scattering contribution and the small-scale Bragg scattering contribution are calculated in parallel. The former is characterized by an exponential decay form related to radar wavenumber, sea surface height standard deviation, incident angle, etc., while the latter is related to the spectral value of the directional sea surface spectrum at the Bragg wavenumber and the polarization scattering matrix, and the contributions of positive and negative Bragg spectral terms are considered at the same time. Finally, in S5, the complex scattering amplitudes of the large / medium / small-scale scatterers are coherently superimposed to obtain the overall sea surface scattering field.

[0055] The resulting synergistic mechanism is as follows: the binding of the three-scale criterion with the grid scale ensures that the numerical expression of scatterers at each scale matches the electromagnetic / system scale; the high-precision propagation function solution brings the phase information of the dynamic physical field into the scattering calculation with controllable error; motion modeling and Bragg phase velocity superposition explicitly inject dynamism into the phase evolution of coherent scatterers; the large-scale exponential decay mechanism and the small-scale Bragg spectral mechanism cover the scattering contributions from different physical sources in parallel and are characterized by a unified polarization channel by the polarization matrix; finally, the phase consistency retained in the above steps is truly transformed into an improvement in the accuracy of the overall scattering field through complex amplitude coherent superposition.

Claims

1. A three-scale hybrid modeling and simulation method for coherent scattering from a dynamic rough sea surface, characterized in that, Includes the following steps: S1: Generate a dynamic sea surface physical field based on the input sea surface parameters and radar parameters; S2: Divide the generated sea surface into three components: pixel-level large scale, wavelength-level medium scale, and subwavelength-level small scale, and generate a grid; S3: Calculate the propagation function using the multi-layer steepest descent smooth coherent Green's function method; S4: Perform motion modeling and scattering modeling for each coherent scatterer; S5: Coherently superimpose the scattering contributions at each scale to obtain the overall sea surface scattering field; The sea surface parameters include the simulated sea surface geometry, number of grid nodes, frictional wind speed, wind direction, peak wave period, and wave direction spectrum. The radar parameters include carrier frequency, bandwidth, pulse repetition frequency, polarization, radar incident angle, and radar azimuth angle.

2. The method for three-scale hybrid modeling and simulation of coherent scattering over a dynamic rough sea surface according to claim 1, characterized in that, The dynamic sea surface physical field in S1 is composed of a discrete quadrilateral grid, including the sea surface height field, the local slope field, and the orbital velocity field.

3. The method for three-scale hybrid modeling and simulation of coherent scattering over a dynamic rough sea surface according to claim 2, characterized in that, The sea surface height field is numerically simulated using the linear wave superposition method. At any spatial location and time, the sea surface height is obtained by superimposing multiple plane advancing wave components. The local slope field is calculated based on the partial derivative of the sea surface height field with respect to the planar spatial coordinates. The local slope field includes a slope component along the first spatial coordinate direction and a slope component along the second spatial coordinate direction. The orbital velocity field is obtained by the gradient of the velocity potential function in the spatial direction; the velocity potential function is related to gravitational acceleration, angular frequency, depth attenuation term and phase term.

4. The method for three-scale hybrid modeling and simulation of coherent scattering over a dynamic rough sea surface according to claim 1, characterized in that, In S2, the condition for dividing the generated sea surface into subwavelength-level small-scale components is that the sea surface feature scale is smaller than the radar carrier wavelength; the condition for dividing the generated sea surface into wavelength-level medium-scale components is that the sea surface feature scale is between the radar carrier wavelength and the radar system range resolution; the condition for dividing the generated sea surface into pixel-level large-scale components is that the sea surface feature scale is larger than the radar system range resolution. The pixel-level large-scale components are modeled using deterministic geometric modeling, the wavelength-level mesoscale components are modeled using Monte Carlo sampling, and the subwavelength-level small-scale components are modeled using statistical modeling.

5. The method for three-scale hybrid modeling and simulation of coherent scattering from a dynamic rough sea surface according to claim 1, characterized in that, The grid size in S2 is seven times the radar carrier wavelength.

6. The method for three-scale hybrid modeling and simulation of coherent scattering over a dynamic rough sea surface according to claim 1, characterized in that, The propagation function in step S3 is obtained by accumulating multiple approximate saddle points obtained from the steepest descent of multiple layers; each accumulated term includes an amplitude weight determined by the approximate saddle point, a complex exponential phase term determined by the phase function, and a Hamming window function.

7. The method for three-scale hybrid modeling and simulation of coherent scattering over a dynamic rough sea surface according to claim 6, characterized in that, The amplitude weight is determined by the Hessian matrix of the phase function at the approximate saddle point; In each iteration, the wave vector of the next iteration is updated in the opposite direction of the gradient of the phase function by a preset step size.

8. The method for three-scale hybrid modeling and simulation of coherent scattering from a dynamic rough sea surface according to claim 1, characterized in that, The motion modeling in S4 includes: For each coherent scatterer, a velocity vector field is first generated from the large-scale spectrum, and then the phase velocity of the corresponding component of the Bragg scattering part is superimposed. The direction of the phase velocity is consistent with the direction of wave propagation.

9. The method for three-scale hybrid modeling and simulation of coherent scattering from a dynamic rough sea surface according to claim 1, characterized in that, The scattering modeling in S4 includes: Calculate the large-scale scattering contribution and the small-scale Bragg scattering contribution for each coherent scatterer; The magnitude of the contribution of large-scale scattering is characterized by exponential decay. Its decay relationship is related to the radar wave number, the standard deviation of sea surface height, and the radar incident angle, and is also related to the amplitude of the large-scale scattering reflection coefficient of the sea surface and the radar azimuth beamwidth. The magnitude of the small-scale Bragg scattering contribution is related to the radar wavenumber, the square of the polarization scattering matrix amplitude, and the spectral value of the zonal sea surface spectrum at the Bragg wavenumber, and also includes the spectral term contributions corresponding to both positive and negative Bragg wavenumbers. The polarization scattering matrix is ​​obtained by multiplying the incident and scattered polarization basis transformation matrices and the diagonal scattering coefficient matrix.

10. The method for three-scale hybrid modeling and simulation of coherent scattering over a dynamic rough sea surface according to claim 1, characterized in that, The overall sea surface scattering field of S5 is obtained by coherently superimposing the complex scattering amplitudes of large-scale scatterers, medium-scale scatterers, and small-scale scatterers.

Images