CN117216954B - Sea surface scattering coefficient rapid prediction method, system, equipment and medium - Google Patents
Sea surface scattering coefficient rapid prediction method, system, equipment and medium Download PDFInfo
- Publication number
- CN117216954B CN117216954B CN202311063177.1A CN202311063177A CN117216954B CN 117216954 B CN117216954 B CN 117216954B CN 202311063177 A CN202311063177 A CN 202311063177A CN 117216954 B CN117216954 B CN 117216954B
- Authority
- CN
- China
- Prior art keywords
- sea surface
- scattering coefficient
- surface scattering
- wave
- simulation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000004088 simulation Methods 0.000 claims abstract description 78
- 239000011159 matrix material Substances 0.000 claims abstract description 60
- 238000001228 spectrum Methods 0.000 claims abstract description 19
- 238000012795 verification Methods 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000000790 scattering method Methods 0.000 claims abstract description 9
- 230000000694 effects Effects 0.000 claims abstract description 8
- 238000012216 screening Methods 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 16
- 238000010276 construction Methods 0.000 claims description 12
- 230000010287 polarization Effects 0.000 claims description 10
- 125000004122 cyclic group Chemical group 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 101100517651 Caenorhabditis elegans num-1 gene Proteins 0.000 claims description 3
- 230000001133 acceleration Effects 0.000 claims description 3
- 125000003275 alpha amino acid group Chemical group 0.000 claims description 3
- 150000001875 compounds Chemical class 0.000 claims description 3
- 230000001419 dependent effect Effects 0.000 claims description 3
- 238000005315 distribution function Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 description 5
- 230000005684 electric field Effects 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A sea surface scattering coefficient rapid prediction method, a system, equipment and a medium, wherein the method comprises the following steps: constructing a simulation parameter matrix, calculating a simulation parameter combined with a sea surface scattering coefficient by adopting a multi-parameter wind-wave hybrid improved Venturi spectrum bin scattering method, constructing an initial sea surface scattering coefficient data set, performing principal component dimension reduction processing on the simulation parameter matrix, extracting principal characteristics, screening principal parameter components, constructing a principal parameter component sea surface scattering coefficient data set, performing data subset division on the principal parameter component sea surface scattering coefficient data set, combining circular verification, solving the optimal parameters of a regression prediction model, constructing a sea surface scattering coefficient nonlinear regression prediction model, and finally evaluating the performance of the model by adopting three indexes; the system, the equipment and the medium are used for realizing a sea surface scattering coefficient rapid prediction method; compared with the prior art, the sea surface scattering coefficient prediction method can realize quick prediction of the sea surface scattering coefficient, is accurate in prediction effect and has good generalization capability.
Description
Technical Field
The invention relates to the technical field of global marine environment monitoring, in particular to a method, a system, equipment and a medium for rapidly predicting sea surface scattering coefficient.
Background
In the field of ocean remote sensing, sea surface electromagnetic scattering property research is a key of sea surface parameter inversion and is also a basis for improving target detection capability in a complex ocean environment, in recent years, along with development of radar technology and remote sensing application, quick and accurate acquisition of sea surface electromagnetic scattering coefficients becomes a research hot spot, and machine learning has great potential in the aspect of quick prediction of sea surface electromagnetic scattering due to superiority in practical problems such as small samples, nonlinearity, high dimensionality, local minimum points and the like; however, because of complex dependency relationship among sea surface electromagnetic scattering, sea wave parameters, radar parameters and other influencing factors, the sea surface scattering coefficient prediction method based on machine learning has high complexity, randomness and high dimensionality, so that the existing sea surface scattering coefficient prediction method based on machine learning is easy to generate an overfitting problem when facing multi-parameter high dimensionality mapping, and the model generalization capability is poor.
The invention discloses a method for estimating sea surface scattering coefficient of a space-based high-frequency radar, which comprises the steps of constructing Fourier series of linear and nonlinear sea surface wave heights, further constructing a relation model between the linear and nonlinear wave height coefficients through a perturbation analysis method, and constructing a relation model of directed sea wave spectrum and the linear wave height coefficients; constructing the electric field intensity of an incident field, constructing components of a near-sea-surface scattered field electric field along X and Z axis directions according to a Fourier series by combining the electric field intensity of the incident field and a perturbation method, and further constructing a vertical polarization field intensity component model of a far-sea-surface scattered field magnetic field; however, due to the linear sea wave height used by the method, the comprehensive effects of various factors on the sea, such as wind speed, wind direction, wind area, wave height, wave direction, wave period and the like, are not considered, so that the model of the method is too simple, details of sea wave fluctuation cannot be captured well, and further the estimated sea scattering coefficient is inaccurate.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a sea surface scattering coefficient rapid prediction method, a system, equipment and a medium.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
a sea surface scattering coefficient rapid prediction method comprises the following steps:
step 1, constructing a simulation parameter matrix according to simulation parameters;
the simulation parameters are divided into two types, namely ocean environment parameters and radar parameters, wherein the ocean environment parameters comprise wind speed, wind direction, wave height, wave direction, wind area, temperature, salinity and wave period; the radar parameters comprise an incident wave frequency, a pitch angle and an azimuth angle;
step 2, calculating the simulation parameters in the step 1 by combining the sea surface scattering coefficients by adopting a multiparameter wind-wave hybrid improved Venturi spectrum bin scattering method, so as to construct an initial sea surface scattering coefficient data set;
step 3, performing main component dimension reduction processing on the simulation parameter matrix in the step 1, extracting main characteristics, screening main parameter components, and constructing a sea surface scattering coefficient data set of the main parameter components;
step 4, dividing a data subset of the sea surface scattering coefficient data set of the main parameter component constructed in the step 3, solving the optimal parameters of the regression prediction model by combining with cyclic verification, and establishing a sea surface scattering coefficient nonlinear regression prediction model;
step 5, determining coefficients (R) by using Root Mean Square Error (RMSE) and model respectively 2 ) And three indexes of Bayesian information quantity (BIC) are used for evaluating the performance of the sea surface scattering coefficient nonlinear regression prediction model established in the step 4;
in the step 1, a simulation parameter matrix is constructed according to simulation parameters, specifically:
constructing a simulation parameter matrix by using the simulation parameters, wherein the simulation parameter matrix is shown as a formula (1):
in the formula (1), M p Represents a simulation parameter matrix, N represents the number of samples, U and theta wi 、H、θ wa 、X、T、S、T S F, θ andrespectively represent wind speed, wind direction, wave height, wave direction, wind area, temperature, salinity, wave period, incident wave frequency, pitch angle and azimuth angle, upper angle mark (N) Representing an N-dimensional column vector;
in the step 2, an initial sea surface scattering coefficient data set is constructed, specifically:
step 2.1, performing geometric modeling on a three-dimensional sea surface large-scale contour based on a multi-parameter wind-wave hybrid improved sea spectrum, and obtaining geometric information of the three-dimensional sea surface large-scale wave, wherein the multi-parameter wind-wave hybrid improved sea spectrum is shown as a formula (2):
in the formula (2), S (omega, theta) wa ) Represents the multi-parameter wind-wave mixing improved sea spectrum, omega is the angular frequency of sea wave, T S For the wave period, U is the wind speed 10m above the sea surface, D (θ wa ) Is a Donelan direction function; the expression of the F (U, X) function is shown in the formula (3):
2.2, on the basis of geometric modeling of three-dimensional sea surface large-scale waves, considering superposition of wind-induced small-scale capillary waves, and calculating the scattering coefficient of each bin by adopting a multi-parameter wind-wave mixed bin scattering method, wherein the scattering coefficient is shown in a formula (4):
wherein p, q respectively represent polarization modes of scattered waves and incident waves, ω represents sea wave angular frequency, g represents gravitational acceleration,k represents wave number>And->The normalized wave vectors corresponding to the incident wave and the scattered wave are respectively represented,represents the sea surface slope distribution function, +.>Representing the slope of the tangent plane of the large-scale wavelets mirror point, ε represents the sea surface dielectric constant, Q pq And W is pq Respectively representing mirror image scattering factors and polarization factors corresponding to large-scale wave surface elements, and subscripts pq Representing horizontal or vertical polarization, S wc (U) represents a wind-induced small-scale capillary wave spectrum;
carrying out set average calculation on the scattering coefficients of all the surface elements to obtain the total scattering coefficient of the whole three-dimensional sea surface, wherein the total scattering coefficient is shown as a formula (5):
in the formula (5), m and n respectively represent the number of the sea surface x-axis and y-axis direction surface elements, sigma pq,mn The scattering coefficients corresponding to the (m, n) th bin obtained by calculation in the formula (4) are represented, and Deltax and Deltay represent the sampling intervals in the x and y axis directions of the sea surface respectively;
step 2.3, combining the simulation parameters in the step 1, and calculating three-dimensional sea surface scattering coefficients under different simulation parameters through the calculation formulas in the step 2.1 and the step 2.2 to construct an initial sea surface scattering coefficient data set;
in the step 3, a main parameter component sea surface scattering coefficient data set is constructed, specifically:
step 3.1, carrying out data standardization processing on the simulation parameter matrix, and obtaining a standardized simulation parameter matrix after the original data is subjected to the standardization processing, wherein the standardized simulation parameter matrix is shown as a formula (6):
in the formula (6), the standardized simulation parameter matrix elementM pij Is the element corresponding to the simulation parameter matrix in the formula (1)>Represents the mean value of the j-th column data of the simulation parameter matrix, D pj Representing the standard deviation of the j-th column data of the simulation parameter matrix;
step 3.2, calculating a covariance matrix of the standardized simulation parameter matrix, solving eigenvalues and eigenvectors of the covariance matrix, and selecting the first k main eigenvalues, wherein k is smaller than 11;
step 3.3, reconstructing a data set according to the screened k main characteristic values to obtain a sea surface scattering coefficient data set of the main parameter component;
in the step 4, a sea surface scattering coefficient nonlinear regression prediction model is established, specifically:
step 4.1, dividing the main parameter component sea surface scattering coefficient data set obtained in the step 3 into a training set and a testing set by adopting a random division mode, and further dividing the training set into (Num-1) training subsets and 1 verification subset by adopting a random equidistant mode, wherein Num represents the total number of the data subsets;
and 4.2, performing cyclic verification on the training subset and the verification subset in the step 4.1, calculating a generalization error mean value of the regression prediction model, selecting a model parameter with the minimum generalization error mean value from the training subset and the verification subset, namely, an optimal parameter, and establishing a sea surface scattering coefficient nonlinear regression prediction model.
In the step 5, the performance of the sea surface scattering coefficient nonlinear regression prediction model is evaluated, specifically:
and 5.1, estimating the error between the predicted value and the true value of the sea surface scattering coefficient nonlinear regression prediction model by adopting a Root Mean Square Error (RMSE), wherein the smaller the RMSE value is, the smaller the predicted error of the model is, the better the prediction capability is, and the calculation mode of the Root Mean Square Error (RMSE) is shown as a formula (7):
in the formula (7), n represents the number of samples extracted from the sea surface scattering coefficient nonlinear regression prediction model, y i Representing the true value of the ith sample in the sea surface scattering coefficient nonlinear regression prediction model,representing a predicted value of an ith sample in the sea surface scattering coefficient nonlinear regression prediction model;
step 5.2, determining the coefficient R by using the model 2 To evaluate the interpretation power of the sea surface scattering coefficient nonlinear regression prediction model, R 2 The value range of R is 0-1 2 The closer to 1, the better the fitting degree of the model to the data is, and the better the variability of most of the data can be explained, namely the better the interpretation ability of the model isThe model determines the coefficient R 2 The calculation mode of (2) is shown as the formula (8):
in the formula (8), the amino acid sequence of the compound,representing the mean value of dependent variables in the sea surface scattering coefficient nonlinear regression prediction model;
and 5.3, estimating the fitting goodness and model complexity of the sea surface scattering coefficient nonlinear regression prediction model by adopting a Bayesian information quantity BIC, wherein the lower the value of the Bayesian information quantity BIC is, the smaller the interpretation variable used by the model is, namely the model fitting effect is better, and the calculation mode of the Bayesian information quantity BIC is shown as a formula (9):
BIC=kln(m)-2ln(L) (9)
in the formula (9), L represents a likelihood function, m represents a sample size in the sea surface scattering coefficient nonlinear regression prediction model, and k represents a parameter number.
A system based on a sea surface scattering coefficient fast prediction method, comprising:
the simulation parameter matrix construction module is used for constructing a simulation parameter matrix according to the simulation parameters;
the initial sea surface scattering coefficient data set construction module adopts a multi-parameter wind-wave hybrid improved Venturi spectrum bin scattering method to calculate sea surface scattering coefficients under different simulation parameters in the step 1;
the main parameter component sea surface scattering coefficient data set construction module is used for carrying out main component dimension reduction processing according to the simulation parameter matrix obtained by the simulation parameter matrix construction module, extracting main characteristics, screening main parameter components and constructing a main parameter component sea surface scattering coefficient data set;
the sea surface scattering coefficient nonlinear regression prediction model construction module is used for dividing a data subset of the sea surface scattering coefficient data set of the main parameter component, which is obtained by the sea surface scattering coefficient data set construction module, and solving the optimal parameters of the regression prediction model by combining with cyclic verification to establish a sea surface scattering coefficient nonlinear regression prediction model;
model performance evaluation module, which adopts Root Mean Square Error (RMSE) and model determination coefficient (R) for sea surface scattering coefficient nonlinear regression prediction model 2 And performing performance evaluation on three indexes of the Bayesian information quantity BIC.
A sea surface scattering coefficient fast prediction apparatus comprising:
the memory is used for storing a computer program of the sea surface scattering coefficient rapid prediction method and is equipment readable by a computer;
and the processor is used for realizing the sea surface scattering coefficient rapid prediction method when executing the computer program.
A computer readable medium storing a computer program which, when executed by a processor, enables the implementation of a method for fast predicting sea surface scattering coefficients.
Compared with the prior art, the invention has the beneficial effects that:
1. according to the method, the simulation parameter matrix is constructed, the main component dimension reduction processing is carried out, the main characteristic is extracted, and the main parameter component is screened out to construct the sea surface scattering coefficient data set of the main parameter component, so that the method has the advantage of rapid prediction.
2. According to the sea surface scattering coefficient prediction method, the sea surface scattering coefficient nonlinear regression prediction model is established to predict the sea surface scattering coefficient, so that the prediction effect is more accurate.
3. Compared with the existing sea surface scattering coefficient prediction method based on machine learning, the scattering coefficient nonlinear regression prediction model constructed by the method disclosed by the invention avoids the problem of overfitting generated during multi-parameter high-dimensional mapping, and therefore, the method has the advantage of strong model generalization capability.
In summary, the method and the device have the advantages of fast prediction speed, accurate prediction effect and strong generalization capability by constructing the simulation parameter matrix, carrying out main component dimension reduction processing, extracting main features, screening main parameter components to construct a main parameter component sea surface scattering coefficient data set, solving the optimal parameters of the regression prediction model and establishing the sea surface scattering coefficient nonlinear regression prediction model to predict the sea surface scattering coefficient.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Fig. 2 is a schematic diagram of the present invention.
Fig. 3 is a comparison chart of a sea surface scattering coefficient prediction result and a prediction error obtained by using the present invention, where fig. 3 (a) is a sea surface scattering coefficient prediction result obtained by using the present invention, fig. 3 (b) is a sea surface scattering coefficient prediction result when the dimension k=11 is reduced, fig. 3 (c) is a sea surface scattering coefficient prediction result when the dimension k=10 is reduced, fig. 3 (d) is a sea surface scattering coefficient prediction result when the dimension k=9 is reduced, and fig. 3 (e) is a sea surface scattering coefficient prediction result when the dimension k=8 is reduced.
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings.
The invention provides an embodiment.
Referring to fig. 1 and 2, a method for rapidly predicting sea surface scattering coefficient includes the steps of:
step 1, constructing a simulation parameter matrix according to simulation parameters;
in this embodiment, the simulation parameters included in the simulation parameter matrix are classified into two types, i.e., marine environment parameters and radar parameters, wherein: marine environmental parameters include: wind speed, wind direction, wave height, wave direction, wind area, temperature, salinity and wave period 8 parameters; the radar parameters include: the method comprises the steps of constructing an N-11-dimensional simulation parameter matrix by utilizing 11 total parameters of 3 parameters of an incident wave frequency, a pitch angle and an azimuth angle, wherein the 11 parameters are shown in a formula (1):
in the formula (1), M p Representing a matrix of simulation parameters,n represents the number of samples, U, θ wi 、H、θ wa 、X、T、S、T S F, θ andrespectively represent wind speed, wind direction, wave height, wave direction, wind area, temperature, salinity, wave period, incident wave frequency, pitch angle and azimuth angle, upper angle mark (N) Representing an N-dimensional column vector; according to the method, the simulation parameter matrix is constructed, the main component dimension reduction treatment is carried out, the main characteristic is extracted, and the main parameter component is screened out to construct a sea surface scattering coefficient data set of the main parameter component, so that the method has the advantage of rapid prediction;
step 2, calculating 11 different simulation parameters in the step 1 by adopting a multiparameter wind-wave hybrid improved Venturi spectrum bin scattering method in combination with the sea surface scattering coefficient, so as to construct an initial sea surface scattering coefficient data set; the method comprises the following steps:
step 2.1, performing geometric modeling on a three-dimensional sea surface large-scale contour based on a multi-parameter wind-wave hybrid improved sea spectrum, and obtaining geometric information of the three-dimensional sea surface large-scale wave, wherein the multi-parameter wind-wave hybrid improved sea spectrum is shown as a formula (2):
in the formula (2), S (omega, theta) wa ) Represents the multi-parameter wind-wave mixing improved sea spectrum, omega is the angular frequency of sea wave, T S For the wave period, U is the wind speed 10m above the sea surface, D (θ wa ) Is a Donelan direction function; the expression of the F (U, X) function is shown in the formula (3):
2.2, on the basis of geometric modeling of three-dimensional sea surface large-scale waves, considering superposition of wind-induced small-scale capillary waves, and calculating the scattering coefficient of each bin by adopting a multi-parameter wind-wave mixed bin scattering method, wherein the scattering coefficient is shown in a formula (4):
wherein p, q respectively represent polarization modes of scattered waves and incident waves, ω represents sea wave angular frequency, g represents gravitational acceleration,k represents wave number>And->The normalized wave vectors corresponding to the incident wave and the scattered wave are respectively represented,represents the sea surface slope distribution function, +.>Representing the slope of the tangent plane of the large-scale wavelets mirror point, ε represents the sea surface dielectric constant, Q pq And W is pq Respectively representing mirror image scattering factors and polarization factors corresponding to large-scale wave surface elements, and subscripts pq Representing horizontal or vertical polarization, S wc (U) represents a wind-induced small-scale capillary wave spectrum;
carrying out set average calculation on the scattering coefficients of all the surface elements to obtain the total scattering coefficient of the whole three-dimensional sea surface, wherein the total scattering coefficient is shown as a formula (5):
in the formula (5), m and n respectively represent the number of the sea surface x-axis and y-axis direction surface elements, sigma pq,mn The scattering coefficients corresponding to the (m, n) th bin obtained by calculation in the formula (4) are represented, and Deltax and Deltay represent the sampling intervals in the x and y axis directions of the sea surface respectively;
step 2.3, combining the 11 simulation parameters in the step 1, and calculating three-dimensional sea surface scattering coefficients under different simulation parameters through the calculation formulas in the step 2.1 and the step 2.2 to construct an initial sea surface scattering coefficient data set;
step 3, performing main component dimension reduction processing on the simulation parameter matrix in the step 1, extracting main characteristics, screening main parameter components, and constructing a sea surface scattering coefficient data set of the main parameter components;
step 3.1, carrying out data standardization processing on the simulation parameter matrix, and obtaining a standardized simulation parameter matrix after the original data is subjected to the standardization processing, wherein the standardized simulation parameter matrix is shown as a formula (6):
in the formula (6), the standardized simulation parameter matrix elementM pij Is the element corresponding to the simulation parameter matrix in the formula (1)>Represents the mean value of the j-th column data of the simulation parameter matrix, D pj Representing the standard deviation of the j-th column data of the simulation parameter matrix;
step 3.2, calculating a covariance matrix of the standardized simulation parameter matrix, solving eigenvalues and eigenvectors of the covariance matrix, and selecting the first k main eigenvalues, wherein k is smaller than 11;
step 3.3, reconstructing a data set according to the screened k main characteristic values to obtain a sea surface scattering coefficient data set of the main parameter component;
step 4, dividing a data subset of the sea surface scattering coefficient data set of the main parameter component constructed in the step 3, solving the optimal parameter of the regression prediction model by combining with cyclic verification, and establishing a sea surface scattering coefficient nonlinear regression prediction model;
step 4.1, dividing the main parameter component sea surface scattering coefficient data set obtained in the step 3 into a training set and a testing set by adopting a random division mode, and further dividing the training set into (Num-1) training subsets and 1 verification subset by adopting a random equidistant mode, wherein Num represents the total number of the data subsets;
step 4.2, performing cyclic verification on the training subset and the verification subset in the step 4.1, calculating a generalization error mean value of the regression prediction model, selecting a model parameter with the minimum generalization error mean value from the training subset and the verification subset, namely, an optimal parameter, and establishing a sea surface scattering coefficient nonlinear regression prediction model; compared with the existing sea surface scattering coefficient prediction method based on machine learning, the scattering coefficient nonlinear regression prediction model constructed by the method disclosed by the invention avoids the problem of overfitting generated during multi-parameter high-dimensional mapping, so that the method has the advantage of strong model generalization capability;
step 5, performing cyclic verification on the sea surface scattering coefficient nonlinear regression prediction model to evaluate the performance, and determining a coefficient R by adopting a Root Mean Square Error (RMSE) and the model respectively 2 And three indexes of the Bayesian information quantity BIC are used for evaluating the performance of the prediction model, and specifically:
and 5.1, estimating the error between the predicted value and the true value of the sea surface scattering coefficient nonlinear regression prediction model by adopting a Root Mean Square Error (RMSE), wherein the smaller the RMSE value is, the smaller the predicted error of the model is, the better the prediction capability is, and the calculation mode of the Root Mean Square Error (RMSE) is shown as a formula (7):
in the formula (7), n represents the number of samples extracted from the sea surface scattering coefficient nonlinear regression prediction model, y i Representing the true value of the ith sample in the sea surface scattering coefficient nonlinear regression prediction model,representing sea surface scattering coefficient nonlinearityA predicted value of an ith sample in the regression prediction model;
step 5.2, determining the coefficient R by using the model 2 To evaluate the interpretation power of the sea surface scattering coefficient nonlinear regression prediction model, R 2 The value range of R is 0-1 2 The closer to 1, the better the fitting degree of the model to the data is, the better the variability of most of the data can be explained, namely, the better the interpretation ability of the model is, and the model determines the coefficient R 2 The calculation mode of (2) is shown as the formula (8):
in the formula (8), the amino acid sequence of the compound,representing the mean value of dependent variables in the sea surface scattering coefficient nonlinear regression prediction model;
and 5.3, estimating the fitting goodness and model complexity of the sea surface scattering coefficient nonlinear regression prediction model by adopting a Bayesian information quantity BIC, wherein the lower the value of the Bayesian information quantity BIC is, the smaller the interpretation variable used by the model is, namely the model fitting effect is better, and the calculation mode of the Bayesian information quantity BIC is shown as a formula (9):
BIC=kln(m)-2ln(L) (9)
in the formula (9), L represents a likelihood function, m represents a sample size in the sea surface scattering coefficient nonlinear regression prediction model, and k represents a parameter number.
Referring to FIG. 3, FIG. 3 shows predicted values of sea surface scattering coefficients obtained by the conventional prediction method and the prediction method of the present invention, and the coefficient R is determined by the RMSE and the model described in step 5 in the embodiment of the present invention 2 Three indexes of Bayesian information quantity BIC evaluate the sea surface scattering coefficient nonlinear regression prediction model; wherein, fig. 3 (a) is a sea surface scattering coefficient prediction result obtained by the prior art method, fig. 3 (b) is a sea surface scattering coefficient prediction result obtained by the prediction method of the present invention when the dimension of dimension k=11 is reduced, and fig. 3 (c) is a sea surface scattering coefficient prediction result obtained by the prediction method of the present invention when the dimension k=11 is reducedFig. 3 (d) is a sea surface scattering coefficient prediction result when the dimension k=10 is reduced, fig. 3 (e) is a sea surface scattering coefficient prediction result when the dimension k=9 is reduced, and fig. 3 (e) is a sea surface scattering coefficient prediction result when the dimension k=8 is reduced, wherein a red mark represents an error between a prediction value and a true value; wherein, the prediction conditions are shown in table 1:
TABLE 1
As can be seen from the data in fig. 3 and table 1, the root mean square error RMSE of the sea surface scattering coefficient predicted by the existing method is 1.5045dB, but the predicted result of the method according to the invention is 1.2950dB,1.1267dB,0.9905dB and 1.1699dB respectively according to the difference of dimension reduction, and the training accuracy is improved by 13.92%,25.11%,34.16% and 22.24% by comparison; in addition, the coefficient R is determined by comparing the model of the conventional prediction method with that of the present invention 2 As can be seen from the evaluation result and the Bayesian information quantity BIC evaluation result, the fitting goodness and the interpretation ability of the prediction method are obviously improved after the dimension is reduced, so that the sea surface scattering coefficient prediction method provided by the invention has the advantages of high prediction speed, accurate prediction effect and strong generalization capability compared with the existing prediction method.
Claims (5)
1. The sea surface scattering coefficient rapid prediction method is characterized by comprising the following steps of:
step 1, constructing a simulation parameter matrix according to simulation parameters;
the simulation parameters are divided into two types, namely ocean environment parameters and radar parameters, wherein the ocean environment parameters comprise wind speed, wind direction, wave height, wave direction, wind area, temperature, salinity and wave period; the radar parameters comprise an incident wave frequency, a pitch angle and an azimuth angle;
step 2, calculating the simulation parameters in the step 1 by combining the sea surface scattering coefficients by adopting a multiparameter wind-wave hybrid improved Venturi spectrum bin scattering method, so as to construct an initial sea surface scattering coefficient data set;
step 3, performing main component dimension reduction processing on the simulation parameter matrix in the step 1, extracting main characteristics, screening main parameter components, and constructing a sea surface scattering coefficient data set of the main parameter components;
step 4, dividing a data subset of the sea surface scattering coefficient data set of the main parameter component constructed in the step 3, solving the optimal parameters of the regression prediction model by combining with cyclic verification, and establishing a sea surface scattering coefficient nonlinear regression prediction model;
step 5, determining the coefficient R by adopting root mean square error RMSE and a model respectively 2 Evaluating the performance of the sea surface scattering coefficient nonlinear regression prediction model established in the step 4 by using three indexes of Bayesian information quantity BIC;
in the step 1, a simulation parameter matrix is constructed according to simulation parameters, specifically:
constructing a simulation parameter matrix by using the simulation parameters, wherein the simulation parameter matrix is shown as a formula (1):
in the formula (1), M p Represents a simulation parameter matrix, N represents the number of samples, U and theta wi 、H、θ wa 、X、T、S、T S F, θ andrespectively represent wind speed, wind direction, wave height, wave direction, wind area, temperature, salinity, wave period, incident wave frequency, pitch angle and azimuth angle, upper angle mark (N) Representing an N-dimensional column vector;
in the step 2, an initial sea surface scattering coefficient data set is constructed, specifically:
step 2.1, performing geometric modeling on a three-dimensional sea surface large-scale contour based on a multi-parameter wind-wave hybrid improved sea spectrum, and obtaining geometric information of the three-dimensional sea surface large-scale wave, wherein the multi-parameter wind-wave hybrid improved sea spectrum is shown as a formula (2):
in the formula (2), S (omega, theta) wa ) Represents the multi-parameter wind-wave mixing improved sea spectrum, omega is the angular frequency of sea wave, T S For the wave period, U is the wind speed 10m above the sea surface, D (θ wa ) Is a Donelan direction function; the expression of the F (U, X) function is shown in the formula (3):
2.2, on the basis of geometric modeling of three-dimensional sea surface large-scale waves, considering superposition of wind-induced small-scale capillary waves, and calculating the scattering coefficient of each bin by adopting a multi-parameter wind-wave mixed bin scattering method, wherein the scattering coefficient is shown in a formula (4):
wherein p, q respectively represent polarization modes of scattered waves and incident waves, ω represents sea wave angular frequency, g represents gravitational acceleration,k represents wave number>And->The normalized wave vectors corresponding to the incident wave and the scattered wave are respectively represented,represents the sea surface slope distribution function, +.>Representing the slope of the tangent plane of the large-scale wavelets mirror point, ε represents the sea surface dielectric constant, Q pq And W is pq Respectively representing mirror image scattering factors and polarization factors corresponding to large-scale wave surface elements, and subscripts pq Representing horizontal or vertical polarization, S wc (U) represents a wind-induced small-scale capillary wave spectrum;
carrying out set average calculation on the scattering coefficients of all the surface elements to obtain the total scattering coefficient of the whole three-dimensional sea surface, wherein the total scattering coefficient is shown as a formula (5):
in the formula (5), m and n respectively represent the number of the sea surface x-axis and y-axis direction surface elements, sigma pq,mn The scattering coefficients corresponding to the (m, n) th bin obtained by calculation in the formula (4) are represented, and Deltax and Deltay represent the sampling intervals in the x and y axis directions of the sea surface respectively;
step 2.3, combining the simulation parameters in the step 1, and calculating three-dimensional sea surface scattering coefficients under different simulation parameters through the calculation formulas in the step 2.1 and the step 2.2 to construct an initial sea surface scattering coefficient data set;
in the step 3, a main parameter component sea surface scattering coefficient data set is constructed, specifically:
step 3.1, carrying out data standardization processing on the simulation parameter matrix, and obtaining a standardized simulation parameter matrix after the original data is subjected to the standardization processing, wherein the standardized simulation parameter matrix is shown as a formula (6):
in the formula (6), the standardized simulation parameter matrix elementM pij Is the element corresponding to the simulation parameter matrix in the formula (1)>Represents the mean value of the j-th column data of the simulation parameter matrix, D pj Representing the standard deviation of the j-th column data of the simulation parameter matrix;
step 3.2, calculating a covariance matrix of the standardized simulation parameter matrix, solving eigenvalues and eigenvectors of the covariance matrix, and selecting the first k main eigenvalues, wherein k is smaller than 11;
step 3.3, reconstructing a data set according to the screened k main characteristic values to obtain a sea surface scattering coefficient data set of the main parameter component;
in the step 4, a sea surface scattering coefficient nonlinear regression prediction model is established, specifically:
step 4.1, dividing the main parameter component sea surface scattering coefficient data set obtained in the step 3 into a training set and a testing set by adopting a random division mode, and further dividing the training set into (Num-1) training subsets and 1 verification subset by adopting a random equidistant mode, wherein Num represents the total number of the data subsets;
and 4.2, performing cyclic verification on the training subset and the verification subset in the step 4.1, calculating a generalization error mean value of the regression prediction model, selecting a model parameter with the minimum generalization error mean value from the training subset and the verification subset, namely, an optimal parameter, and establishing a sea surface scattering coefficient nonlinear regression prediction model.
2. The method for rapidly predicting sea surface scattering coefficient according to claim 1, wherein the performance of the sea surface scattering coefficient nonlinear regression prediction model is evaluated in step 5, specifically:
and 5.1, estimating the error between the predicted value and the true value of the sea surface scattering coefficient nonlinear regression prediction model by adopting a Root Mean Square Error (RMSE), wherein the smaller the RMSE value is, the smaller the predicted error of the model is, the better the prediction capability is, and the calculation mode of the Root Mean Square Error (RMSE) is shown as a formula (7):
in the formula (7), n represents the number of samples extracted from the sea surface scattering coefficient nonlinear regression prediction model, y i Representing the true value of the ith sample in the sea surface scattering coefficient nonlinear regression prediction model,representing a predicted value of an ith sample in the sea surface scattering coefficient nonlinear regression prediction model;
step 5.2, determining the coefficient R by using the model 2 To evaluate the interpretation power of the sea surface scattering coefficient nonlinear regression prediction model, R 2 The value range of R is 0-1 2 The closer to 1, the better the fitting degree of the model to the data is, the better the variability of most of the data can be explained, namely, the better the interpretation ability of the model is, and the model determines the coefficient R 2 The calculation mode of (2) is shown as the formula (8):
in the formula (8), the amino acid sequence of the compound,representing the mean value of dependent variables in the sea surface scattering coefficient nonlinear regression prediction model;
and 5.3, estimating the fitting goodness and model complexity of the sea surface scattering coefficient nonlinear regression prediction model by adopting a Bayesian information quantity BIC, wherein the lower the value of the Bayesian information quantity BIC is, the smaller the interpretation variable used by the model is, namely the model fitting effect is better, and the calculation mode of the Bayesian information quantity BIC is shown as a formula (9):
BIC=kln(m)-2ln(L)(9)
in the formula (9), L represents a likelihood function, m represents a sample size in the sea surface scattering coefficient nonlinear regression prediction model, and k represents a parameter number.
3. A system based on the sea surface scattering coefficient rapid prediction method according to any one of claims 1-2, characterized by comprising:
the simulation parameter matrix construction module is used for constructing a simulation parameter matrix according to the simulation parameters;
the initial sea surface scattering coefficient data set construction module is used for calculating sea surface scattering coefficients under different simulation parameters in the step 1 by adopting a multi-parameter wind-wave hybrid improved Venturi spectrum bin scattering method;
the main parameter component sea surface scattering coefficient data set construction module is used for carrying out main component dimension reduction processing according to the simulation parameter matrix obtained by the simulation parameter matrix construction module, extracting main characteristics, screening main parameter components and constructing a main parameter component sea surface scattering coefficient data set;
the sea surface scattering coefficient nonlinear regression prediction model construction module is used for dividing a data subset of the sea surface scattering coefficient data set of the main parameter component, which is obtained by the sea surface scattering coefficient data set construction module, and solving the optimal parameters of the regression prediction model by combining with cyclic verification to establish a sea surface scattering coefficient nonlinear regression prediction model;
model performance evaluation module, which adopts Root Mean Square Error (RMSE) and model determination coefficient (R) for sea surface scattering coefficient nonlinear regression prediction model 2 And performing performance evaluation on three indexes of the Bayesian information quantity BIC.
4. A sea surface scattering coefficient fast prediction apparatus, comprising:
a memory for storing a computer program of a method for rapid sea surface scattering coefficient prediction according to any one of claims 1-2, a device readable by a computer;
a processor for implementing a method for rapidly predicting sea surface scattering coefficient as claimed in any one of claims 1-2 when executing said computer program.
5. A computer readable medium, characterized in that the computer readable storage medium stores a computer program, which when executed by a processor is capable of implementing a sea surface scattering coefficient fast prediction method according to any of claims 1-2.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311063177.1A CN117216954B (en) | 2023-08-23 | 2023-08-23 | Sea surface scattering coefficient rapid prediction method, system, equipment and medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311063177.1A CN117216954B (en) | 2023-08-23 | 2023-08-23 | Sea surface scattering coefficient rapid prediction method, system, equipment and medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117216954A CN117216954A (en) | 2023-12-12 |
CN117216954B true CN117216954B (en) | 2024-03-26 |
Family
ID=89039847
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311063177.1A Active CN117216954B (en) | 2023-08-23 | 2023-08-23 | Sea surface scattering coefficient rapid prediction method, system, equipment and medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117216954B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5978344B1 (en) * | 2015-05-07 | 2016-08-24 | 東芝電波プロダクツ株式会社 | Radar image simulation device, sea surface clutter simulation method, and sea surface clutter simulation program |
CN112130122A (en) * | 2020-09-01 | 2020-12-25 | 武汉大学 | Method for estimating sea surface scattering coefficient of space-based high-frequency radar |
CN114004158A (en) * | 2021-11-02 | 2022-02-01 | 西安电子科技大学 | Sea surface electromagnetic scattering prediction method based on genetic algorithm optimization support vector machine |
CN115421120A (en) * | 2022-08-04 | 2022-12-02 | 中国人民解放军91977部队 | Modeling method for real sea surface electromagnetic scattering environment |
CN115906667A (en) * | 2022-12-29 | 2023-04-04 | 西安电子科技大学 | Marine environment parameter inversion model construction method and device |
-
2023
- 2023-08-23 CN CN202311063177.1A patent/CN117216954B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5978344B1 (en) * | 2015-05-07 | 2016-08-24 | 東芝電波プロダクツ株式会社 | Radar image simulation device, sea surface clutter simulation method, and sea surface clutter simulation program |
CN112130122A (en) * | 2020-09-01 | 2020-12-25 | 武汉大学 | Method for estimating sea surface scattering coefficient of space-based high-frequency radar |
CN114004158A (en) * | 2021-11-02 | 2022-02-01 | 西安电子科技大学 | Sea surface electromagnetic scattering prediction method based on genetic algorithm optimization support vector machine |
CN115421120A (en) * | 2022-08-04 | 2022-12-02 | 中国人民解放军91977部队 | Modeling method for real sea surface electromagnetic scattering environment |
CN115906667A (en) * | 2022-12-29 | 2023-04-04 | 西安电子科技大学 | Marine environment parameter inversion model construction method and device |
Non-Patent Citations (3)
Title |
---|
不同海域多频段海杂波特性差异与主要影响因素分析;吴涛;中国博士学位论文全文数据库 (信息科技辑);20200731;第I136-62页 * |
双尺度海面电磁散射特性主影响因素分析;刘悦 等;2023年全国微波毫米波会议论文汇编(三);20230514;第1-3页 * |
基于文氏改进谱的二维粗糙海面模型及其电磁散射研究;吴庚坤 等;物理学报;20141231;第63卷(第13期);第134203-1-13页 * |
Also Published As
Publication number | Publication date |
---|---|
CN117216954A (en) | 2023-12-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102109495A (en) | Method for classifying types of mixed seabed sediment based on multi-beam sonar technology | |
Huang et al. | Wave height estimation from X-band nautical radar images using temporal convolutional network | |
Liu et al. | Multitarget state extraction for the PHD filter using MCMC approach | |
CN111898749A (en) | Unmanned aerial vehicle track prediction method based on neural network | |
CN111159937A (en) | Near-shore nonlinear sinusoidal wave microwave scattering characteristic analysis method | |
Zhao et al. | Research on evaporation duct height prediction based on back propagation neural network | |
Palmer et al. | Complexity in the atmosphere | |
Zhang et al. | Improved Continuous Wavelet Transform for Modal Parameter Identification of Long‐Span Bridges | |
Chen et al. | LFM signal detection and estimation based on deep convolutional neural network | |
CN115097451A (en) | Sea wave multi-parameter inversion method and system based on SAR satellite data | |
CN114740441A (en) | Low-slow small-target radar echo identification method based on small samples | |
CN114218860A (en) | Laser radar wind measurement motion compensation method and system based on machine learning | |
CN114004158A (en) | Sea surface electromagnetic scattering prediction method based on genetic algorithm optimization support vector machine | |
CN117216954B (en) | Sea surface scattering coefficient rapid prediction method, system, equipment and medium | |
Wang et al. | Radar HRRP target recognition via statistics‐based scattering centre set registration | |
CN112766381A (en) | Attribute-guided SAR image generation method under limited sample | |
CN115774248A (en) | Sea surface small target detection method and system based on multiple characteristics of graph | |
CN113960551B (en) | Clutter image generation method and target detection method for SAR image | |
CN115308705A (en) | Multi-pose extremely narrow pulse echo generation method based on generation countermeasure network | |
CN111428420B (en) | Method and device for predicting sea surface flow velocity, computer equipment and storage medium | |
Liu et al. | Mining weather data using fuzzy cluster analysis | |
CN110135280B (en) | Multi-view SAR automatic target recognition method based on sparse representation classification | |
CN113762203A (en) | Cross-domain self-adaptive SAR image classification method, device and equipment based on simulation data | |
Liu et al. | Arctic Sea-ice type recognition based on the surface wave investigation and monitoring instrument of the China-French Ocean satellite | |
CN110806444A (en) | Seabed sediment recognition and classification method based on shallow stratum profiler and SVM |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |