Bragg peak screening, pseudo peak determination and missing value joint estimation method for sea state inversion and sea state inversion method

Abstract

The invention discloses a Bragg peak screening, pseudo peak determination and missing value joint estimation method for sea state inversion and a sea state inversion method, and belongs to the technical field of sea state inversion. The objective of the invention is to solve the problem of high false peak misjudgment rate of a Bragg peak detection value in existing sea state inversion. According to the method, the Bragg peak offset is obtained based on the maximum ocean current velocity of the area needing to be detected, the first-order peak appearance range is determined, the average noise amplitude is calculated based on the background Doppler unit, the initial detection threshold value is determined, and the extreme points in the range are counted through the threshold value to achieve matching. Selecting a pair of candidate Bragg peaks of which the theoretical frequency difference value with the positive and negative Bragg peaks is closest to the theoretical difference value to obtain candidate Bragg peaks, determining a detection window range according to the positions of the candidate Bragg peaks, and judging the candidate peaks beyond and within the range and exceeding a dynamic detection threshold value as false peaks; and then, false peaks are eliminated, spatial correlation and statistical inference are fused through an LW-EM algorithm, Bragg peak joint estimation is realized, and then sea state inversion is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of sea state inversion, and relates to a sea state inversion estimation method for the phenomenon of Bragg peak missed detection or misjudgment under ionospheric clutter interference. Background Art

[0002] High-frequency ground-wave radar's wide-area monitoring capabilities can cover exclusive economic zones and disputed waters, acquiring ocean dynamic parameters in real time and providing data support for maritime traffic control and resource development. Particularly in sensitive waters, its all-weather monitoring capabilities can effectively identify illegal fishing, oil and gas theft, and other activities, helping to enhance the country's maritime law enforcement capabilities. This technology's long-term continuous observation capabilities can provide a scientific basis for addressing climate change and protecting biodiversity. Its environmental significance is becoming increasingly prominent in the context of carbon neutrality and sustainable development. The technology can also provide information on sea surface currents and waves, significantly improving the timeliness and accuracy of disaster warnings.

[0003] High-frequency surface wave radar (HFSWR), a core technology for marine environmental monitoring, receives electromagnetic scattering echoes from the ocean surface, enabling the dynamic inversion of key ocean parameters such as wind speed, significant wave height, and currents. Between 1966 and 1972, Wait and Barrick proposed the Bragg scattering principle and established first- and second-order scattering equations, providing a solid theoretical foundation for the subsequent inversion of ocean state parameters. Its core mechanism is based on the analysis of the Bragg scattering peak in the radar range-Doppler spectrum (RD spectrum). This characteristic peak is formed by the resonant scattering of short sea surface gravity waves and radar electromagnetic waves. Its frequency offset, spectral peak amplitude, and spectral energy distribution are directly related to ocean dynamic parameters.

[0004] At present, the maximum value method and the matching peak method are two commonly used methods for detecting Bragg peaks. The basic principle of the matching peak method is to use the fact that the frequency difference between the positive and negative Bragg peaks is fixed to 2 in theory. This characteristic ensures that the positive and negative frequency difference remains constant regardless of the Bragg peak's shift. By determining the Bragg peak's range, setting a detection threshold, searching for extreme points within the range, and then matching the statistical extreme points, we select a pair of Bragg peaks whose positive and negative frequency difference is closest to the theoretical difference.

[0005] However, interference from ionospheric clutter severely limits the reliability and accuracy of this technology. Ionospheric clutter, generated by the interaction of electromagnetic waves with ionospheric irregularities, has a spectral characteristic that overlaps highly with the ocean Bragg peaks. This causes the ocean information in the radar echo signal to be obscured or distorted by the clutter energy. This phenomenon directly undermines the physical basis for extracting ocean parameters, exposing the inversion of ocean state to the risk of systematic deviations or even failure.

[0006] At the same time, after ionospheric clutter suppression, some range cells still have residual clutter contamination. The above two methods will cause the Bragg peak detection value to have false peak misjudgment and effective peak missed detection, resulting in a high false peak misjudgment rate and a high effective peak missed detection rate. Such abnormal data will directly affect the inversion accuracy of sea state parameters.

[0007] To address the aforementioned issues with existing technologies, a method is urgently needed that can accurately detect Bragg peaks, effectively eliminate false peaks, and jointly estimate Bragg resonance parameters in the presence of ionospheric clutter interference or after ionospheric clutter suppression. This method can repair data affected by clutter and residual clutter, improve the spatial continuity and inversion accuracy of ocean state inversion data, ensure the reliability and accuracy of high-frequency ground wave radar in ocean state monitoring, and provide higher-quality technical support for marine environmental monitoring and related applications. Summary of the Invention

[0008] The present invention aims to solve the problem that the Bragg peak detection value in the existing sea state inversion has a high false peak misjudgment rate, thereby affecting the accuracy of the sea state inversion.

[0009] A Bragg peak screening method for sea state inversion, comprising:

[0010] The maximum value of the Bragg peak offset is obtained based on the maximum velocity of the ocean current in the area to be detected. , determine the range of the first-order peak; calculate the average noise amplitude based on all background Doppler units of the high-frequency ground wave radar , and determine the threshold for preliminary detection of Bragg peaks. The extreme points within the range of Bragg peaks are statistically recorded using this threshold to search for extreme points. The statistical extreme points are matched to select a pair of Bragg peaks that are closest to the theoretical frequency difference between the positive and negative Bragg peaks, and then obtain candidate Bragg peaks.

[0011] Furthermore, the first-order peak appears in the following range:

[0012] The positive first-order Bragg peak is , the negative first-order Bragg peak is ; is the theoretical frequency.

[0013] Furthermore, the theoretical frequency difference between the positive and negative Bragg peaks is , is the theoretical frequency.

[0014] A method for determining a pseudo-peak of a Bragg peak for sea state inversion comprises the following steps:

[0015] The position of the candidate Bragg peak obtained based on the Bragg peak screening method for sea state inversion is , the corresponding spectrum amplitude is ; Established The detection window at the center ,in is the frequency window width, are positive integer coefficients, is the Doppler resolution;

[0016] The candidate Bragg peaks outside the detection window are judged as Bragg pseudo peaks;

[0017] In the detection window Statistics satisfy the range Frequency points , According to the spectrum amplitude Determine the dynamic detection threshold; when it meets When , the current candidate peak is determined to be a Bragg pseudo peak. It is the decision threshold set according to the statistical characteristics of ocean echo.

[0018] Furthermore, according to the spectrum amplitude Determined dynamic detection threshold , is the amplitude margin parameter.

[0019] A missing value joint estimation method for Bragg peaks in sea state inversion is disclosed. The method uses the aforementioned pseudo peak determination method for Bragg peaks in sea state inversion to determine and remove pseudo peaks. Then, the LW-EM algorithm is used to fuse spatial correlation and statistical inference to achieve joint estimation of Bragg peaks. The specific process includes the following steps:

[0020] Step S31, initialization parameters:

[0021] The radar transmits pulse electromagnetic waves and receives echoes, and divides the detection range into There are equally spaced distance segments, each of which is a distance unit; let The observed amplitude value of each distance unit is , forming a sequence , define missing data as , the missing data are pseudo peaks or missed Bragg peaks;

[0022] Step S32, calculating the expected step and the maximization step:

[0023] Expectation step, or E step: Calculate the local posterior expectation ; is the posterior weight, is the spatial correlation window function, is the set of parameters to be estimated for the kth iteration update; 、 denote the mean and variance of the j-th position, respectively. Indicates the The prior probability that the position is a true Bragg peak, is the probability density function of the normal distribution, is the impulse function;

[0024] The maximization step is Step 1: Update parameters; use adaptive sliding window estimation:

[0025]

[0026] 、 is the missing value estimation result obtained in the kth and k+1th iterations;

[0027] Step S33, repeat the iterative process of S32 until the iteration converges, and when the iteration converges, the final and ; Then the missing values ​​are jointly estimated, and the missing values ​​are estimated as follows: , Estimate results for missing values.

[0028] Furthermore, the spatial correlation window function ,in is the window width parameter, and j is the position order.

[0029] Furthermore, the convergence condition of step S33 is as follows:

[0030]

[0031] in, To set the convergence error, Represents the missing value estimation result in the kth iteration.

[0032] A sea state inversion method performs sea state inversion on Bragg peak parameters obtained based on a missing value joint estimation method for Bragg peaks used for sea state inversion.

[0033] Furthermore, the radial velocity calculation result obtained by sea state inversion is , is the radar wavelength, is the difference between the theoretical Bragg peak frequency and the observed frequency.

[0034] The beneficial effects of the present invention are:

[0035] 1. The present invention significantly improves the Bragg peak detection accuracy in sea echo data processed by ionospheric clutter detection, effectively reducing the false peak misjudgment rate and the effective peak missed detection rate.

[0036] 2. In terms of the accuracy of joint estimation of sea state parameters, the estimation is more accurate, and the joint estimation of multiple parameters such as radial flow velocity, wind direction, and wind speed is realized at the same time, which solves the problem of parameter distortion under ionospheric interference and improves the overall accuracy of sea state inversion.

[0037] 3. In terms of robustness, the system jointly estimates and fills in missing Bragg peaks (after missing or false peaks are removed), and is highly robust to different sea conditions (such as current velocity changes and wave height fluctuations), providing technical support for the all-weather, high-precision application of high-frequency ground wave radar in ocean remote sensing. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is the RD spectrum of the measured data on August 23, 2023. The red line is the Bragg peak detection result.

[0039] Figure 2 It is the RD spectrum and detection result after ionospheric clutter suppression processing.

[0040] Figure 3 It is the amplitude on the left side obtained by performing Bragg pseudo-peak detection and estimating the data using the LW-EM method. The red circle represents the original Bragg peak detection result, and the estimated data is the blue broken line.

[0041] Figure 4 is the right amplitude obtained by performing Bragg pseudo-peak detection and LW-EM method to estimate the data.

[0042] Figure 5 It is the left frequency obtained by performing Bragg pseudo-peak detection and LW-EM method to estimate the data.

[0043] Figure 6 is the right-side frequency obtained by performing Bragg pseudo-peak detection and LW-EM method to estimate the data.

[0044] Figure 7 is the estimated 3D model of the front Bragg peak.

[0045] Figure 8 is the estimated Bragg peak 3D model.

[0046] Figure 9 This is a top view of the 3D model of Bragg Peak before estimation.

[0047] Figure 10 This is a top view of the estimated Bragg peak 3D model.

[0048] Figure 11 is the inversion result of radial flow velocity. DETAILED DESCRIPTION

[0049] In order to specifically address the shortcomings of the existing technologies in Bragg peak detection accuracy, joint parameter estimation, and robustness of sea state inversion, the present invention proposes a Bragg resonance parameter joint estimation and sea state inversion method. By constructing a parameter joint optimization model and utilizing the physical correlation constraints between Bragg scattering parameters, the abnormal Bragg peaks detected are iteratively corrected and the missing peaks are reconstructed. A multi-stage processing strategy is adopted to first establish a pseudo-peak recognition mechanism based on spectral feature criteria, and then implement algorithms such as LW-EM to achieve multi-parameter joint optimal estimation, ultimately achieving improved sea state inversion performance. Specific implementation method one:

[0051] This embodiment is a sea state inversion method based on joint estimation of Bragg resonance parameters, which includes four main parts: a Bragg peak screening method for sea state inversion corresponding to step S1, a pseudo-peak determination method (including pseudo-peak removal) for Bragg peaks in sea state inversion corresponding to step S2, a missing value joint estimation method for Bragg peaks in sea state inversion corresponding to step S3, and a sea state inversion method corresponding to step S4.

[0052] The specific process of this embodiment includes the following steps:

[0053] Step S1 : detecting Bragg peaks using a matching peak method, and determining candidate Bragg peaks using the same offset characteristics of the Bragg peaks on both sides.

[0054] (1) Determine the range of the Bragg peak: Due to the existence of ocean surface currents, the Bragg peak is A certain degree of offset will occur. This offset is related to the ocean current velocity, which makes the actual detection value of the Bragg peak have a certain detection range. The maximum value of the Bragg peak offset can be obtained by calculating the maximum ocean current velocity in the area to be detected. , then the range of the first-order peak is defined as: the positive first-order Bragg peak is , the negative first-order Bragg peak is .

[0055] (2) Determine the threshold for Bragg peak detection: The background Doppler unit is the frequency unit in the range-Doppler spectrum (RD spectrum) of the high-frequency ground wave radar (HFSWR) that does not contain the Bragg scattering peak or other target echo signals. It mainly reflects the energy distribution of background signals such as environmental noise and incoherent clutter. Select all background Doppler units to calculate the average noise amplitude On this basis, the set minimum signal-to-noise ratio is added as the threshold for preliminary detection of Bragg peaks.

[0056] (3) Search for extreme points within the range: By detecting the threshold of the Bragg peak, the extreme points within the range where the Bragg peak exists are statistically recorded. Generally, multiple extreme points can be searched.

[0057] (4) Perform peak matching to select candidate Bragg peaks: The theoretical frequency difference between positive and negative Bragg peaks is fixed to Regardless of the Bragg peak's shift, its positive and negative frequency difference remains constant. Therefore, we match the extreme points counted in the previous step and select a pair of Bragg peaks whose positive and negative frequency differences are closest to the theoretical difference. If they meet reasonable error conditions, they are identified as candidate Bragg peaks. Otherwise, the Bragg peak is considered undetectable.

[0058] Step S2: Establish a pseudo-peak recognition mechanism based on spectrum feature criteria to screen and eliminate abnormal detection peaks.

[0059] Bragg pseudo-peak detection is performed based on the Doppler characteristics of the Bragg peak at a fixed distance unit. Each Bragg peak is judged based on the distribution characteristics of the peak neighborhood energy, and pseudo-Bragg peaks caused by clutter interference or noise are eliminated. The specific process includes the following steps:

[0060] Step S21, establishing a central detection window:

[0061] Assume that the candidate Bragg peak position obtained by the initial detection is , the corresponding spectrum amplitude is . Established The detection window at the center ,in is the frequency window width, are positive integer coefficients, is the Doppler resolution. The dynamic detection threshold is defined as:

[0062]

[0063] in, It is the amplitude margin parameter, which is selected according to different data and can be set to a fixed value of about 4dB.

[0064] In step S22, candidate Bragg peaks outside the detection window are determined to be Bragg pseudo peaks and are removed.

[0065] In the detection window Statistics satisfy the range Frequency points .

[0066]

[0067] Where, is the indicator function.

[0068] Set the judgment threshold according to the statistical characteristics of ocean echoes , when satisfied When the current candidate peak is determined to be a Bragg pseudo peak, the algorithm performs a removal operation. Different thresholds can be set for the RD spectrum data in different clutter environments. Corresponding to positive and negative Bragg peaks.

[0069] This method can be used to further screen the Bragg peak detection results of the RD spectrum after ionospheric clutter suppression, remove the Bragg pseudo-peaks, and provide a basis for the subsequent joint estimation of Bragg resonance parameters.

[0070] Step S3, based on the Locally Weighted Expectation-Maximization (LW-EM) algorithm, achieves joint estimation of Bragg resonance parameters by fusing spatial correlation and statistical inference. For the missing data that may be generated after S2 removes pseudo-peaks, the LW-EM algorithm is used to fuse spatial correlation and statistical inference to achieve joint estimation of Bragg peaks. At the same time, the physical correlation between Bragg scattering parameters is used to iteratively correct the residual outliers after S2 processing to ensure that the parameters conform to the laws of ocean dynamics. The specific process includes the following steps:

[0071] Step S31, initialization parameters:

[0072] The radar transmits pulse electromagnetic waves and receives echoes, and divides the detection range into There are equidistant distance segments, each of which is a distance unit. The observed amplitude value of each distance unit is , forming a sequence , define missing data as , the missing data are pseudo peaks or missed Bragg peaks. Introducing spatial correlation window function ,in is the window width parameter, j is the position order. Introduce the local Gaussian mixture model:

[0073]

[0074] In the formula is the parameter set to be estimated, 、 denote the mean and variance of the j-th position, respectively. Indicates the The prior probability that the position is a true Bragg peak, is the probability density function of the normal distribution, is the impulse function.

[0075] Step S32, calculating the expectation step (E step) and the maximization step (M step):

[0076] Expectation step (E step): Calculate the local posterior expectation;

[0077]

[0078] The posterior weight is:

[0079]

[0080] in is the updated parameter set to be estimated.

[0081] Maximization step ( Step 1): Update parameters;

[0082] Adaptive sliding window estimation:

[0083]

[0084] Step S33, repeat the iterative process of S32 until the iteration converges, and when the iteration converges, the final and ; Then the missing values ​​are jointly estimated, and the missing values ​​are estimated as follows:

[0085]

[0086] in, is the missing value estimation result, and is the iteration result.

[0087] This improved method captures the distance correlation of sea state parameters through an exponential decay window and uses cubic spline interpolation for initial value filling, which improves the robustness of the model. It also uses dynamic convergence and a stopping criterion based on relative mean square error (RMSE):

[0088]

[0089] in, To set the convergence error, Represents the missing value estimation result in the kth iteration.

[0090] Step S4, perform sea state inversion:

[0091] Steps S2 and S3 provide high-quality Bragg peak parameters for S4 sea state inversion through the progressive processing of "pseudo-peak removal-parameter optimization". The present invention establishes a radial velocity inversion error analysis model to evaluate the joint optimization effect of the pseudo-peak identification mechanism after ionospheric clutter processing and the LW-EM parameter estimation method in strong interference scenarios.

[0092] The radial flow velocity is calculated as

[0093]

[0094] in, is the radar wavelength, is the difference between the theoretical Bragg peak frequency and the observed frequency.

[0095] Aiming at the Bragg peak missed detection and misjudgment phenomenon in some distance units in sea state inversion data, the present invention uses a Bragg pseudo-peak detection mechanism based on spectral feature criterion and the LW-EM algorithm to jointly estimate the Bragg resonance parameters, thereby realizing the recovery of Bragg peak contaminated data and ensuring the inversion accuracy of sea state parameters.

[0096] Example:

[0097] The measured data on August 23, 2023 is selected, and its high-frequency ground wave radar RD spectrum is as follows Figure 1 As shown in the figure, there is a large area of ​​high-intensity ionospheric clutter at 100km to 150km, which almost submerges the Bragg peak. There is clutter with a shallow amplitude at 225km, but the amplitude of the Bragg peak is already small within this distance, and there is no need to suppress the ionospheric clutter in this part.

[0098] Image fusion clutter suppression processing based on optimization algorithm is implemented for 100-150km distance cells. The processing results are as follows Figure 2 As shown in Figure 2, experiments show that ionospheric clutter energy in this region is effectively suppressed, allowing the previously obscured Bragg scattering peak to fully reveal itself. However, in the vicinity of 150 km, the signal intensity of the left Bragg peak in the original signal is weakened due to its weak amplitude. Therefore, a Bragg peak detection algorithm is required to extract targeted parameters to accurately assess the characteristics of the local ocean echo.

[0099] The Bragg peak detection results of the radar measured data after ionospheric clutter suppression are used to perform Bragg pseudo-peak detection and LW-EM method to estimate the data. The amplitude results are as follows: Figure 3 and Figure 4 As shown, the frequency results are as follows Figure 5 and Figure 6 shown.

[0100] The red circle represents the original Bragg peak detection result, and the estimated data is the blue broken line. There is no original data in some distance units, which indicates that the Bragg peak was not detected by the matching peak method or the Bragg pseudo-peak was eliminated by Bragg pseudo-peak detection.

[0101] The left Bragg peak has a large range of zero values ​​at the 50-70 range bins. The estimated frequency conforms to the continuity of the Bragg peak, with slow frequency changes and no sudden changes. The amplitude conforms to the correlation between the Bragg peak amplitude and the range bin, also without sudden changes. The estimated amplitude range is between -60 and -65, which is within the normal amplitude range of the Bragg peak. The left Bragg peak result is better. The right Bragg peak has small missing values ​​near the 40, 52, and 63 range bins. The estimated frequency changes also follow the Bragg peak frequency variation pattern. The estimated amplitude changes also have a certain range correlation, compensating for the missing Bragg peak caused by ionospheric clutter suppression.

[0102] The 3D model of the Bragg peak before and after estimation is as follows Figure 7 and Figure 8 As shown, the top view is as follows Figure 9 and Figure 10 As shown in the figure, after the joint parameter estimation, the frequency and amplitude characteristics of the Bragg peaks on both sides are consistent with the characteristics of ocean echoes. The continuity of the Bragg peaks before and after estimation is improved, and the results use other data to support the original data characteristics. The data truncation of the left Bragg peak at the 40-70 distance unit is perfectly restored, and the fragmented missing Bragg peak on the right is also repaired.

[0103] The radial velocity inversion results are as follows: Figure 11 As shown in the figure, the solid line is the radial velocity inversion result after processing, and the dashed line is the radial velocity inversion result using the Bragg peak extracted from the original RD spectrum. It can be clearly seen that compared to direct Bragg peak detection, the radial velocity inversion result of the processed data meets the requirements of radial velocity in conventional sea areas, and its variation characteristics are consistent with the characteristics of ocean dynamics. There is no velocity jump, and the velocity magnitude and direction of each distance unit are continuous and correlated.

[0104] In summary, the pseudo-peak identification mechanism based on spectral feature criteria and the LW-EM method can realize the joint estimation of Bragg resonance parameters, and the repaired data conforms to the characteristics of ocean echoes, so that the Bragg peak data can be better used in subsequent ocean state inversion work.

[0105] Each embodiment in this specification is described in a related manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiment is generally similar to the method embodiment, so the description is relatively simple. For related parts, refer to the description of the method embodiment.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention are included in the scope of protection of the present invention.

Claims

1. A Bragg peak screening method for sea state inversion, characterized in that: include: The maximum value of the Bragg peak offset is obtained based on the maximum velocity of the ocean current in the area to be detected. , determine the range of the first-order peak; calculate the average noise amplitude based on all background Doppler units of the high-frequency ground wave radar , and determine the threshold for preliminary detection of Bragg peaks. The extreme points within the range of Bragg peaks are statistically recorded using this threshold to search for extreme points. The statistical extreme points are matched to select a pair of Bragg peaks that are closest to the theoretical frequency difference between the positive and negative Bragg peaks, and then obtain candidate Bragg peaks.

2. A Bragg peak screening method for sea state inversion according to claim 1, characterized in that: The first-order peak appears in the following range: The positive first-order Bragg peak is , the negative first-order Bragg peak is ; is the theoretical frequency.

3. A Bragg peak screening method for sea state inversion according to claim 1 or 2, characterized in that: The theoretical frequency difference between the positive and negative Bragg peaks is , is the theoretical frequency.

4. A method for determining pseudo-peaks of Bragg peaks used in ocean state inversion, characterized in that: The following steps are involved: The position of the candidate Bragg peak obtained by the Bragg peak screening method for sea state inversion according to any one of claims 1 to 3 is , the corresponding spectrum amplitude is ; Established The detection window at the center ,in is the frequency window width, are positive integer coefficients, is the Doppler resolution; The candidate Bragg peaks outside the detection window are judged as Bragg pseudo peaks; In the detection window Statistics satisfy the range Frequency points , According to the spectrum amplitude Determine the dynamic detection threshold; when it meets When , the current candidate peak is determined to be a Bragg pseudo peak. It is the decision threshold set according to the statistical characteristics of ocean echo.

5. The method for determining a pseudo peak of a Bragg peak for sea state inversion according to claim 4, characterized in that: According to the spectrum amplitude Determined dynamic detection threshold , is the amplitude margin parameter.

6. A missing value joint estimation method for Bragg peaks in ocean state inversion, characterized by: The pseudo peak determination method for Bragg peak inversion of sea state according to claim 4 or 5 is used to determine and eliminate pseudo peaks, and then the spatial correlation and statistical inference are integrated through the LW-EM algorithm to achieve joint estimation of Bragg peaks. The specific process includes the following steps: Step S31, initialization parameters: The radar transmits pulse electromagnetic waves and receives echoes, and divides the detection range into There are equally spaced distance segments, each of which is a distance unit; let The observed amplitude value of each distance unit is , forming a sequence , define missing data as , the missing data are pseudo peaks or missed Bragg peaks; Step S32, calculating the expected step and the maximization step: Expectation step, or E-step: Calculate the local posterior expectation ; is the posterior weight, is the spatial correlation window function, is the set of parameters to be estimated for the kth iteration update; 、 denote the mean and variance of the j-th position, respectively. Indicates the The prior probability that the position is a true Bragg peak, is the probability density function of the normal distribution, is the impulse function; The maximization step is Step 1: Update parameters; use adaptive sliding window estimation: 、 is the missing value estimation result obtained in the kth and k+1th iterations; Step S33, repeat the iterative process of S32 until the iteration converges, and when the iteration converges, the final and ; Then the missing values ​​are jointly estimated, and the missing values ​​are estimated as follows: , Estimate results for missing values.

7. A missing value joint estimation method for Bragg peaks in sea state inversion according to claim 6, characterized in that: The spatial correlation window function ,in is the window width parameter, and j is the position order.

8. A missing value joint estimation method for Bragg peaks in sea state inversion according to claim 6, characterized in that: The convergence conditions of step S33 are as follows: in, To set the convergence error, Represents the missing value estimation result in the kth iteration.

9. A sea state inversion method, characterized in that: Sea state inversion is performed on the Bragg peak parameters obtained based on the missing value joint estimation method for Bragg peak inversion of sea state described in claim 6.

10. A sea state inversion method according to claim 9, characterized in that: The radial velocity calculation result obtained by sea state inversion is: , is the radar wavelength, is the difference between the theoretical Bragg peak frequency and the observed frequency.

Images