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

Abstract

This invention relates to a method for Bragg peak screening, spurious peak identification, and joint estimation of missing values ​​for ocean state inversion, belonging to the field of ocean state inversion technology. To address the high false peak misclassification rate in existing Bragg peak detection methods for ocean state inversion, this invention determines the range of first-order peak occurrence by obtaining the Bragg peak offset based on the maximum ocean current velocity in the detection area. It calculates the average noise amplitude based on background Doppler cells to determine a preliminary detection threshold. This threshold is used to statistically match extreme points within the range, selecting a pair whose frequency difference is closest to the theoretical difference between positive and negative Bragg peaks to obtain candidate Bragg peaks. The detection window range is determined based on their positions, and candidate peaks outside or exceeding the dynamic detection threshold within the range are judged as spurious peaks. False peaks are then removed, and the LW-EM algorithm is used to fuse spatial correlation and statistical inference to achieve joint estimation of Bragg peaks, thereby realizing ocean state inversion.

Description

Technical Field

[0001] This invention belongs to the field of ocean state inversion technology and relates to an ocean state inversion estimation method for the phenomenon of missed detection or misjudgment of Bragg peak under ionospheric clutter interference. Background Technology

[0002] High-frequency ground-wave radar's wide-area monitoring capabilities can cover exclusive economic zones and disputed waters, acquiring real-time ocean dynamic parameters and providing data support for maritime traffic control and resource development. Especially in sensitive waters, its all-weather monitoring capabilities can effectively identify illegal fishing, oil and gas theft, and other activities, contributing to the enhancement of national maritime law enforcement capabilities. The technology's long-term continuous observation capabilities can provide scientific evidence for addressing climate change and protecting biodiversity; its environmental significance is increasingly prominent in the context of carbon neutrality and sustainable development. Simultaneously, this technology can provide sea surface current and wave information, significantly improving the timeliness and accuracy of disaster early warning.

[0003] High-frequency ground wave radar (HFSWR), as a core technology for marine environmental monitoring, can dynamically retrieve key ocean parameters such as sea surface wind speed, significant wave height, and ocean current field by receiving electromagnetic scattering echoes from the ocean surface. From 1966 to 1972, Wait and Barrick proposed the Bragg scattering principle and established first and second-order scattering equations, providing a solid theoretical foundation for subsequent sea state parameter retrieval. 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 gravity waves from the sea surface and radar electromagnetic waves, and its frequency shift, peak amplitude, and energy distribution in the spectral region are directly related to ocean dynamic parameters.

[0004] Currently, the maximum value method and the matched peak method are two commonly used techniques for detecting Bragg peaks. The basic principle of the matched peak method is that the frequency difference between positive and negative Bragg peaks is theoretically fixed at 2. This characteristic means that the difference between the positive and negative frequencies of the Bragg peak remains constant regardless of its shift. By determining the range of the Bragg peak, setting a detection threshold, searching for extreme points within the range, and then matching the statistically analyzed extreme points, a pair of Bragg peaks whose positive and negative frequency differences are closest to the theoretical difference is selected.

[0005] However, ionospheric clutter severely restricts the reliability and accuracy of this technology. Ionospheric clutter is generated by the interaction between electromagnetic waves and ionospheric irregularities. Its spectral characteristics highly overlap with the ocean's Bragg peak, causing ocean information in radar echo signals to be masked or distorted by clutter energy. This phenomenon directly undermines the physical basis of ocean parameter extraction, making ocean state inversion face systematic biases or even failure risks.

[0006] Meanwhile, even after ionospheric clutter suppression, some distance cells still have residual clutter contamination. Both of these methods can lead to false peak misjudgment and missed detection of valid peaks in the Bragg peak detection values, resulting in a high false peak misjudgment rate and a high valid peak missed detection rate. Such abnormal data will directly affect the accuracy of sea state parameter inversion.

[0007] To address the problems existing in the current technology, there is an urgent need for a method capable of high-precision detection of Bragg peaks, effective removal of spurious peaks, and joint estimation of Bragg resonance parameters, both in the presence of ionospheric clutter interference and after ionospheric clutter suppression. This method would repair data affected by clutter and residual clutter, improve the spatial continuity and inversion accuracy of sea state inversion data, ensure the reliability and accuracy of high-frequency ground wave radar in sea state monitoring, and provide better technical support for marine environmental monitoring and related applications. Summary of the Invention

[0008] This invention addresses the problem of high false peak misjudgment rate in existing sea state inversion methods for Bragg peak detection, which affects the accuracy of sea state inversion.

[0009] A method for selecting Bragg peaks for ocean state inversion includes:

[0010] The maximum value of the Bragg peak shift is obtained based on the maximum ocean current velocity in the area to be detected. The range of first-order peak occurrence was determined; the average noise amplitude was calculated based on all background Doppler elements of the high-frequency ground wave radar. The threshold for preliminary detection of Bragg peaks is determined, and 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, and a pair of Bragg peaks that are closest to the theoretical frequency difference between positive and negative Bragg peaks are selected to obtain candidate Bragg peaks.

[0011] Furthermore, the range of the first-order peak is as follows:

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

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

[0014] A method for determining the pseudo-peak of Bragg peak in ocean state inversion includes the following steps:

[0015] The positions of the candidate Bragg peaks obtained based on the aforementioned Bragg peak screening method for sea state inversion are as follows: The corresponding spectral amplitude is Establish with The detection window centered on ,in For frequency window width, The coefficients are positive integers. For Doppler resolution;

[0016] Candidate Bragg peaks outside the detection window are identified as Bragg spurious peaks.

[0017] In the detection window Within the range of statistics, Number of frequency points , To be based on the spectral amplitude A defined dynamic detection threshold; when the following conditions are met. When the current candidate peak is determined to be a Bragg spurious peak, The judgment threshold is set based on the statistical characteristics of ocean echoes.

[0018] Furthermore, based on the spectral amplitude Determined dynamic detection threshold , This is the amplitude margin parameter.

[0019] A joint estimation method for missing values ​​of Bragg peaks in sea state inversion is proposed. This method utilizes a pseudo-peak identification method for Bragg peaks in sea state inversion to identify and remove pseudo-peaks. Then, the LW-EM algorithm is used to fuse spatial correlation and statistical inference to achieve joint estimation of the Bragg peaks. The specific process includes the following steps:

[0020] Step S31, Initialize parameters:

[0021] Radar transmits pulsed electromagnetic waves and receives the echoes, dividing the detection range into sections based on the echo delay. There are 3 equally spaced distance segments, each of which is a distance unit; let the 1st distance segment be a distance unit. The amplitude observation value of each distance cell is , forming a sequence Define missing data as Missing data indicates spurious peaks or missed Bragg peaks;

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

[0023] The expectation step, or E-step: Calculate the local posterior expectation. ; For posterior weights, For spatial correlation window functions, This is the set of parameters to be estimated for the k-th iteration update; , Let represent the mean and variance at position j, respectively. Indicates the first The prior probability that the location is the true Bragg peak. Let be the probability density function of a normal distribution. Let be the impulse function;

[0024] Maximize steps Step: Update parameters; use adaptive sliding window estimation:

[0025]

[0026] , These are the missing value estimation results obtained from the k-th and (k+1)-th iterations;

[0027] Step S33: Repeat the iteration process of S32 until the iteration converges. Once the iteration converges, the final result is obtained. and Then, a joint estimation of the missing values ​​is performed, and the missing value estimates are as follows: , This is the result of missing value estimation.

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

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

[0030]

[0031] in, To set the convergence error, This represents the estimation result of missing values ​​in the k-th iteration.

[0032] A sea state inversion method is provided, which performs sea state inversion on Bragg peak parameters obtained based on the aforementioned method for joint estimation of missing values ​​of Bragg peaks for sea state inversion.

[0033] Furthermore, the radial velocity calculation results obtained from the sea state inversion are as follows: , For radar wavelength, This represents the difference between the theoretical Bragg peak frequency and the observed frequency.

[0034] The beneficial effects of this invention are:

[0035] 1. This invention significantly improves the detection accuracy of Bragg peaks in sea state 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. At the same time, it realizes the joint estimation of multiple parameters such as radial current velocity, wind direction, and wind speed, 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 joint estimation and filling of missing Bragg peaks (after missing detection or removal of spurious peaks) provides strong robustness to different sea conditions (such as changes in current velocity and wave height fluctuations), providing technical support for the all-weather, high-precision application of high-frequency ground wave radar in marine remote sensing. Attached Figure Description

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

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

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

[0041] Figure 4 The right-hand amplitude is obtained by performing Bragg pseudo-peak detection and estimating the data using the LW-EM method.

[0042] Figure 5 The left-hand frequency is obtained by performing Bragg pseudo-peak detection and estimating the data using the LW-EM method.

[0043] Figure 6 The right-hand frequency is obtained by performing Bragg pseudo-peak detection and estimating the data using the LW-EM method.

[0044] Figure 7 It is a 3D model for estimating the pre-Bragg peak.

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

[0046] Figure 9 It is a top view of the estimated 3D model of the pre-Bragg peak.

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

[0048] Figure 11 This is the result of radial velocity inversion. Detailed Implementation

[0049] To address the shortcomings of existing technologies in Bragg peak detection accuracy, joint parameter estimation, and robustness of sea state inversion, this invention proposes a method for joint estimation of Bragg resonance parameters and sea state inversion. By constructing a joint parameter optimization model and utilizing the physical correlation constraints between Bragg scattering parameters, iterative correction and missing peak reconstruction are performed on detected anomalous Bragg peaks. A multi-stage processing strategy is employed: first, a pseudo-peak identification mechanism based on spectral feature criteria is established; then, algorithms such as LW-EM are implemented to achieve joint optimal estimation of multiple parameters, ultimately improving the performance of sea state inversion. Specific implementation method one:

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

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

[0053] Step S1: The Bragg peak is detected using the matching peak method, and the candidate Bragg peak is determined by using the same offset characteristics of the Bragg peaks on both sides.

[0054] (1) Determine the range of the Bragg peak: Due to the presence of ocean surface currents, the Bragg peak relative to the theoretical frequency A certain degree of shift will occur, and this shift is related to the ocean current velocity, resulting in a certain detection range for the actual detected value of the Bragg peak. The maximum value of the Bragg peak shift can be obtained by calculating based on the maximum ocean current velocity in the area to be detected. The range of first-order peaks is defined as follows: a positive first-order Bragg peak is... The negative first-order Bragg peak is .

[0055] (2) Determine the threshold for Bragg peak detection: Background Doppler cells are frequency cells in the range-Doppler spectrum (RD spectrum) of high-frequency ground wave radar (HFSWR) that do not contain Bragg scattering peaks or other target echo signals. They mainly reflect the energy distribution of background signals such as environmental noise and incoherent clutter. The average noise amplitude is calculated by selecting all background Doppler cells. Based on this, a minimum signal-to-noise ratio is set as the threshold for the initial detection of the Bragg peak.

[0056] (3) Extreme points within the search range: By detecting the threshold of the Bragg peak, the extreme points within the range of the Bragg peak 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 at 1. Regardless of the shift in the Bragg peak, its positive and negative difference remains constant. Therefore, the extreme points statistically analyzed in the previous steps are matched, and a pair of Bragg peaks whose positive and negative frequency differences are closest to the theoretical difference are selected. If they meet reasonable error conditions, they are determined to be candidate Bragg peaks; otherwise, the Bragg peak at that location is determined to be undetectable.

[0058] Step S2: Establish a pseudo-peak identification mechanism based on spectral feature criteria to screen and remove abnormal detection peaks.

[0059] Bragg spurious peaks are detected based on the Doppler characteristics of the Bragg peak at a fixed distance cell. Each Bragg peak is judged based on the distribution characteristics of the peak neighborhood energy, and spurious Bragg peaks caused by clutter interference or noise are removed. The specific process includes the following steps:

[0060] Step S21, establish the central detection window:

[0061] Let the position of the candidate Bragg peak obtained in the initial detection be . The corresponding spectral amplitude is Establish with The detection window centered on ,in For frequency window width, The coefficients are positive integers. For Doppler resolution. Define the dynamic detection threshold as:

[0062]

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

[0064] Step S22: Candidate Bragg peaks outside the detection window are identified as Bragg pseudo-peaks and are removed.

[0065] In the detection window Within the range of statistics, Number of frequency points .

[0066]

[0067] In the formula, This is an indicator function.

[0068] A threshold is set based on the statistical characteristics of ocean echoes. When satisfied When a candidate peak is identified as a Bragg spurious peak, it is removed. The algorithm allows for setting different threshold values ​​for 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 Bragg spurious peaks, and provide a basis for 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 any missing data that may arise after removing spurious peaks in S2, the LW-EM algorithm is used to fuse spatial correlation and statistical inference to achieve joint estimation of the Bragg peaks. Simultaneously, the physical correlation between Bragg scattering parameters is utilized to iteratively correct the residual outliers after S2 processing, ensuring that the parameters conform to ocean dynamics. The specific process includes the following steps:

[0071] Step S31, Initialize parameters:

[0072] Radar transmits pulsed electromagnetic waves and receives the echoes, dividing the detection range into sections based on the echo delay. The distance is divided into several equally spaced segments, each of which is a distance unit. Let the first segment be a distance unit. The amplitude observation value of each distance cell is , forming a sequence Define missing data as Missing data may be spurious peaks or undetected Bragg peaks. A spatial correlation window function is introduced. ,in Let be the window width parameter, and j be the position order. Introduce a local Gaussian mixture model:

[0073]

[0074] In the formula For the set of parameters to be estimated, , Let represent the mean and variance at position j, respectively. Indicates the first The prior probability that the location is the true Bragg peak. Let be the probability density function of a normal distribution. Let be the impulse function.

[0075] Step S32, calculate the expected step (E step) and the maximization step (M step):

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

[0077]

[0078] The posterior weights are:

[0079]

[0080] in This is the updated set of parameters to be estimated.

[0081] Maximize steps ( Step 1: Update the parameters;

[0082] Adaptive sliding window estimation is used:

[0083]

[0084] Step S33: Repeat the iteration process of S32 until the iteration converges. Once the iteration converges, the final result is obtained. and Then, a joint estimation of the missing values ​​is performed, and the missing value estimates are as follows:

[0085]

[0086] in, The results of missing value estimation. and This is the result of the iteration.

[0087] This improved method captures the distance correlation of sea state parameters through an exponentially decaying window, while using cubic spline interpolation for initial value filling to enhance the model's robustness. Furthermore, it employs dynamic convergence and utilizes a stopping criterion based on relative mean square error (RMSE).

[0088]

[0089] in, To set the convergence error, This represents the estimation result of missing values ​​in the k-th iteration.

[0090] Step S4, perform sea state inversion:

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

[0092] The radial flow velocity calculation results are as follows

[0093]

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

[0095] This invention addresses the issue of missed and misjudged Bragg peaks in some distance cells of ocean state inversion data. It employs a Bragg pseudo-peak detection mechanism based on spectral feature criteria and the LW-EM algorithm to jointly estimate Bragg resonance parameters, thereby restoring Bragg peak contamination data and ensuring the accuracy of ocean state parameter inversion.

[0096] Example:

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

[0098] Image fusion clutter suppression processing based on an optimization algorithm was implemented for cells with a range of 100-150km. The processing results are as follows: Figure 2 As shown in the figure, experiments demonstrate that ionospheric clutter energy is effectively suppressed in this region, and the previously obscured Bragg scattering peak is fully revealed. However, in the vicinity of 150 km, the intensity of the Bragg peak signal decreases after suppression due to its weak amplitude on the left side of the original signal. Therefore, targeted parameter extraction using a Bragg peak detection algorithm is necessary to accurately assess the local ocean echo characteristics.

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

[0100] The red circles represent the original Bragg peak detection results, and the estimated data is represented by the blue broken line. The absence of original data in some distance cells indicates that the Bragg peak was not detected by the matching peak method or that the Bragg spurious peak was removed by the Bragg spurious peak detection method.

[0101] The left-side Bragg peak exhibits a large range of zero values ​​in the 50-70 distance cell range. After estimation, its frequency conforms to the continuity of a Bragg peak, changing slowly without abrupt changes. Its amplitude also conforms to the correlation between the Bragg peak amplitude and distance cells, again without abrupt changes. The estimated amplitude range is between -60 and -65, consistent with the normal amplitude range of a Bragg peak. The results for the left-side Bragg peak are superior. The right-side Bragg peak shows small gaps near distance cells 40, 52, and 63. After estimation, its frequency changes also follow the pattern of Bragg peak frequency changes, and the estimated amplitude changes show some distance correlation, compensating for the partial Bragg peak gaps caused by ionospheric clutter suppression.

[0102] The 3D model of the Bragg peak before and after is estimated as follows: Figure 7 and Figure 8 As shown, its top-down view is as follows Figure 9 and Figure 10 As shown, it can be seen that after joint parameter estimation, the frequency and amplitude characteristics of the Bragg peaks on both the left and right sides conform to the characteristics of ocean echoes. The continuity of the Bragg peaks is improved before and after estimation, and the results are supported by other data and have the characteristics of the original data. The data truncation of the left Bragg peak at the 40-70 distance cell is perfectly repaired, and the fragmentation of the right Bragg peak is also repaired.

[0103] Radial velocity inversion results are as follows Figure 11 As shown, the solid line represents the processed radial velocity inversion result, and the dashed line represents the radial velocity inversion result from the Bragg peak extracted from the original RD spectrum. It can be clearly seen that, compared to directly detecting the Bragg peak, the radial velocity inversion result from the processed data meets the requirements for radial velocity in conventional sea areas, and its variation characteristics conform to ocean dynamics, without velocity jumps. The magnitude and direction of the velocity in each distance cell exhibit continuity and correlation.

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

[0105] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

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

Claims

1. A joint estimation method for missing values ​​of Bragg peaks in sea state inversion, characterized in that, A method for identifying and removing spurious peaks in ocean state inversion, including: Bragg peak selection: The maximum Bragg peak offset is obtained based on the maximum ocean current velocity in the area to be detected. The range of first-order peak occurrence was determined; the average noise amplitude was calculated based on all background Doppler elements of the high-frequency ground wave radar. The threshold for preliminary detection of Bragg peaks is determined, and extreme points within the range of Bragg peaks are statistically recorded using this threshold to search for extreme points. These extreme points are then matched to select a pair of Bragg peaks whose theoretical frequency difference is closest to the theoretical difference between positive and negative Bragg peaks, thus obtaining candidate Bragg peaks. The positions of the obtained candidate Bragg peaks are... The corresponding spectral amplitude is ; Establish with The detection window centered on ,in For frequency window width, The coefficients are positive integers. To achieve Doppler resolution; candidate Bragg peaks outside the detection window are identified as Bragg spurious peaks; within the detection window... Within the range of statistics, Frequency points , To be based on the spectral amplitude A defined dynamic detection threshold; when the following conditions are met. When the current candidate peak is determined to be a Bragg spurious peak, The judgment threshold is set based on the statistical characteristics of ocean echoes; Then, the LW-EM algorithm is used to fuse spatial correlation and statistical inference to achieve joint estimation of the Bragg peak; the specific process includes the following steps: Step S31, Initialize parameters: The radar transmits pulsed electromagnetic waves and receives echoes, dividing the detection range into sections based on the echo delay. There are 3 equally spaced distance segments, each of which is a distance unit; let the 1st distance segment be a distance unit. The amplitude observation value of each distance cell is , forming a sequence Define missing data as Missing data indicates spurious peaks or missed Bragg peaks; Step S32, calculate the expected step and the maximization step: The expectation step, or E-step: Calculate the local posterior expectation. ; For posterior weights, For spatial correlation window functions, This is the set of parameters to be estimated for the k-th iteration update; , Let represent the mean and variance at position j, respectively. Indicates the first The prior probability that the location is the true Bragg peak. Let be the probability density function of a normal distribution. Let be the impulse function; Maximize steps Step: Update parameters; use adaptive sliding window estimation: ， ， , These are the missing value estimation results obtained from the k-th and (k+1)-th iterations; Step S33: Repeat the iteration process of S32 until the iteration converges. Once the iteration converges, the final result is obtained. and Then, a joint estimation of the missing values ​​is performed, and the missing value estimates are as follows: , This is the result of missing value estimation.

2. The method for joint estimation of missing values ​​of Bragg peaks in sea state inversion according to claim 1, characterized in that, The range of first-order peaks is as follows: The first-order Bragg peak is The negative first-order Bragg peak is ; This is the theoretical frequency.

3. A method for joint estimation of missing values ​​of Bragg peaks in sea state inversion according to claim 1 or 2, characterized in that, The theoretical frequency difference between the positive and negative Bragg peaks is 100%. , This is the theoretical frequency.

4. The method for joint estimation of missing values ​​of Bragg peaks in sea state inversion according to claim 1, characterized in that, According to the spectral amplitude Determined dynamic detection threshold , This is the amplitude margin parameter.

5. The method for joint estimation of missing values ​​of Bragg peaks in sea state inversion according to claim 1, characterized in that, The spatial correlation window function ,in is the window width parameter, and j is the position order.

6. The method for joint estimation of missing values ​​of Bragg peaks in sea state inversion according to claim 1, characterized in that, The convergence condition for step S33 is as follows: in, To set the convergence error, This represents the estimation result of missing values ​​in the k-th iteration.

7. A method for ocean state inversion, characterized in that, The sea state inversion is performed on the Bragg peak parameters obtained based on the missing value joint estimation method for Bragg peaks for sea state inversion as described in claim 1.

8. The sea state inversion method according to claim 7, characterized in that, The radial velocity calculation result obtained from the sea state inversion is: , For radar wavelength, This represents the difference between the theoretical Bragg peak frequency and the observed frequency.

Images