A Vector Flow Inversion Method Combining Single-Station Ground Wave Radar and Spaceborne SAR

Abstract

This invention relates to the field of ocean current remote sensing, and specifically to a method for vector flow inversion combining single-station ground-wave radar (GWR) and spaceborne SAR. The method mainly includes: calculating the frequency shift of the first-order Bragg peak of the GWR; inverting the radial current of the sea surface from the GWR; inverting the radial current of the sea surface from the SAR using the Doppler centroid anomaly method; interpolating the radial current results from the GWR to the SAR data grid using bilinear interpolation; inverting the single-station vector flow based on the radial current results from the GWR; projecting the single-station vector flow data with high confidence onto the SAR range direction; calculating the mean difference between the projected current velocity and the SAR radial current velocity; and then calibrating the SAR radial flow; finally, inverting the vector flow by combining the radial current results from the GWR and the calibrated SAR radial flow. This method overcomes the limitations of single-method ocean current observation, achieving high-precision inversion of combined vector flow, and is of great significance for marine environmental monitoring, disaster forecasting, and shipping safety.

Description

Technical Field

[0001] This invention relates to the field of ocean current remote sensing, and specifically to a method for inverting sea surface vector currents by combining high-frequency ground wave radar and spaceborne synthetic aperture radar (SAR). Background Technology

[0002] Accurate monitoring of ocean currents is of great significance for maritime navigation safety, disaster early warning, and marine engineering operation and maintenance. High-frequency ground wave radar and spaceborne synthetic aperture radar (SAR) are two major remote sensing methods capable of simultaneously achieving large-scale, long-distance marine environmental monitoring.

[0003] Ground-wave radar utilizes the diffraction and propagation characteristics of high-frequency electromagnetic waves on the sea surface to achieve continuous, over-the-horizon observation of the sea surface and real-time inversion of ocean current information; however, its spatial resolution is relatively low, at the kilometer level, making it difficult to obtain fine-grained parameters. Spaceborne SAR, with its high spatial resolution at the meter or sub-meter level, can acquire detailed two-dimensional images of the sea surface, thus enabling the acquisition of high spatial resolution ocean current information; however, SAR has low temporal resolution, making it difficult to achieve continuous observation of areas of interest even through satellite networking, failing to meet emergency observation needs. Therefore, both ground-wave radar and SAR possess global observation capabilities in space, enabling large-scale spatiotemporal matching in nearshore or offshore areas, and demonstrating the feasibility of collaborative detection.

[0004] Ground-wave radar (GWR) uses Doppler shift information in the first-order echo spectrum to calculate radial current velocity. Vector current acquisition is mostly done using bistatic or multistatic radar; however, single-station radar is sometimes used. In such cases, methods such as least squares can be used to invert the vector current, but this sacrifices some spatial or angular resolution, resulting in low accuracy. SAR primarily uses two methods to invert ocean currents: in-orbit interferometry (ATI) and Doppler centroid anomaly (DCA). The ATI method inverts radial current velocity through the phase difference of dual-antenna images; its applicability in practical observations is limited by antenna configuration. The DCA method uses the relationship between sea surface motion and Doppler shift in SAR echoes to invert radial current velocity, suitable for large and medium-scale flow fields; however, Doppler shifts caused by wind and waves can also affect the inverted radial current. Therefore, GWR and SAR can acquire radial flow field results from different lines of sight. Consider using GWR vector current results to optimize SAR radial current, thereby obtaining higher-precision vector current results when both are used in conjunction to observe ocean currents.

[0005] In summary, the collaborative observation of ground wave radar and SAR can obtain higher precision vector flow field results. Based on this, this application will integrate the advantages of ground wave radar and SAR in observing sea conditions and develop a vector flow inversion method that combines single-station ground wave radar and spaceborne SAR to achieve high-precision and high-resolution inversion of joint vector flow. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] This invention aims to provide a vector flow inversion method that combines single-station ground wave radar and spaceborne SAR to achieve large-scale, high-precision ocean current vector observation, thereby improving the accuracy of ocean current field monitoring and meeting the urgent need for reliable current field data for safe navigation, disaster early warning, and engineering operation and maintenance at sea.

[0008] (II) Technical Solution

[0009] This invention includes the following steps:

[0010] Step 1, Radial flow inversion from ground wave radar and radial flow inversion from SAR:

[0011] Step 1.1: Perform radial flow inversion from ground wave radar. After acquiring the range-Doppler echo spectrum data from the ground wave radar, directly calculate the frequency shift of the first-order Bragg peak. The formula is: In the formula, Indicates the radial velocity of the ocean current. This represents the difference between the actual first-order Bragg peak frequency and the theoretical frequency. This indicates the wavelength of the electromagnetic waves emitted by the radar.

[0012] Step 1.2: Achieve SAR radial flow inversion. Utilize the Doppler centroid anomaly method, i.e., invert sea surface current velocity by analyzing the relationship between sea surface motion and the Doppler frequency shift in the SAR echo. The Doppler frequency shift in the SAR echo... It can be represented as: ,in, The frequency shift caused by the relative motion between the SAR platform and the Earth's surface, i.e., the predicted Doppler, can be calculated using satellite orbit and attitude data. The frequency shift caused by sea surface motion is the result of the combined effects of ocean currents, wind fields, and ocean waves. This indicates the frequency shift of electromagnetic pointing error caused by the satellite platform or antenna attitude being affected by seasonal atmospheric effects and solar radiation, resulting in the antenna pointing deviating from the theoretical direction. This represents the frequency shift caused by inaccurate estimations and unknown biases of non-geophysical terms. It is removed by utilizing the Earth Empirical Physical Model (CDOP). Doppler shift in wind and waves The remaining Doppler frequency shift is generated by ocean current motion.

[0013] The CDOP model function is: ,in, These represent the wind speed and direction at a depth of 10 meters above the sea surface, respectively. Indicates the incident angle of the SAR. This represents the polarization mode of the SAR. Therefore, the inverted radial velocity is: , ,in, Radial flow velocity, For SAR radar electromagnetic wave number, This is the Doppler shift caused by ocean currents.

[0014] Step 2: Spatial matching of ground wave radar and SAR: Spatial registration is performed between the synchronized ground wave radar radial current results and the SAR area. Since the latitude and longitude grids of ground wave radar and SAR data observed in the same sea area are not the same, the radial current results from the ground wave radar need to be matched into the SAR coordinate grid. Here, bilinear interpolation is used between the ground wave radar latitude and longitude matrix and the radial current results to interpolate the radial current results into the uniform latitude and longitude grid of the SAR, ensuring that both methods maintain the same detection range and latitude and longitude grid. The formula for bilinear interpolation is: In the formula, Represents the SAR latitude and longitude points to be matched. , , , Representing four ground-wave radar latitude and longitude points surrounding the SAR latitude and longitude point, the radial flow values ​​at the four points are as follows: , , , .

[0015] Step 3: Constrain and calibrate the SAR radial flow using the vector flow results from a single-station ground wave radar.

[0016] First, the vector flow is inverted using the least squares method.

[0017] Secondly, to ensure the reliability of the vector flow inversion results from a single-station ground wave radar, a quality assessment is necessary. This mainly involves selecting data points with high confidence from the entire flow field obtained in the initial inversion and discarding unreliable points. ,in, For radial velocity residual, For the eastward velocity component of the vector flow, For the northward velocity component of the vector flow, This is the azimuth angle of the radial flow of the ground wave radar. The value is 1 if the condition is met, and 0 otherwise. This represents the standard deviation of the flow direction within a 3x3 grid window. Finally, grid points with a reliability greater than 0.9 are selected, and high-confidence vector flow fields and their spatial locations are output, effectively extracting the reliable portion of the inversion results.

[0018] Finally, the radial velocity of the SAR is calibrated using the mean difference. The high-confidence vector flow is projected onto the SAR range direction to obtain the projected radial component of the ground-wave radar. Then, the mean difference between the SAR and ground-wave radar projected radial components is calculated, and this mean is subtracted from the original SAR velocity as a correction factor to obtain the calibrated velocity. ,in, This indicates the calibrated flow rate. This represents the initial radial velocity obtained using the DCA method. Represents a high-confidence vector flow. Indicates the number of valid data points. This represents the angle between the ground wave radar vector flow and the SAR range direction.

[0019] Step 4: Perform vector flow inversion by combining the radial flow from the single-station ground wave radar and the calibrated SAR radial flow. By unifying the radial flow from the ground wave radar and the calibrated SAR radial flow under the stream function framework, the velocity components are solved using Taylor series expansion, and finally the vector flow field is constructed.

[0020] Stream function It is a function describing the movement of ocean currents, and can be defined by the following formula: , For the ground wave radar radial flow matched in the overlapping region and the calibrated SAR radial flow, , The relationship between the radial velocity and any observation point can be expressed, and the stream function can be expanded using Taylor series: , ,in, for powers of, for powers of, It corresponds The coefficient of the term, Let be the highest order of the Taylor series, set to 3. This includes ground-wave radar and calibrated SAR radial flow at the same observation point. A linear equation can be established for each observation point, and the overdetermined linear equations are finally solved using the least squares method to obtain the stream function coefficients, allowing for the calculation of... and Components, calculate the vector flow: , ,in, The vector flow velocity obtained from the final joint inversion, This indicates the direction of the vector flow.

[0021] (III) Beneficial Effects

[0022] The advantages of this invention are as follows:

[0023] This invention innovatively combines the advantages of both ground-wave radar and SAR remote sensing methods for observing the sea surface. By spatiotemporally registering the radial flow results from ground-wave radar with SAR data, radial flow results from ground-wave radar that match the SAR data for the same region are obtained. Then, the radial flow retrieved from SAR is constrained and calibrated using vector flow results from a high-confidence single-station ground-wave radar to improve the accuracy of the SAR radial flow. Finally, the vector flow is retrieved by combining the radial flow from ground-wave radar with the calibrated SAR radial flow using a stream function. Attached Figure Description

[0024] Figure 1 The flowchart of the vector flow inversion method for the combined single-station ground wave radar and spaceborne SAR provided by the present invention is shown.

[0025] Figure 2 The high-confidence vector flow results of single-station ground wave radar provided by this invention.

[0026] Figure 3 The absolute error results between the calibrated SAR radial flow and actual ocean current data provided by this invention.

[0027] Figure 4 The absolute error results between the jointly inverted vector current velocity and the actual ocean current data provided by this invention.

[0028] Figure 5 The absolute error results between the vector current direction obtained by the joint inversion provided by this invention and the actual ocean current data. Detailed Implementation

[0029] To make the objectives, contents, and advantages of the present invention clearer, the following description is provided in conjunction with the appendix. Figure 1 The specific embodiments of the present invention will be described in further detail below:

[0030] Taking a set of synchronized ground-wave radar and SAR observation data as an example, the selected area is the west coast of the United States, and the time is 23:30 (UTC) on May 5, 2019. The ground-wave radar data comes from the radial current product of the WERA radar deployed in this sea area, operating at a frequency of 8.328 MHz; the SAR data comes from the Sentinel-1A satellite, is of OCN data type, and uses the interferometric wide swath mode with a spatial resolution of 1 km × 1 km. The time difference between the SAR imaging time and the ground-wave radar data is less than 30 minutes.

[0031] Step 1, inversion of radial flow from ground wave radar and SAR radial flow:

[0032] Step 1.1: Perform radial flow inversion from ground wave radar.

[0033] Step 1.2, perform SAR radial flow inversion. Doppler frequency shift in SAR echo. It can be represented as: ,in, The frequency shift caused by the relative motion between the SAR platform and the Earth's surface, i.e., the predicted Doppler, can be calculated using satellite orbit and attitude data. The frequency shift caused by sea surface motion is the result of the combined effects of ocean currents, wind fields, and ocean waves. This indicates the frequency shift of electromagnetic pointing error caused by the satellite platform or antenna attitude being affected by seasonal atmospheric effects and solar radiation, resulting in the antenna pointing deviating from the theoretical direction. This indicates the frequency shift caused by inaccurate estimations and unknown biases of non-geophysical terms. Since the OCN product data from Sentinel-1 SAR directly provides the Doppler frequency shift... And predicting Doppler Therefore, no additional calculations are required; while the electromagnetic pointing error frequency shift and unknown deviation Land correction is needed to remove the frequency shift, which involves calculating the average Doppler frequency value of the land area in the SAR image and uniformly removing it from the entire scene data. After removing the frequency shift, the Doppler frequency of the sea surface is calculated. At this point, it is necessary to use the Earth Empirical Physical Model (CDOP) to remove the Doppler shift of wind and waves. The remaining Doppler frequency shift is generated by ocean current motion.

[0034] The CDOP model function is: .in, These represent the wind speed and direction at a depth of 10 meters above the sea surface, respectively. Indicates the incident angle of the SAR. This indicates the polarization mode of the SAR. Therefore, the inverted radial velocity is: , .in, Radial flow velocity, For SAR radar electromagnetic wave number, This is the Doppler shift caused by ocean currents.

[0035] Step 2 involves spatially registering the synchronized ground wave radar radial current results with the SAR data. Since the latitude and longitude grids of the ground wave radar and SAR data observed in the same sea area are not identical, the ground wave radar radial current results need to be matched into the SAR coordinate grid after acquisition. Subsequently, bilinear interpolation is performed with the ground wave radar latitude and longitude matrix and the radial current results to interpolate the radial current results into the uniform latitude and longitude grid of the SAR, ensuring that both methods maintain the same detection range and latitude and longitude grid. The formula for bilinear interpolation is: In the formula, Represents the SAR latitude and longitude points to be matched. , , , Representing four ground-wave radar latitude and longitude points surrounding the SAR latitude and longitude point, the radial flow values ​​at the four points are as follows: , , , .

[0036] Step 3: Constrain and calibrate the SAR radial flow using the vector flow results from a single-station ground wave radar.

[0037] First, the least squares method is used to invert the vector flow of a single-station ground wave radar. For each matched radial flow observation point of the ground wave radar, neighborhood data of a certain window size are selected around it. Based on the fact that the radial flow velocity is the projection of the ocean current vector onto the radar beam direction, the corresponding equation is established: .in, The radial velocity of the ground wave radar. The azimuth angle of the radial flow of the ground wave radar. For vector flow, The direction of the vector flow is then determined. Next, the linear equations within this window are solved using least squares. By traversing all valid radial flow observation data through a sliding window, the initial vector flow field of the single-station ground wave radar area can be obtained.

[0038] Secondly, to ensure the reliability of the vector flow inversion results from a single-station ground wave radar, a quality assessment is required: ,in, For radial velocity residual, For the eastward velocity component of the vector flow, This represents the northward velocity component of the vector flow. The value is 1 if the condition is met, and 0 otherwise. This represents the standard deviation of the flow direction within a 3x3 grid window. Finally, grid points with a reliability greater than 0.9 are selected, and high-confidence vector flow fields and their spatial locations are output, effectively extracting the reliable portion of the inversion results. The high-confidence vector flow results from a single-station ground wave radar are shown below. Figure 2 As shown.

[0039] Finally, the radial velocity of the SAR is calibrated using the mean difference. The radial component of the ground-wave radar projection is obtained by projecting the high-confidence vector flow from the ground-wave radar onto the SAR range direction. Then, the mean difference between the SAR and ground-wave radar radial components is calculated, and this mean is subtracted from the original SAR velocity as a correction factor to obtain the calibrated velocity. ,in, This indicates the calibrated flow rate. This represents the initial radial velocity obtained using the DCA method. Represents a high-confidence vector flow. Indicates the number of valid data points. This represents the angle between the ground-wave radar vector current and the SAR range direction. The absolute error results between the calibrated SAR radial current and actual ocean current data are as follows: Figure 3 As shown.

[0040] Step 4: Perform vector flow inversion by combining the radial flow from the single-station ground wave radar and the radial flow from the calibrated SAR. By unifying the radial flow from the ground wave radar and the calibrated SAR radial flow under the stream function framework, the velocity components are solved using Taylor series expansion, and finally, the vector flow is synthesized.

[0041] Stream function It is a function describing the movement of ocean currents, and can be defined by the following formula: , For the ground wave radar radial flow matched in the overlapping region and the calibrated SAR radial flow, , The relationship between the radial velocity and any observation point can be expressed, and the stream function can be expanded using Taylor series: , ,in, for powers of, for powers of, It corresponds The coefficient of the term, Let be the highest order of the Taylor series, set to 3. This includes ground-wave radar and calibrated SAR radial flow at the same observation point. A linear equation can be established for each observation point, and the overdetermined linear equations are finally solved using the least squares method to obtain the stream function coefficients, allowing for the calculation of... and Components, calculate the vector flow: , ,in, The vector flow velocity obtained from the final joint inversion, This indicates the direction of the vector flow.

[0042] The absolute error results of the joint inversion of vector current velocity and actual ocean current data are as follows: Figure 4 As shown, the root mean square error of the current velocity is 0.13 m / s. The absolute error between the jointly inverted vector current direction and the actual ocean current data is shown in the figure. Figure 5 As shown, the root mean square error of the flow direction is 19.37°.

[0043] In summary, the vector currents retrieved by the proposed method have high accuracy in sea state observation.

Claims

1. A method for vector flow inversion combining ground-based single station ground wave radar and space-borne SAR, characterized in that, The method comprises the following steps: (1) Ground wave radar radial flow inversion and SAR radial flow inversion: for ground wave radar, the frequency shift of the first-order Bragg peak is directly calculated to invert the sea surface radial flow; for SAR, the Doppler centroid anomaly method is used, i.e., the relationship between the sea surface movement and the Doppler shift in the SAR echo is analyzed to invert the SAR radial flow velocity; (2) Spatial registration: the synchronous ground wave radar radial flow result is spatially registered with the SAR data, the ground wave radar radial flow result is interpolated into the SAR latitude and longitude grid by using the bilinear interpolation to obtain the ground wave radar radial flow result under the SAR latitude and longitude grid; (3) SAR radial flow calibration: according to the fact that the radial flow velocity of the ground wave radar is the projection of the sea current vector in the direction of the radar beam, an equation between the ground wave radar radial flow and the vector flow is established, the initial single-station ground wave radar vector flow is inverted by solving the equation set by using the least square method; quality evaluation is needed, the data points with high confidence of the single-station ground wave radar vector flow are selected, and the unreliable points are removed; the projection radial component of the ground wave radar is obtained by projecting the data with high confidence in the single-station ground wave radar vector flow inversion result to the SAR range direction, the mean value of the difference between the projection radial component of the ground wave radar and the SAR radial flow velocity obtained in step (1) is calculated, and the mean value is subtracted from the SAR radial flow velocity as a correction to obtain the calibrated SAR flow velocity; (4) Joint vector flow inversion: the matched ground wave radar radial flow in step (2) and the calibrated SAR radial flow obtained in step (3) are uniformly constrained into the flow function framework, the velocity components are solved by using the Taylor series expansion, and finally the vector flow field is synthesized.

2. The method according to claim 1, wherein, The ground wave radar radial current inversion in step (1) and the SAR radial current inversion: sea surface radial current inversion of ground wave radar The calculation formula is: , wherein, represents the difference between the actual first-order Bragg peak frequency and the theoretical frequency, represents the wavelength of the electromagnetic wave emitted by the radar; in the method of SAR using Doppler centroid anomaly to invert radial current, the Doppler frequency shift in the SAR echo can be expressed as: , wherein, represents the frequency shift caused by the relative motion between the SAR platform and the earth's surface, i.e. the predicted Doppler, which can be calculated by satellite orbit and attitude data, represents the Doppler frequency shift caused by the sea surface motion, which is the result of the combined action of sea current, wind field and sea surface motion of sea wave, represents the electromagnetic pointing error frequency shift caused by the deviation of the antenna pointing direction from the theoretical direction due to the influence of seasonal atmospheric effect and solar radiation on the satellite platform or antenna attitude, represents the frequency shift caused by inaccurate estimation and unknown deviation of non-geophysical items; since the OCN product data of Sentinel-1 SAR directly provides the Doppler frequency shift and the predicted Doppler , additional calculation is not required; while the electromagnetic pointing error frequency shift and the frequency shift caused by inaccurate estimation and unknown deviation of non-geophysical items need to be removed by land calibration; after removing the above frequency shifts, the Doppler frequency shift caused by the sea surface motion is calculated , at this time, the wind wave Doppler frequency shift needs to be removed by using the empirical geophysical model CDOP , and the remaining Doppler frequency shift is the sea current motion, and the CDOP model function is: , wherein, respectively represent the wind speed and wind direction at the sea surface 10 meters, represents the incidence angle of SAR, represents the polarization mode of SAR; the inverted radial current velocity is: , , wherein, is the radial current velocity, is the wave number of SAR radar electromagnetic wave, is the Doppler frequency shift caused by the sea current motion.

3. The method according to claim 1, wherein, The spatial registration in step (2): for the SAR longitude and latitude points to be matched The four ground wave radar range-beam coordinate system longitude and latitude points around it are respectively , , , The corresponding radial flow values are respectively , , , The formula for bilinear interpolation to calculate the radial flow value is: The final result of the ground wave radar radial flow under the SAR latitude and longitude grid is obtained.

4. The method according to claim 1, wherein, The SAR radial flow calibration in step (3) is as follows: The formula of the ground wave radar radial flow velocity is: wherein, is the ground wave radar radial flow velocity, is the azimuth angle of the ground wave radar radial flow, is the vector flow, is the flow direction of the vector flow; and the formula of the single-station vector flow result of the screening confidence is: wherein, is the radial velocity residual, is the eastward flow velocity component of the vector flow, is the northward flow velocity component of the vector flow, represents 1 when the condition is met, otherwise 0; represents the flow direction standard deviation in a 3*3 grid window, is the confidence result; finally, the grid points with a confidence greater than 0.9 are screened, and the reliable part in the inversion result is effectively extracted; the formula of the calibrated SAR radial flow velocity is: wherein, represents the calibrated flow velocity, represents the initial radial flow velocity obtained by using the DCA method, represents the part of the high-confidence single-station ground wave radar vector flow, represents the number of effective data points, represents the angle between the ground wave radar vector flow and the SAR range direction.

5. The method according to claim 1, wherein, The joint vector flow inversion in step (4): stream function is defined as , , for the matched ground wave radar radial flow and calibrated SAR radial flow in the overlapping area, and the radial velocity relationship of any observation point can be expressed, and the stream function is expanded by Taylor series: , wherein, is the power of , is the power of , is the coefficient corresponding to , is the highest order of Taylor series, which is set to 3; contains the ground wave radar and SAR radial flow of the same observation point, and a linear equation can be established for each observation point, and finally the over-determined linear equation group is solved by the least square method to obtain the stream function coefficient, that is, the and components can be calculated, and the vector flow is obtained: , wherein, is the vector flow velocity obtained by joint inversion, is the vector flow direction obtained by joint inversion.

Images