SAR shallow water depth inversion method and system based on local wave direction

Abstract

The present invention discloses a SAR shallow sea water depth inversion method and system based on local wave direction, which belongs to the field of synthetic aperture radar signal processing and the field of marine remote sensing application technology. The method comprises obtaining a SAR image with clear sea wave characteristics in the study area; dividing the SAR image into sub-image blocks, estimating the local wave direction of each sub-image block based on the image spectrum; calculating the reference window angle of each sub-image block based on the local wave direction, and setting the window angle range and step size of the adaptive multi-angle window based on the reference window angle; calculating each image spectrum of the image sequence divided under different window angles for each sub-image block, estimating the sea wave wavelength corresponding to each image spectrum, and performing wavelength fusion based on Kalman filtering; applying the linear dispersion relationship to the fused wavelength to invert the underwater terrain. The present invention improves the accuracy and robustness of underwater terrain detection, and provides support for the engineering application of SAR shallow sea underwater terrain detection.

Description

Technical Field

[0001] The present invention belongs to the field of synthetic aperture radar (SAR) signal processing and marine remote sensing application, and in particular relates to a SAR shallow sea water depth inversion method and system based on local wave direction. Background Art

[0002] Shallow underwater topography has important guiding significance for human development activities such as coastal engineering construction and marine resource development. It is one of the basic contents of marine science research. As an important element in studying the dynamic balance between seawater and topography, it is particularly critical to the study of topographic dynamics.

[0003] Synthetic Aperture Radar (SAR) can instantly image a large area of ​​sea in a short period of time. The high-resolution data obtained contains rich shallow sea underwater topography information, which is used to realize dynamic monitoring of underwater topography changes. The advantage of SAR shallow sea underwater topography inversion is that it can not only reduce the cost of underwater topography detection, but also obtain first-hand water depth information of shallow sea areas and sensitive sea areas that ships cannot enter, and it is not affected by water quality and performs well in coastal waters with a limited water depth of less than 100m.

[0004] However, the inversion of underwater terrain in the existing technology still has two problems: high error and poor robustness. Studies have shown that the spectral peak position is limited when estimating the wavelength of waves using sub-images divided under a single window angle. When the spectral peak is not sharp, the wavelength estimation is inaccurate, resulting in high underwater terrain inversion errors. On the other hand, the wavelength inversion results are easily affected by noise and other marine phenomena, the inversion results are unstable, the anti-interference ability is poor, and the water depth inversion results are unreliable.

[0005] Therefore, in order to improve the accuracy and robustness of SAR shallow sea underwater terrain detection, the existing methods need to be improved. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a SAR shallow sea water depth inversion method and system based on local wave direction, which makes full use of the differences in estimating the wavelength of the waves at different divided window angles to improve the accuracy and stability of the estimation of the wavelength of the waves. Since there are differences in the sub-image blocks obtained under different window angles, when the window angle is consistent with the local wave direction of the waves, the spectrum peak energy is more concentrated, and the wavelength can be estimated more accurately, and the water depth can be inverted. The window angle is used as the reference window angle. Based on the reference window angle, the water depth is inverted by adaptively fusing multiple window angles, which solves the problem of high error and poor robustness of SAR shallow sea underwater terrain inversion in the prior art.

[0007] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0008] The SAR shallow sea depth inversion method based on local wave direction includes the following steps:

[0009] Step 1: Obtain a SAR image with clear wave characteristics in the study area;

[0010] Step 2: Divide the SAR image into sub-image blocks, and estimate the local wave direction of each sub-image block based on the image spectrum, where M and N are both natural numbers greater than 1;

[0011] Step 3: Calculate the reference window angle of each sub-image block based on the local wave direction, and set the window angle range and step size of the adaptive multi-angle window based on the reference window angle;

[0012] Step 4: Calculate the image spectra of the divided image sequence under different window angles for each sub-image block, estimate the wave wavelength corresponding to each image spectrum, and perform wavelength fusion based on Kalman filtering;

[0013] Step 5: Apply the linear dispersion relationship to the fused wavelengths to invert the underwater topography.

[0014] On the other hand, the present invention provides a SAR shallow sea water depth inversion system based on local wave direction, comprising:

[0015] Image acquisition module, used to obtain SAR images with clear wave characteristics in the study area;

[0016] Image segmentation module is used to divide SAR images into sub-image blocks, and estimate the local wave direction of each sub-image block based on the image spectrum, where M and N are both natural numbers greater than 1;

[0017] A window angle setting module calculates a reference window angle of each sub-image block based on the local wave direction, and sets a window angle range and a step size of the adaptive multi-angle window based on the reference window angle;

[0018] The Kalman filter fusion module is used to calculate the image spectra of the divided image sequence under different window angles for each sub-image block, estimate the wave wavelength corresponding to each image spectrum, and perform wavelength fusion based on the Kalman filter;

[0019] The water depth inversion module is used to apply the linear dispersion relationship to the fused wavelengths to invert the underwater topography.

[0020] In a third aspect, the present invention provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned SAR shallow sea water depth inversion method based on local wave direction.

[0021] In a fourth aspect, the present invention provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the aforementioned SAR shallow sea water depth inversion method based on local wave direction.

[0022] The beneficial effects of the present invention are:

[0023] 1. The present invention starts from SAR shallow sea underwater terrain detection and realizes large-area, near-real-time shallow sea underwater terrain detection, fully considering the influence of underwater terrain on wave parameters, not only considering the change of wave wavelength, but also additionally introducing local wave direction to provide support for underwater terrain detection, and introducing reference window angle based on local wave direction, thereby improving the accuracy of underwater terrain detection by improving the accuracy of wavelength estimation.

[0024] 2. The present invention improves the robustness of water depth inversion by fusing the wavelength sequence estimated by multi-angle windows based on local wave directions during SAR shallow sea underwater terrain detection, and provides support for the engineering application of SAR shallow sea underwater terrain detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a flow chart of the SAR shallow sea water depth inversion method based on local wave direction of the present invention;

[0026] Figure 2 This is a schematic diagram of the implementation principle of the SAR shallow sea water depth inversion method based on local wave direction of the present invention;

[0027] Figure 3 A schematic diagram of a SAR image of a research area provided by an embodiment of the present invention;

[0028] Figure 4 A schematic diagram of nautical chart data of a research area provided by an embodiment of the present invention;

[0029] Figure 5 This is a result diagram of the local wave direction estimation of each sub-image block of the ocean wave according to an embodiment of the present invention;

[0030] Figure 6 A schematic diagram of underwater terrain inverted by an embodiment of the present invention;

[0031] Figure 7 This is an error scatter plot drawn based on the inversion results according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0033] like Figure 1As shown, the present invention provides a SAR shallow sea water depth inversion method based on local wave direction, which mainly improves the prior art method from the following aspects. Due to the different window angle selections of each sub-image block, there are differences in the wavelength estimation results. The present invention calculates the reference window angle of the corresponding position based on the local wave direction estimated by each sub-image block initially divided. When the window angle is consistent with the local wave direction of the wave, the wavelength estimation is more accurate. At this time, the window angle is the reference window angle; based on the reference window angle, an adaptive multi-angle window is used to fuse the wavelength estimated by the sub-image under the multi-angle window as the wavelength estimated at the position; based on the divided sub-image sequence under the multi-angle window, the wavelength sequence of each sub-image block is estimated, and according to the set window angle range, the sub-image sequence at different angles is obtained. It should be noted that the window angle of each sub-image block is different, depending on the local wave direction; the wavelength sequence estimated at different window angles is fused by Kalman filtering to obtain the fused wavelength of each sub-image for subsequent water depth inversion. Kalman filtering is performed on the wavelength estimated by the sub-image sequence obtained at different window angles for each sub-image block to obtain the fused wavelength of each sub-image position. Specifically, the method includes the following steps:

[0034] Step 1: Obtain a SAR image with clear wave characteristics in the study area;

[0035] Step 2: Divide the SAR image into sub-image blocks, and estimate the local wave direction of each sub-image block based on the image spectrum, where M and N are both natural numbers greater than 1;

[0036] Step 3: Calculate the reference window angle of each sub-image block based on the local wave direction, and set the window angle range and step size of the adaptive multi-angle window based on the reference window angle;

[0037] Step 4: Calculate the image spectrum of the divided image sequence under different angle windows for each sub-image block, estimate the wave wavelength corresponding to each image spectrum, and perform wavelength fusion based on Kalman filtering;

[0038] Step 5: Apply the linear dispersion relationship to the fused wavelengths to invert the underwater topography.

[0039] In some embodiments, a research area is selected, where the underwater topography often changes greatly, frequent in-situ detection is difficult to implement, and remote sensing data is needed to provide large-area near-real-time underwater topography detection data. Since it is necessary to use the changes in wave characteristics to invert the water depth, the area is required to have well-developed long swells.

[0040] Since the wave characteristics of SAR images are easily affected by speckle noise, the acquired SAR images should first be preprocessed by Lee filtering, and then the SAR images should be divided into sub-image blocks, , are all natural numbers greater than 1. They are used to detect underwater terrain from SAR images in the following flow chart. Figure 2 shown.

[0041] For each sub-image block, the image spectrum is calculated based on Fourier transform:

[0042] (1)

[0043] in, Represents the coordinates of the sub-image block in the horizontal and vertical directions, and Sub-image block Direction and The width of the direction, represents a sub-image block, represents the exponential function, represents the imaginary unit, and express , Directional wave number;

[0044] The local wave direction of each sub-image block is calculated according to the peak position of the image spectrum :

[0045] (2)

[0046] in, and represents the peak coordinates in the wavenumber domain space, and arctan represents the inverse tangent function.

[0047] The wavelength estimation results are different due to different window angles for each sub-image block. The reference window angle at the corresponding position is calculated based on the estimated local wave direction of each sub-image block divided initially. The window angle when the local wave direction of the sea wave is consistent with the window angle is recorded as the reference window angle. Indicates that an adaptive multi-angle window is set based on the reference window angle, and the wavelength estimated by the sub-image block under the multi-angle window is fused as the wavelength estimated at the corresponding position. The reference window angle is taken as the center, and the window angle range is set before and after it. , set the step size Evenly generated window angles, where is the number of windows, . Use formula (3) to calculate the angle corresponding to each window , get the adaptive angles corresponding to the multiple windows at the position .

[0048] (3)

[0049] in, The serial number representing the adaptive window angle;

[0050] According to the window angle range set and step length , get the corresponding sub-image sequence of each sub-image block under multiple window angles , among which, among which, Indicates the sequence number of the sub-image block into which the SAR image is initially divided. , Indicates the angles corresponding to different windows at this position;

[0051] Calculate the image spectrum of each sub-image sequence at different angles, and calculate the wavelength of the main wave according to the peak position of the image spectrum. :

[0052] (4)

[0053] in, and Represents the peak coordinates in the wavenumber domain.

[0054] Calculate the wavelength sequence corresponding to each position , perform Kalman filtering based on the wavelength sequence corresponding to each position to obtain the fusion wavelength of each sub-image , the state equation and observation equation of Kalman filter are:

[0055] (5)

[0056] (6)

[0057] in, represents the state vector, represents the observation vector, yes Time has come The state transfer matrix at time , for The observation matrix at time, and is uncorrelated noise.

[0058] When the waves are in deep water, the propagation of the waves is basically not affected by the underwater terrain, and the linear diffusion relationship in the deep water area is satisfied. The wavelength estimated by the deep water sub-image block and the diffusion relationship in the deep water area are used to estimate the wave period of the deep water waves. , which is approximated as the ocean wave period in the study area, and it is assumed that the ocean wave period remains unchanged in the entire study area.

[0059] (7)

[0060] in, is the acceleration due to gravity, Indicates the main wave wavelength.

[0061] The water depth of each sub-image block is inverted based on the linear diffusion relationship, and finally the underwater topography of the area to be detected is given.

[0062] (8)

[0063] in, For water depth, is the main wave length, is the deep-water wave period, is the hyperbolic tangent function.

[0064] On the other hand, the present invention provides a shallow sea water depth inversion system based on a local wave direction adaptive window of a SAR image, the system comprising:

[0065] Image acquisition module, used to obtain SAR images with clear wave characteristics in the study area;

[0066] Image segmentation module is used to divide SAR images into sub-image blocks, and estimate the local wave direction of each sub-image block based on the image spectrum, where M and N are both natural numbers greater than 1;

[0067] A window angle setting module calculates a reference window angle of each sub-image block based on the local wave direction, and sets a window angle range and a step size of the adaptive multi-angle window based on the reference window angle;

[0068] The Kalman filter fusion module is used to calculate the image spectra of the divided image sequence under different window angles for each sub-image block, estimate the wave wavelength corresponding to each image spectrum, and perform wavelength fusion based on the Kalman filter;

[0069] The water depth inversion module is used to apply the linear dispersion relationship to the fused wavelengths to invert the underwater topography.

[0070] The following is a further detailed description of a SAR shallow water depth inversion method and system based on local wave direction proposed by the present invention through specific embodiments:

[0071] like Figure 2 As shown, first, this embodiment selects Australia's Rottnest Island as the study area. This area is exposed to the ocean all year round, has well-developed long swells, and has large terrain changes. A scene of GF3 SAR data obtained in this area is selected to intercept the study area, such as Figure 3As shown, download the nautical chart data of the area. Here we use GMRT (Global Multi-Resolution Topography) data, such as Figure 4 shown.

[0072] Then we start to detect underwater terrain with the data. First, we perform Lee filtering on the data to minimize the impact of speckle noise on underwater terrain detection. sub-image blocks, M and N are both natural numbers greater than 1, and the initial setting is , according to formula (1) calculate the The image spectrum of the sub-image block is estimated according to formula (2): The local wave direction of the sub-image block ,like Figure 5 As shown, calculate its base window angle Set the window angle range and step length , calculate the The sub-image block is adaptive to multiple window angles, which is given by formula (3). Then, for each sub-image block in the sub-image sequence, the main wavelength is calculated based on formula (4) to obtain the corresponding wavelength sequence.

[0073] Then, according to the Kalman filter algorithm given in (5)-(6), the wavelength sequence estimated under different angle windows of each sub-image block is Kalman filtered to calculate the first The fusion wavelengths of the sub-image blocks under different angle windows. If it does not meet ,but , continue to estimate the fusion wavelength of the next sub-image block, if it satisfies , the wavelength estimation result of the whole study area can be obtained. The sub-image block of the deep water area in the data is selected, and the deep water wave period is calculated based on formula (7), which is approximated as the wave period of the study area, and it is assumed that the wave period of the whole study area is constant.

[0074] Finally, according to the fusion wavelength and estimated wave period of each sub-image block, the linear diffusion relationship given by formula (8) is applied to estimate the water depth of each sub-image block, and finally the underwater topography of the area is obtained, as shown in Figure 6 Compare the chart data and draw an error scatter plot, as shown in Figure 7As shown in the figure, the mean absolute error (MAE), mean relative error (MRE) and root mean square error (RMSE) are calculated respectively. The MAE is 3.67m, the MRE is 12.75%, and the RMSE is 4.72m, which effectively reduces the error of underwater terrain inversion based on SAR images in this area.

[0075] This completes the SAR shallow sea underwater topography detection.

[0076] In a third aspect, the present invention provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned SAR shallow sea water depth inversion method based on local wave direction.

[0077] In a fourth aspect, the present invention provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the aforementioned SAR shallow sea water depth inversion method based on local wave direction.

[0078] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A SAR shallow water depth inversion method based on local wave direction, characterized in that: The steps include: Step 1: Obtain a SAR image with clear wave characteristics in the study area; Step 2: Divide the SAR image into sub-image blocks, and estimate the local wave direction of each sub-image block based on the image spectrum, where M and N are both natural numbers greater than 1; Step 3, calculating the reference window angle of each sub-image block based on the local wave direction, and setting the window angle range and step size of the adaptive multi-angle window based on the reference window angle; including: The reference window angle of the corresponding position of the sub-image block is calculated based on the estimated local wave direction of each sub-image block, where the window angle when the local wave direction of the wave is consistent with the window angle is recorded as the reference window angle, and express; At the base window angle Based on this, an adaptive multi-angle window is set, and the wavelength estimated by the sub-image block under the multi-angle window is fused as the wavelength estimated at the corresponding position: Centered on the left and right sides, set the window angle range , by step length Evenly generated window angles, where , use formula (3) to calculate the adaptive angle corresponding to each window : （3） The serial number representing the adaptive window angle; Step 4: Calculate the image spectra of the divided image sequence under different window angles for each sub-image block, estimate the wave wavelength corresponding to each image spectrum, and perform wavelength fusion based on Kalman filtering; Step 5: Apply the linear dispersion relationship to the fused wavelengths to invert the underwater topography.

2. The SAR shallow sea water depth inversion method based on local wave direction according to claim 1 is characterized in that: The step 1 also includes applying Lee filtering to pre-process the acquired SAR image.

3. The SAR shallow sea water depth inversion method based on local wave direction according to claim 1 is characterized in that: In step 2, the image spectrum is calculated based on Fourier transform for each sub-image block. : （1） in, Represents the coordinates of the sub-image block in the horizontal and vertical directions, and For sub-image blocks along Direction and The width of the direction, represents a sub-image block, represents the exponential function, represents the imaginary unit, and Indicates along Direction and The wave number of the direction; According to the image spectrum The local wave direction of each sub-image block is calculated by the peak position of the spectrum : （2） in, and represents the peak coordinates in the wavenumber domain space, and arctan represents the inverse tangent function.

4. The SAR shallow sea water depth inversion method based on local wave direction according to claim 1 is characterized in that: In step 4, according to the window angle range set and step length , get the corresponding sub-image sequence of each sub-image block under multiple window angles ,in, Indicates the sequence number of the sub-image block into which the SAR image is initially divided. , Indicates the adaptive angle corresponding to different windows at the sub-image block position; The image spectrum of each sub-image block under different window angles is calculated respectively, and the wavelength of the main wave of the wave is calculated according to the peak position of the image spectrum. : （4） in, and Represents the peak coordinates in the wavenumber domain.

5. The SAR shallow sea water depth inversion method based on local wave direction according to claim 1 is characterized in that: In step 4, the wavelength sequence corresponding to the position of each sub-image block is calculated. , based on Kalman filtering, the fusion wavelength of each sub-image block is obtained The state equation and observation equation of Kalman filter are: （5） （6） in, represents the state vector, represents the observation vector, yes Time has come The state transfer matrix at time , for The observation matrix at time, and is uncorrelated noise.

6. The SAR shallow sea water depth inversion method based on local wave direction according to claim 1 is characterized in that: In step 5, when the sea waves are in the deep water area, the wavelength estimated by the deep water sub-image block and the deep water area diffusion relationship are used to estimate the deep water wave period. , which is approximated as the wave period of the study area, and it is assumed that the wave period of the entire study area remains unchanged: （7） in, is the acceleration due to gravity, Indicates the main wave wavelength; The water depth of each sub-image block is inverted based on the linear diffusion relationship, and the underwater topography of the area to be detected is finally obtained: （8） in, For water depth, is the main wave length, is the deep-water wave period, is the hyperbolic tangent function.

7. A SAR shallow water depth inversion system based on local wave direction, characterized in that: include: Image acquisition module, used to obtain SAR images with clear wave characteristics in the study area; Image segmentation module is used to divide SAR images into sub-image blocks, and estimate the local wave direction of each sub-image block based on the image spectrum, where M and N are both natural numbers greater than 1; The window angle setting module calculates the reference window angle of each sub-image block based on the local wave direction, and sets the window angle range and step size of the adaptive multi-angle window based on the reference window angle; including: The reference window angle of the corresponding position of the sub-image block is calculated based on the estimated local wave direction of each sub-image block, where the window angle when the local wave direction of the sea wave is consistent with the window angle is recorded as the reference window angle, and express; At the base window angle Based on this, an adaptive multi-angle window is set, and the wavelength estimated by the sub-image block under the multi-angle window is fused as the wavelength estimated at the corresponding position: Centered on the left and right sides, set the window angle range , by step length Evenly generated window angles, where , use formula (3) to calculate the adaptive angle corresponding to each window : （3） The serial number representing the adaptive window angle; The Kalman filter fusion module is used to calculate the image spectra of the divided image sequence under different window angles for each sub-image block, estimate the wave wavelength corresponding to each image spectrum, and perform wavelength fusion based on the Kalman filter; The water depth inversion module is used to apply the linear dispersion relationship to the fused wavelengths to invert the underwater topography.

8. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; Wherein, when one or more programs are executed by the one or more processors, the one or more processors implement the SAR shallow sea water depth inversion method based on local wave direction as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that: Executable instructions are stored thereon, and when the instructions are executed by the processor, the processor can implement the SAR shallow sea water depth inversion method based on local wave direction as described in any one of claims 1-6.

Images