Smooth filtering method based on multi-beam narrowband warning line integration

Abstract

The invention relates to the technical field of sonar signal processing, in particular to a smooth filtering method based on multi-beam narrowband warning line integration, which comprises the following steps of: (1) acquiring multi-beam narrowband data; (2) pre-processing the acquired multi-beam narrowband data, and supplementing a data value of a half-beam size before and after the data; (3) converting the multi-beam narrowband data into projection data through line integral transformation, and filtering the projection data in a projection domain; and (4) carrying out back projection on the filtered projection data to obtain reconstructed narrowband data, and carrying out linear integral smooth filtering on the multi-beam narrowband data to obtain a high-resolution space energy spectrum.

Description

Technical Field

[0001] This invention relates to the field of sonar signal processing technology, specifically to a smoothing filtering method based on multi-beam narrowband warning line integration. Background Technology

[0002] Multipath propagation in a sonar system refers to the phenomenon where sound waves emitted from a source, during propagation, encounter different medium interfaces or obstacles, resulting in reflection, refraction, scattering, and other phenomena that cause the sound signal to reach the receiving point through multiple paths. Different propagation paths lead to differences in the received signal's time, phase, and amplitude, thus affecting signal quality and the accuracy of target detection. In actual operating conditions, the sonar detection process is simultaneously affected by various noises. Noise is mainly divided into marine environmental noise and platform self-noise. These various noise sources together constitute the sonar's receiving noise model, and the resulting multipath effects pose challenges to sonar signal detection.

[0003] The phenomenon of multipath propagation of towed platform noise is particularly evident. The effects of multipath propagation of towed platform noise and strong interference side lobes / grating lobes exhibit characteristics inconsistent with the real target signal in the multi-beam narrowband spatial spectrum. That is, the real signal exhibits spatial consistency characteristics, with all frequency components of the signal basically located near the same spatial beam, while the multipath propagation of towed platform noise and interference such as strong interference side lobes / grating lobes are distributed across different beams in the narrowband beam space, exhibiting a "non-vertical, tilted arc" characteristic.

[0004] Traditional high-resolution narrowband early warning methods using towed linear arrays directly weight average or sum the spatial energy spectra of each sub-band when obtaining the multi-beam sub-bands. This results in the multipath propagation of towed platform noise and the superposition of strong interference sidelobes / grating lobes introduced by array model mismatch onto the broadband spatial energy spectrum, thus weakening the algorithm's ability to detect weak signals. Line integrals are an important concept in mathematical analysis, used to calculate the cumulative effect of function values ​​or distributions along a certain path. Therefore, these spurious interferences can be suppressed by using line integral smoothing filtering in the narrowband beam space, achieving purification of the true target signal spectrum characteristics in the narrowband beam space. Then, by weighted averaging or summing the sub-bands, a high-resolution spatial energy spectrum after suppressing "multipath propagation of towed platform noise and strong interference sidelobes / grating lobes" can be obtained. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a smoothing filtering method based on multi-beam narrowband warning line integration.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a smoothing filtering method based on multi-beam narrowband warning line integration, characterized in that its steps are as follows:

[0007] (1) Acquire multi-beam narrowband data;

[0008] (2) Preprocess the acquired multi-beam narrowband data by adding half-beam size data values ​​before and after the data;

[0009] (3) Convert the multibeam narrowband data into projection data by line integral transform, and then filter the projection data in the projection domain;

[0010] (4) Backproject the filtered projection data to obtain the reconstructed narrowband data.

[0011] In some embodiments, step (1) is specifically implemented as follows:

[0012] ① Calculate the array manifold matrix: , This represents the total number of array elements in the linear array. For the spacing between array elements, For the speed of sound, , To handle the upper and lower limits of the frequency band, ;

[0013] ② Utilize the array manifold matrix to analyze the frequency domain array data of each channel. By performing guided weighting, the phase-compensated array output matrix is ​​obtained:

[0014]

[0015] in, This is the conjugate transpose of a complex matrix. For corresponding Frequency, The diagonal matrix formed by the array manifold matrices corresponding to the directions is called the focusing matrix, and is defined as:

[0016] ;

[0017] ③ By focusing the frequency domain array data using the focusing matrix, the steering power spectral density matrix is ​​obtained:

[0018]

[0019] in for The covariance matrix of frequency points ,in For frequency domain data snapshots;

[0020] ④ Calculate the optimal weight vector and beam output for adaptive waveform restoration:

[0021]

[0022] Array beam output data for:

[0023]

[0024] in To find the inverse of the power spectral density matrix, It is a unit array.

[0025] In some embodiments, step (2) is specifically performed as follows:

[0026] Add a value equal to half the number of beams to the beginning and end of the column for multibeam narrowband data, add a value equal to half the number of frequency points to the beginning and end of the row, add a value equal to the second half the number of beams to the beginning of the column, and so on.

[0027] Suppose that for a size of Two-dimensional multibeam data , For the number of beams, The number of frequency points refers to the data after preprocessing. Size is .

[0028] In some embodiments, step (3) is specifically implemented as follows:

[0029] Multibeam narrowband data is converted into projected data using Radon transform.

[0030]

[0031] in, It is the distance from the straight line to the origin, and the projection angle. Set to 0 to 180 degrees. It is the Dirac function, representing the function along the straight line. integral;

[0032] Narrowband data is transformed to obtain projected data values , It is a function of distance and angle.

[0033] In some embodiments, according to step (3), the value corresponding to the direction with a projection angle of 0° is retained, and the values ​​in other directions are filtered.

[0034] In some embodiments, step (4) is specifically implemented as follows:

[0035] Perform an inverse line integral transform on the filtered data.

[0036]

[0037] in, It refers to the projection angle, which ranges from 0 to 180 degrees. yes Point at angle The projection position below, This is a filtering function used to eliminate high-frequency noise. The filtering function uses a Ram-Lak filter.

[0038]

[0039] Two-dimensional data are obtained after the inverse Radon transform. ,right Extract the data to obtain the desired result. Multi-beam data of varying sizes.

[0040] Compared with the prior art, the beneficial effects of the present invention are: by using the vertical line integral, the proportion of non-vertical direction after integration is reduced, the noise energy of the arc-shaped distribution is eliminated, and the influence of strong interference side lobes / gate lobes is reduced.

[0041] The line integral smoothing method proposed in this application is designed based on the multipath and grating lobe phenomena existing in the towed array platform. After applying line integral smoothing filtering to multibeam narrowband data, this method can obtain a high-resolution spatial energy spectrum.

[0042] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. The embodiments of this application will provide a detailed description and understanding of the application. Attached Figure Description

[0043] Figure 1 This is a flowchart of the method of the present invention;

[0044] Figure 2 Diagram showing multipath propagation and grating lobe phenomena in multibeam narrowband;

[0045] Figure 3 The distribution of the results of the Radon transform;

[0046] Figure 4 The result is shown in the figure after line integral filtering.

[0047] Figure 5 For spatial energy spectrum;

[0048] Figure 6 The result is processed using a traditional background equalization method;

[0049] Figure 7 This is a graph comparing the algorithm performance. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Please see Figures 1 to 4 This invention provides a technical solution: a smoothing filtering method based on multi-beam narrowband warning line integration, the technical steps and specific implementation methods of which are as follows:

[0052] (1) Acquisition of multi-beam narrowband data:

[0053] 1) Calculate the array manifold matrix: , This represents the total number of array elements in the linear array. For the spacing between array elements, For the speed of sound, , To handle the upper and lower limits of the frequency band, .

[0054] 2) Utilize the array manifold matrix to analyze the frequency domain array data of each channel. By performing guided weighting, the phase-compensated array output matrix is ​​obtained.

[0055]

[0056] in, This is the conjugate transpose of a complex matrix. For corresponding Frequency, The diagonal matrix formed by the array manifold matrices corresponding to the directions is called the focusing matrix, and is defined as:

[0057] ;

[0058] 3) The frequency domain array data is focused using a focusing matrix to obtain the steering power spectral density matrix.

[0059]

[0060] in for The covariance matrix of frequency points ,in For frequency domain data snapshots;

[0061] 4) Calculate the optimal weight vector and beam output for adaptive waveform restoration.

[0062]

[0063] Array beam output data for

[0064]

[0065] in To find the inverse of the power spectral density matrix, It is a unit array.

[0066] (2) Preprocessing of multi-beam narrowband data

[0067] To prevent data from exceeding limits during online integration and to ensure the accuracy of subsequent processing, a data value equal to half the beam number is added before and after the column of multi-beam narrowband data; a data value equal to half the frequency point number is added before and after the row; a data value equal to the second half of the beam number is added before the column, and so on. For a data of size... Two-dimensional multibeam data , For the number of beams, The number of frequency points refers to the data after preprocessing. Size is .

[0068] (3) Line integral transform and filtering

[0069] Multibeam narrowband data is converted into projected data using Radon transform.

[0070]

[0071] in, It is the distance from the straight line to the origin, and the projection angle. Typically, the temperature ranges from 0 to 180 degrees. It is the Dirac function, representing the function along the straight line. integral.

[0072] Narrowband data is transformed to obtain projected data values , It is a function of distance and angle. This method retains the value corresponding to the 0° direction and filters the values ​​in other directions. Figure 3 For example, the area with the highest brightness is This indicates that it is perpendicular to The region with the largest value represents the distribution area of ​​multipath propagation. This range... The value is set to 0. and The value of the region is set to a threshold as an adjustment factor. and mean , The threshold value is set to 0.1. Projected values ​​exceeding the threshold are set to 0, while the rest are retained, resulting in the final filtered data. .

[0073] (4) Inverse line integral transform

[0074] Perform an inverse line integral transform on the filtered data.

[0075]

[0076] in, It refers to the projection angle, which ranges from 0 to 180 degrees. yes Point at angle The projection position below. This is a filtering function used to eliminate high-frequency noise. The filtering function uses a Ram-Lak filter.

[0077]

[0078] Two-dimensional data are obtained after the inverse Radon transform. ,right Extract the data to obtain the desired result. Multi-beam data of varying sizes.

[0079] According to the technical solution of this application, the noise and interference model of the towing platform shows that the noise multipath propagation and strong interference sidelobe / grating lobe phenomenon exhibit a "non-vertical, inclined arc" characteristic in the narrowband spatial distribution. The final spatial energy spectrum is obtained by summing the energy according to the orientation. Therefore, by using the vertical line integral and reducing the proportion of the non-vertical direction after integration, the arc-shaped noise energy can be eliminated, and the influence of strong interference sidelobe / grating lobe can be reduced.

[0080] The Radon transform is a special type of line integral. Its core principle is to integrate two-dimensional data along a set of straight lines to obtain projected data containing both the projected values ​​and the projected orientation. This projected data can be viewed as the line integrals of the data at different angles. Based on the idea of ​​transform domain filtering, the projected domain data can be filtered to eliminate the effects of multipath propagation and interference. The filtering process can be carried out in two ways. First, because the real signal exhibits consistency in the multi-beam narrowband, it is continuous in the frequency domain and invariant in the beam domain. Noise multipath propagation and strong interference sidelobe / grating lobe effects exhibit a tilted arc-shaped characteristic. After projection through the line integral, they converge at larger angle values. By weakening the line integral values ​​in non-perpendicular directions and retaining the line integral values ​​in the perpendicular directions, the multipath propagation of noise and grating lobe phenomena (such as...) can be effectively suppressed. Figure 2The image shows the effect of noise, exhibiting significant multipath propagation and grating lobe phenomena, requiring post-processing correction. Secondly, considering that the size of the projection domain data corresponds to the continuous value of the integration angle, the impact of multipath propagation does not cover the entire multibeam narrowband region. When reducing the line integral value in non-perpendicular directions, the mean and coefficient factor of that direction are used as a threshold to filter out larger projection values. This cleans up the continuous linear region formed by multipath propagation while retaining useful information in that azimuth. Finally, the projection values ​​are converted back to narrowband data using the inverse Radon transform.

[0081] The proposed line integral smoothing method is designed based on the multipath and grating lobe phenomena present in towed array platforms. After applying line integral smoothing filtering to multibeam narrowband data, the method yields a high-resolution spatial energy spectrum. This work has been validated using simulation and real sea trial data.

[0082] The implementation process is as follows: First, to avoid data out of bounds, the multi-beam narrowband data is preprocessed by adding half-beam-sized data values ​​before and after the data; second, the multi-beam narrowband data is converted into projection data through line integral transform, and the projection data is filtered in the projection domain; finally, the filtered projection data is back-projected to obtain the reconstructed narrowband data.

[0083] like Figure 5 As shown, the spatial energy spectrum of the multibeam narrowband data after line integral filtering has a much higher azimuth resolution than the original method, which indicates that the present invention effectively enhances the system's ability to detect weak targets.

[0084] like Figure 6 The image shown is the result obtained by processing using the traditional background equalization method;

[0085] like Figure 7 As shown, the spatial energy spectrum of multibeam narrowband data after line integral filtering is shown. Blue represents the phenomenon under multipath propagation and strong interference sidelobes / grating lobes, red represents the result of traditional background equalization processing, and yellow represents the effect of processing by this intellectual achievement. After post-processing using this method, the azimuth resolution capability of the algorithm far exceeds that of traditional methods, which indicates that this intellectual achievement effectively enhances the system's ability to detect weak targets.

[0086] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smoothing filtering method based on multi-beam narrowband warning line integration, characterized in that: The steps are as follows: (1) Acquire multi-beam narrowband data; (2) Preprocess the acquired multi-beam narrowband data by adding half-beam size data values ​​before and after the data; (3) Convert the multibeam narrowband data into projection data by line integral transform, and then filter the projection data in the projection domain; (4) Backproject the filtered projection data to obtain the reconstructed narrowband data.

2. The smoothing filtering method based on multi-beam narrowband warning line integration according to claim 1, characterized in that: According to step (1), the specific method is as follows: ① Calculate the array manifold matrix: , This represents the total number of array elements in the linear array. For the spacing between array elements, For the speed of sound, , To handle the upper and lower limits of the frequency band, ; ② Utilize the array manifold matrix to analyze the frequency domain array data of each channel. By performing guided weighting, the phase-compensated array output matrix is ​​obtained: in, This is the conjugate transpose of a complex matrix. For corresponding Frequency point, The diagonal matrix formed by the array manifold matrices corresponding to the directions is called the focusing matrix, and is defined as: ； ③ By focusing the frequency domain array data using the focusing matrix, the steering power spectral density matrix is ​​obtained: in for The covariance matrix of frequency points ,in For frequency domain data snapshots; ④ Calculate the optimal weight vector and beam output for adaptive waveform restoration: Array beam output data for: in To find the inverse of the power spectral density matrix, It is a unit array.

3. The smoothing filtering method based on multi-beam narrowband warning line integration according to claim 2, characterized in that: According to step (2), the specific method is as follows: Add a value equal to half the number of beams to the beginning and end of the column for multibeam narrowband data, add a value equal to half the number of frequency points to the beginning and end of the row, add a value equal to the second half the number of beams to the beginning of the column, and so on. Suppose that for a size of Two-dimensional multibeam data , For the number of beams, The number of frequency points refers to the data after preprocessing. Size is .

4. The smoothing filtering method based on multi-beam narrowband warning line integration according to claim 3, characterized in that: According to step (3), the specific method is as follows: Multibeam narrowband data is converted into projected data using Radon transform. in, It is the distance from the straight line to the origin, and the projection angle. Set to 0 to 180 degrees. It is the Dirac function, representing the function along the straight line. integral; Narrowband data is transformed to obtain projected data values , It is a function of distance and angle.

5. The smoothing filtering method based on multi-beam narrowband warning line integration according to claim 4, characterized in that: According to step (3), the value corresponding to the direction with a projection angle of 0° is retained, and the values ​​in other directions are filtered.

6. The smoothing filtering method based on multi-beam narrowband warning line integration according to claim 5, characterized in that: According to step (4), the specific method is as follows: Perform an inverse line integral transform on the filtered data. in, It refers to the projection angle, which ranges from 0 to 180 degrees. yes Point at angle The projection position below, This is a filtering function used to eliminate high-frequency noise. The filtering function uses a Ram-Lak filter. The inverse Radon transform yields two-dimensional data. ,right Extract the data to obtain the desired result. Multi-beam data of varying sizes.

Images