A deconvolution imaging method for shift-variable array multi-beam synthetic aperture sonar

Abstract

The present invention belongs to the field of measurement technology, and specifically relates to a shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method. The method includes: performing conventional beamforming processing on the array element signals received by the multi-beam synthetic aperture sonar in the horizontal direction to obtain multiple array conventional beamforming outputs; performing aperture synthesis processing on the multiple array conventional beamforming outputs in the track direction to obtain conventional multi-beam synthetic aperture sonar imaging outputs; traversing and calculating the point spread function matrix according to the multi-beam synthetic aperture sonar array flow type and preset beam angle; establishing a linear equation group of difference functions between conventional beamforming and the azimuth of the target sound source; iteratively estimating the target azimuth function according to the estimation criterion of the minimum difference function, and obtaining the shift-variable array multi-beam synthetic aperture sonar deconvolution imaging results. The present invention can improve the degree of refinement of multi-beam synthetic aperture sonar imaging and effectively improve the flexibility of application scenarios of multi-beam synthetic aperture sonar.

Description

Technical Field

[0001] The invention belongs to the field of measurement technology, and in particular relates to a shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method. Background Art

[0002] Multibeam bathymetric sonar is a widely used underwater target imaging device that can obtain imaging results in three-dimensional space. However, it is limited by conventional beamforming theory, and its imaging resolution in the horizontal direction is heavily dependent on the array aperture. The resolution in the track direction drops sharply with the increase of detection distance, and the imaging quality is poor in working scenarios such as small array scale and long-distance detection. The development of multibeam synthetic aperture technology can improve the above problems to a certain extent. In particular, the introduction of deconvolution beamforming technology can achieve imaging effects that break through the Cramer-Rao bound, so that multibeam synthetic aperture sonar can obtain a constant imaging resolution in the track direction, further reduce the beam width in the horizontal direction, and improve the main-sidelobe ratio of beamforming.

[0003] The commonly used deconvolution technology at this stage requires that the array element arrangement has the property of linear shift invariance, or has the property of shift invariance on the cosine axis after coordinate transformation, and requires the preset beam angle to be uniformly distributed on the cosine axis. However, for the typical application of multi-beam synthetic aperture sonar, although its linearly arranged array elements show the property of shift invariance on the cosine axis after coordinate transformation, for its most commonly used working scenarios, its beam angle should be uniformly distributed or uniformly distributed on the tangent axis, which does not meet the prerequisite requirement of shift invariance on the cosine axis, thus limiting the solution of key problems such as flexible allocation of sonar beam angles, motion attitude compensation, and non-uniform array calibration. Summary of the invention

[0004] The present invention aims at the limitation that the existing multi-beam synthetic aperture sonar deconvolution imaging method requires the system to have linear shift invariance, pre-constructs a point spread function matrix of arbitrary beam angles, uses the fastest negative gradient projection method to iteratively search the target sound source orientation, and obtains the least squares solution of the target sound source orientation estimation, thereby improving the refinement of multi-beam synthetic aperture sonar imaging. The present invention can perform deconvolution operations on arbitrarily distributed preset beam angles, effectively improving the flexibility of application scenarios of multi-beam synthetic aperture sonar.

[0005] The present invention provides a shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method, comprising the following steps:

[0006] Step (1): performing conventional beamforming processing on the array element signals received by the multi-beam synthetic aperture sonar in the horizontal direction to obtain multiple array conventional beamforming outputs;

[0007] Step (2): performing aperture synthesis processing on the conventional beamforming outputs of the multiple arrays in the track upward direction to obtain conventional multi-beam synthetic aperture sonar imaging output;

[0008] Step (3): traverse and calculate the point spread function matrix according to the multi-beam synthetic aperture sonar array flow pattern and the preset beam angle;

[0009] Step (4): Establish a linear equation group of difference functions between conventional beamforming and the target sound source azimuth; perform iterative estimation of the target azimuth function according to the estimation criterion of minimum difference function, and obtain the deconvolution imaging result of the shift-variable array multi-beam synthetic aperture sonar.

[0010] Preferably, in step (1), the conventional beamforming output is traversed according to the preset beam angle and time to obtain a sequence of conventional beamforming outputs of all arrays: , expressed as:

[0011] ;

[0012] in, To preset the beam angle, To traverse the target location; is the number of array elements of multi-beam synthetic aperture sonar, is the horizontal array element spacing, is the echo signal wavelength, For the moment.

[0013] Preferably, in step (2), the conventional multi-beam synthetic aperture sonar imaging output is expressed as :

[0014] ;

[0015] in, is the total number of array entries; It is a single array output; For the Horizontal receiving array.

[0016] Preferably, in step (3), the point spread functions of all the azimuth angles of the target are traversed to obtain a shifted array point spread function matrix, which is expressed as:

[0017] ;

[0018] in, is the shift array point spread function matrix; , … Indicates the 1st to the The location of the target, Preset the total number of beam angles for the imaging system; is the point spread function of the target location, with the superscript Transpose the matrix.

[0019] Preferably, in step (4), the establishment of the linear equation group of the difference function between conventional beamforming and the target sound source orientation includes:

[0020] (1) When the scanning target is When the position is The angle-dimensional data in the angular dimension is used to obtain the angle-dimensional conventional beamforming output matrix of the multi-beam synthetic aperture sonar image: ;

[0021] (2) Construct a linear equation system of the difference function between the positive convolution output and the conventional beamforming output:

[0022] ;

[0023] in, is the difference function, is the target orientation function to be estimated, is the 2-norm.

[0024] Preferably, in step (4), the iterative estimation of the target orientation function uses the fastest negative gradient projection method to iteratively search for the target sound source orientation to obtain a least squares solution for the target sound source orientation estimation.

[0025] The present invention proposes a shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method, which breaks through the linear shift-invariant array limitation of the Richardson-Lucy deconvolution technology of multi-beam synthetic aperture sonar, and extends the deconvolution technology to the condition of any preset beam angle. Compared with the traditional Richardson-Lucy deconvolution algorithm, the method of the present invention does not require the beam angle to be evenly distributed on the cosine axis, and is more suitable for the actual working scene of the multi-beam synthetic aperture sonar with even distribution of the angle axis and the tangent axis. The present invention establishes a linear equation group of conventional beam forming and target sound source orientation by pre-constructing a point spread function matrix of an arbitrary beam angle, and uses the fastest negative gradient projection method to iteratively search the target sound source orientation, thereby obtaining the least squares solution of the target sound source orientation estimation. The present invention can improve the refinement of multi-beam synthetic aperture sonar imaging and effectively improve the flexibility of the application scenarios of multi-beam synthetic aperture sonar. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A schematic diagram of the movement of a shift-variable array multi-beam synthetic aperture sonar in an embodiment of the present invention;

[0027] Figure 2This is a flow chart of a shift-variable array multi-beam synthetic aperture sonar imaging algorithm in an embodiment of the present invention;

[0028] Figure 3 A comparison of the shift-variable array beamforming algorithms in the embodiments of the present invention;

[0029] Figure 4 is the shift-invariant array point spread function in an embodiment of the present invention;

[0030] Figure 5 is the shift-variant array point spread function in an embodiment of the present invention;

[0031] Figure 6 The following is a comparison of imaging effects of different beamforming methods on multiple targets in the embodiments of the present invention;

[0032] Figure 7 The variation trend of the beam width with the number of iterations in the embodiment of the present invention;

[0033] Figure 8 is a conventional multi-beam synthetic aperture time-azimuth diagram in an embodiment of the present invention;

[0034] Fig. 9 It is a time-azimuth diagram of the shift-variant array deconvolution multi-beam synthetic aperture in an embodiment of the present invention;

[0035] Fig.10 It is a multi-beam synthetic aperture track imaging slice in an embodiment of the present invention;

[0036] Fig.11 It is a horizontal imaging slice of a multi-beam synthetic aperture in an embodiment of the present invention;

[0037] Fig.12 It is the multi-beam bathymetric sonar three-dimensional imaging thinning in the embodiment of the present invention;

[0038] Fig.13 It is the conventional multi-beam synthetic aperture three-dimensional imaging thinning in the embodiment of the present invention;

[0039] Fig.14 This is the shift-variant array deconvolution multi-beam synthetic aperture three-dimensional imaging thinning in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0041] like Figure 1As shown in the figure, the shifting array multi-beam synthetic aperture sonar sails along the moving track upward, and the transmitting array elements of the real aperture array illuminate the scanning pixel position at multiple positions. After illuminating the target at multiple positions, the receiving elements of multiple arrays receive the echo signal. The distance between different transmitting positions and scanning pixels is represented by The distance between different receiving primitives and scanned pixels is represented by Indicated by and The scanning delay at different times can be calculated. After the real aperture array moves upward along the track, a virtual synthetic aperture is formed, and the imaging resolution along the track is constant.

[0042] Embodiment 1 The present invention relates to a shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method, the process Figure 2 As shown, first select the horizontal receiving linear array primitive located at the same track position, calculate its steering vector weight, and obtain the horizontal conventional beamforming output of the currently selected linear array after weighted accumulation. Use the same conventional beamforming method to traverse all track linear arrays to obtain multiple array conventional beamforming outputs. Perform aperture synthesis processing on the beamforming outputs of the same preset beam angles corresponding to multiple arrays in the track direction to obtain conventional multi-beam synthetic aperture three-dimensional imaging results. For the obtained three-dimensional spatial imaging results, decompose them into two parts: track dimension imaging and horizontal dimension imaging results. Perform shifted array deconvolution operation on the extracted horizontal dimension imaging results to obtain more refined horizontal imaging output. Then rearrange the track dimension imaging results and the shifted array deconvolution output in three-dimensional space, and finally obtain the three-dimensional imaging output of the shifted array multi-beam synthetic aperture sonar deconvolution. The specific steps are as follows:

[0043] Step (1): Perform conventional beamforming processing on the array element signals received by the multi-beam synthetic aperture sonar in the horizontal direction to obtain multiple array conventional beamforming outputs.

[0044] Taking the uniform linear array multi-beam synthetic aperture sonar as an example, the conventional beamforming output of the uniform linear array can be expressed as the matrix multiplication of the signal correlation matrix and the steering vector. is the time series of the array element signal received by the array, is the guidance vector weight:

[0045] ;

[0046] ;

[0047] in, The preset beam angle is The conventional beamforming output at is the number of array elements of multi-beam synthetic aperture sonar, is the horizontal array element spacing, is the echo signal wavelength, represents the conjugate transpose of the steering vector; is the conjugate transpose of the array element signal.

[0048] By traversing the conventional beamforming output according to the preset beam angle and time, the sequence of conventional beamforming outputs of all arrays can be obtained. The conventional beamforming output of a uniform linear array for a CW pulse signal can be expressed as:

[0049] ;

[0050] in, is the azimuth of the traversal target.

[0051] Step (2): Perform aperture synthesis processing on the conventional beamforming outputs of multiple arrays in the track upward direction to obtain conventional multi-beam synthetic aperture sonar imaging output.

[0052] Assume that the motion of the carrier is From the moment At the end of time, the aperture synthesis output of a single array element can be calculated as:

[0053] ;

[0054] in, To scan the target slant distance, is the symbol of the imaginary part, is a synthetic aperture period, is the speed of sound in water, is the center frequency of the received signal, The scanning target delay is calculated as follows:

[0055] Assume that the transmitting array element is at the coordinate position when the array motion starts , the scanned pixel is located at , when the sonar carrier moves at a speed In uniform motion, The distance from the moment the primitive is emitted to the scanned pixel position It can be expressed as:

[0056] .

[0057] Given any receiving element of the receiving array, such as The first horizontal receiving array The array element has the coordinates at the initial moment: , then in The distance from the scanning pixel to the receiving array element at any time It can be expressed as:

[0058] .

[0059] The two-way propagation distance between the transmitter and the receiver can be determined by the delay and the speed of sound in water It is expressed as:

[0060] .

[0061] By traversing the aperture synthesis output of a single array element according to the scanning distance in the track direction, the aperture synthesis output sequence of all scanning positions of the array element can be obtained. ,in is the track coordinate of the scanning point. The synthetic aperture imaging result of the linear frequency modulation signal can be expressed as follows, where is the signal bandwidth:

[0062] ;

[0063] in, The scanned pixel is located at The slant distance of the scanned target at .

[0064] When the system has a total When the array is a stripe array, the output of a single element can be expressed as , then the preset beam angle is The track-to-synthetic output at position is expressed as :

[0065] .

[0066] This output is the conventional multi-beam synthetic aperture sonar imaging output. After a simple coordinate transformation, the Cartesian coordinate system can be obtained ( ) under 3D sonar image:

[0067] ;

[0068] .

[0069] Step (3): traverse and calculate the point spread function matrix according to the multi-beam synthetic aperture sonar array flow pattern and the preset beam angle.

[0070] For uniform linear array multi-beam synthetic aperture sonar, when the traversal target is located at When the point spread function It can be expressed as:

[0071] .

[0072] The analysis shows that the point spread function is equivalent to the array directivity function, and The axis has the characteristic of linear shift invariance, that is, However, for beam angles in Axis or In the case of uniform distribution on the axis, its linear shift invariance feature does not hold.

[0073] The point spread function of all target azimuth angles is required After traversal, the shift array point spread function matrix can be obtained When the imaging system preset beam angle totals When the point spread function matrix is It can be expressed as:

[0074] ;

[0075] Among them, the superscript is the matrix transpose, , … Indicates the 1st to the The location of a target.

[0076] Step (4): Establish a linear equation system for conventional beamforming and the orientation of the target sound source.

[0077] When the scan target is When the position is The angle-dimensional data in the multi-beam synthetic aperture sonar image is obtained by using the angle-dimensional conventional beamforming output matrix .

[0078] To construct the linear equations of the difference function between the positive convolution output and the conventional beamforming output, it is necessary to iteratively estimate the target orientation function according to the estimation criterion of the minimum difference function:

[0079] ;

[0080] in, is the difference function, is the shift array point spread function matrix, is the target position function to be estimated, is the conventional beamforming output matrix, is the 2-norm.

[0081] Step (5): Use the fastest negative gradient projection method to iteratively search for the target sound source direction, obtain the least squares solution of the target sound source direction estimation, and obtain the deconvolution imaging result of the shifted array multi-beam synthetic aperture sonar.

[0082] The core idea of ​​this method is to search for the direction with the fastest negative gradient descent, and iteratively search for the target position function according to a specific step size. , until the number of iterations is completed or the residual is calculated. Indicates The iterative search results, and its initial value The iterative process is:

[0083] (1) By The result of the iteration Calculate the residual vector :

[0084] ;

[0085] (2) Calculate and search for negative gradient vector:

[0086] ;

[0087] (3) Calculate the projection vector and determine the search path:

[0088] ;

[0089] (4) Calculate the auxiliary vector:

[0090] ;

[0091] (5) Calculate the search step size:

[0092] ;

[0093] (6) Obtain the Iteration search results:

[0094] .

[0095] The orientation estimation function obtained after a certain number of iterations This is the output of the shift-variable array multi-beam synthetic aperture sonar deconvolution imaging. Output results of conventional beamforming with multi-beam synthetic aperture The angle data of the three-dimensional space image can be reconstructed to obtain the shift-variable array multi-beam synthetic aperture sonar deconvolution imaging:

[0096] .

[0097] Example 2 A simulation experiment was carried out using the method of the present invention.

[0098] Sonar imaging in the horizontal direction requires the use of a beamforming algorithm. At present, the most commonly used method is the conventional beamforming (CBF) method, which can image any preset beam angle without the need to follow the linear shift-invariant principle. The imaging effect is stable, but the beam width is large and the sidelobe level is high. The introduction of the deconvolution algorithm can reduce the beam width and improve the main-sidelobe ratio on the basis of CBF. However, the widely used Richardson-Lucy (RL) method is based on the linear shift-invariant characteristics of the system, which limits the flexible application scenarios of the system. Set the multi-element echo of the simulated dual-target, and the positions of the targets are and , the two targets are coherent sound sources and the number of snapshots is 1, the received signal signal-to-noise ratio . Echo signal center frequency , the number of array elements is 32, the array is arranged at half wavelength intervals, the preset beam angle is -90° to 90°, and it is evenly distributed at intervals of 0.1°. Figure 3 As shown in the figure, when the preset beam azimuth is When the axes are evenly distributed, the output of CBF can distinguish the locations of the two targets, but its main lobe is wider and the side lobes are higher. Since the RL method does not satisfy the linear shift-invariant property and has a larger non-uniformity at the outer beam position, a depression appears at the extreme value, forming a double pseudo peak (x-coordinates are 58.6 and 61.2). The shift-variable array imaging method (NNLS) proposed in the present invention can effectively calibrate the location of the target, and the beam width and side lobe level are comparable to those of the RL method, and are greatly improved compared to CBF. Under the premise of maintaining imaging stability, it can effectively improve the performance of the beamforming algorithm.

[0099] Figure 4 Shows the beam angle when the preset When the beam angle is uniformly distributed, it can be observed that the point spread function matrix has the characteristic of linear shift invariance. In this case, the RL algorithm can be used for deconvolution operation. When the axes are evenly distributed, it can be observed that the point spread function matrix is ​​as follows: Figure 5 As shown in Figure 1, it shows the characteristics of nonlinear shift invariance, and its non-uniformity is more obvious in the case of large beam angles. At this time, the RL method will no longer be applicable, and the imaging result will be Figure 3 The pseudo peak situation in .

[0100] Figure 6 The effects of different beamforming algorithms on horizontal imaging are compared in this paper. The simulation parameters and Figure 3Similarly, the multiple signal subspace algorithm (MUSIC) has better beamwidth characteristics than the CBF method, but it requires prior knowledge such as the number of known sources, and the imaging effect of coherent sources is not good. Compared with the MUSIC algorithm, the minimum variance distortionless response algorithm (MVDR) does not require prior knowledge of the number of sources, and can effectively reduce the beamwidth compared to the CBF method, but its sidelobe level is slightly higher. Although the beamwidth of the NNLS method proposed in the present invention is slightly inferior to that of the MUSIC and MVDR methods at -3dB, its beamwidth decreases faster, and its sidelobe level is better than that of the MUSIC and MVDR methods. It can effectively reduce the beamwidth and improve the main-to-sidelobe ratio while maintaining the robustness of the CBF algorithm.

[0101] Figure 7 The variation of the beam width of the NNLS method proposed in this invention with the number of iterations is shown in FIG. 1 . It can be found that the convergence speed of this method is faster than that of the CBF method. The beam width can be reduced to within 50 iterations. The method of the present invention has good engineering application value.

[0102] Example 3 The following is a further demonstration of the application effect of the present invention in combination with measured data: two adjacent spherical small targets with upward trajectory are selected to verify the imaging effects of different methods in three-dimensional space. The diameter of the spherical target is 13 cm, the distance between the centers of the two spheres is 20 cm, the working distance is 13 m, and the transducer array is a two-dimensional array structure with Receiving primitives. The conventional multi-beam synthetic aperture imaging algorithm and the shift-variant array deconvolution multi-beam synthetic aperture algorithm are used to process the sonar data, and the time azimuth diagram is intercepted to observe the imaging effect. Figure 8 The time-azimuth diagram of the conventional multi-beam synthetic aperture imaging method is shown. It can be found that the imaging result is severely expanded in the horizontal beam angle, and the beam angle reaches , which will cause its horizontal beam footprint to reach 80cm at a distance of 13m, and the target scale will be seriously distorted.

[0103] Fig. 9 The time-azimuth diagram of the shift-variable array deconvolution multi-beam synthetic aperture imaging method proposed in the present invention is shown. It can be found that the horizontal beam width is significantly reduced, the energy concentration is significantly improved, the noise background suppression ability is significantly enhanced, and the beam angle is reduced to , which proves that the method of the present invention can effectively reduce the beam width, improve the main-sidelobe ratio, and suppress image background noise.

[0104] The three-dimensional imaging results are further sliced ​​to show the imaging effects of different methods in the track direction and horizontal direction. Fig.10The slicing effect of multi-beam synthetic aperture imaging in the track direction is demonstrated. When the imaging algorithm of multi-beam bathymetric sonar is used, the dual targets cannot be distinguished. The conventional multi-beam synthetic aperture imaging algorithm can distinguish the dual targets in the track direction and has a good main-sidelobe ratio. The shift-variable array deconvolution multi-beam synthetic aperture imaging method can further reduce the sidelobe level in the track direction. However, due to its principle limitation, its beam width in the track direction has no obvious improvement compared with the conventional method, and can only suppress the background noise to a limited extent.

[0105] Fig.11 The slicing effect of multi-beam synthetic aperture sonar horizontal imaging is demonstrated. It can be found that the method proposed in the present invention can significantly improve the beam width and reduce the sidelobe level compared with the conventional multi-beam synthetic aperture method, and the sidelobe is more uniform, which is helpful to improve the background noise of three-dimensional imaging.

[0106] Fig.12 , 13 , 14 compare the thinning results of different imaging algorithms based on energy detection, which can more intuitively show the three-dimensional spatial position and scale of the target. Fig.12 It can be found that the 3D imaging thinning result using the multi-beam bathymetric sonar imaging algorithm cannot distinguish the positions of the two targets in the track direction, the target scales and numbers are aliased, the horizontal scale is severely widened, and the imaging result is much larger than the preset 13 cm diameter scale. Fig.13 It can be found that the energy thinning results of conventional multi-beam synthetic aperture 3D imaging show that the imaging after track-wise aperture synthesis processing can distinguish the location and number of preset targets, and the target scale is not greatly distorted. However, its horizontal imaging results are not significantly improved compared with multi-beam bathymetric sonar. Fig.14 It can be found that the energy thinning result after the shifted array deconvolution multi-beam synthetic aperture three-dimensional imaging processing not only maintains the imaging effect of aperture synthesis in the track direction and distinguishes the location and scale of the target, but also effectively reduces the beam width in the horizontal direction, thereby alleviating the problem of beam footprint expansion to a certain extent.

[0107] The above analysis and comparison prove that the method proposed in the present invention can break through the linear shift-invariant array limitation of the Richardson-Lucy deconvolution technology of multi-beam synthetic aperture sonar, and extend the application of multi-beam synthetic aperture deconvolution technology to any preset beam angle. While obtaining a constant track resolution, it effectively reduces the horizontal beam width, suppresses background noise, and improves the flexibility of application scenarios of multi-beam synthetic aperture sonar.

Claims

1. A shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method, comprising the following steps: Step (1): performing conventional beamforming processing on the array element signals received by the multi-beam synthetic aperture sonar in the horizontal direction to obtain multiple array conventional beamforming outputs; Step (2): performing aperture synthesis processing on the conventional beamforming outputs of the multiple arrays in the track upward direction to obtain conventional multi-beam synthetic aperture sonar imaging output; Step (3): According to the multi-beam synthetic aperture sonar array flow pattern and the preset beam angle, the point spread function matrix is ​​traversed and calculated, which is expressed as: Step (4): Establish a linear equation group of difference functions between conventional beamforming and the target sound source azimuth; iteratively estimate the target azimuth function according to the estimation criterion of minimum difference function, and obtain the deconvolution imaging result of the shift-variable array multi-beam synthetic aperture sonar; The establishment of the linear equations of the difference function between conventional beamforming and the target sound source orientation includes: (1) When the scanning target is When the position is The angle-dimensional data in the angular dimension is used to obtain the angle-dimensional conventional beamforming output matrix of the multi-beam synthetic aperture sonar image: ; (2) Construct a linear equation system of the difference function between the positive convolution output and the conventional beamforming output: ； in, is the shift array point spread function matrix; is the difference function, is the target position function to be estimated, is the 2-norm; To scan the target slant distance, is the track coordinate of the scanning point; It is the conventional multi-beam synthetic aperture sonar imaging output; is the preset beam angle.

2. The method for deconvolution imaging of a shift-variable array multi-beam synthetic aperture sonar according to claim 1, characterized in that: In step (1), the conventional beamforming output is traversed according to the preset beam angle and time to obtain the sequence of conventional beamforming outputs of all arrays. , expressed as: ； in, To traverse the target location; is the number of array elements of multi-beam synthetic aperture sonar, is the horizontal array element spacing, is the echo signal wavelength, For the moment.

3. The shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method according to claim 2, characterized in that: In step (2), conventional multi-beam synthetic aperture sonar imaging output It is expressed as: ； in, is the total number of array entries; It is a single array output; For the Horizontal receiving array.

4. The shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method according to claim 3, characterized in that: In step (3), the point spread functions of all target azimuth angles are traversed to obtain the shifted array point spread function matrix, which is expressed as: ； in, , … Indicates the 1st to the The location of the target, Preset the total number of beam angles for the imaging system; is the point spread function of the target location, with the superscript Transpose the matrix.

5. The shift-variable array multi-beam synthetic aperture sonar deconvolution imaging method according to any one of claims 1 to 4, characterized in that: In step (4), the iterative estimation of the target orientation function uses the fastest negative gradient projection method to iteratively search for the target sound source orientation to obtain the least squares solution of the target sound source orientation estimation.

Images