Underwater acoustic imaging method for realizing wavelength magnitude relative minimum resolution

Abstract

The invention relates to the technical field of underwater acoustic imaging, in particular to an underwater acoustic imaging method for achieving wavelength magnitude relative minimum resolution, a target scattering spatial spectrum is collected through a circumferential synthetic aperture test system, the traditional lambda / 4 resolution limit is broken through in combination with a super-resolution image processing technology, lambda is 17.64 mm (the center frequency is 85kHz), lambda / 4 is 4.4 mm, and the resolution ratio of the target scattering spatial spectrum is larger than that of the target scattering spatial spectrum. The minimum 3dB bandwidth of an image point after super-resolution processing reaches 0.1 mm and is improved by 54 times compared with that of a traditional method, and wavelength magnitude resolution is achieved through three-dimensional coordinate calibration (the error is smaller than 0.01 m), pulse compression (the distance precision is improved to 0.5 mm or below) and a two-dimensional deconvolution super-resolution algorithm (lambda / 4 = 4.4 mm limit is broken through after 10 times of iteration, and the final resolution reaches 0.1 mm). A 64-element high-frequency hydrophone vertical array and a directional sound source are used for constructing an imaging system, and through target rotation sampling, three-dimensional coordinate calibration, pulse compression and two-dimensional deconvolution super-resolution processing, the resolution superior to lambda / 4 is achieved. Test results show that the method can improve the resolution to 0.1 mm.

Description

Technical Field

[0001] The present invention relates to the field of underwater acoustic imaging technology, and specifically to an underwater acoustic imaging method for achieving a relatively minimum resolution at the wavelength level, so as to achieve high-precision target detection with a relatively minimum resolution at the wavelength level. Background Art

[0002] Underwater acoustic imaging technology has important application values in the fields of marine exploration, underwater target recognition, environmental monitoring, etc. As a core index, the imaging resolution directly determines the recognition ability of the system for tiny targets. The traditional underwater acoustic imaging theory holds that the theoretical physical limit of the minimum resolution of an imaging system is λ / 4 (λ is the acoustic wave wavelength), and this limit is jointly determined by the acoustic wave wavelength and the spatial spectrum coverage ability of the acquisition system. When the target size is less than λ / 4, it is difficult to distinguish with conventional imaging methods, resulting in limited detection ability for fine structures.

[0003] Existing underwater acoustic imaging technologies mainly improve the resolution by increasing the acoustic wave frequency (shortening the wavelength) or increasing the array aperture. However, high-frequency signals attenuate severely in water, limiting the detection range; large-aperture arrays face problems such as difficult installation and high costs. In addition, imaging algorithms such as traditional conventional beamforming (CBF) are limited by the incomplete acquisition of the spatial spectrum of scattered echoes, and the actual resolution often approaches λ / 4 but is difficult to break through. For example, the resolution of the conventional CBF method is only 5.4 mm at a center frequency of 85 kHz, approaching λ / 4 (4.4 mm) but not breaking through, and it is sensitive to multi-path interference and array attitude deviation, resulting in blurred target edges. With the development of super-resolution imaging theory, improving the resolution through signal processing algorithms has become a research hotspot, but existing methods have problems such as being sensitive to noise, high computational complexity, and poor three-dimensional imaging continuity. Especially in a complex multi-path environment, the effect of improving the resolution is limited.

[0004] Traditional methods cannot achieve full spatial spectrum acquisition (such as the angular sampling interval > 0.5°), resulting in the loss of high-frequency information. However, the present invention obtains the complete scattered spatial spectrum of the target through circumferential rotation sampling at an interval of 0.1°, providing sufficient data support for super-resolution processing. Aiming at the above problems, the present invention proposes an underwater acoustic imaging method combining circumferential synthetic aperture technology and two-dimensional deconvolution super-resolution algorithm. By constructing an experimental system with the ability to acquire the full spatial spectrum, cooperating with high-precision three-dimensional coordinate calibration and pulse compression preprocessing, the quality of the original data is effectively improved; the super-resolution algorithm is used to post-process the reconstructed image, breaking through the λ / 4 resolution limit and achieving fine imaging at the wavelength level. This method significantly improves the imaging performance through technological innovation while maintaining the existing hardware conditions, and has important practical application significance.

[0005] Traditional methods result in missing spatial spectra due to a large angular sampling interval (e.g., > 0.5°), while the present invention achieves full-spectrum acquisition through precise rotational sampling at 0.1°. Existing super-resolution algorithms are limited by vertical array sampling in three-dimensional imaging and have the problem of discontinuous z-axis energy. The present invention improves this defect through high-precision calibration and spectral energy retention strategies. Summary of the Invention

[0006] Provided is an underwater acoustic imaging method capable of achieving a relative minimum resolution better than λ / 4, which solves the problems of limited resolution and poor adaptability to complex environments in traditional technologies.

[0007] The technical solution adopted by the present invention to solve its technical problems is: an underwater acoustic imaging method for achieving a relative minimum resolution at the wavelength level, comprising the following steps:

[0008] Build a circumferential synthetic aperture test system, including a directional sound source, a 64-element high-frequency hydrophone vertical array, a rotatable target, and a stepping motor control unit. The directional sound source emits a chirp signal with a center frequency of 85 kHz and a bandwidth of 30 kHz. The element spacing of the vertical array is 7.5 cm, and the aperture is 4.8 m.

[0009] Perform three-dimensional coordinate calibration on the target: Receive the direct sound signal from the sound source through a standard hydrophone at the target to determine the relative position relationship between the sound source and the target, and obtain the three-dimensional absolute coordinates of the sound source. Transmit a signal from the sound source at the target, and collect the signal by the cross high-frequency listeners on the vertical array to calibrate the relative position between the array and the target, and obtain the three-dimensional absolute coordinates of the array. Use the vertical array to collect the direct sound signal of the target sound source, calculate the time-delay residuals of each channel and perform compensation. The three-dimensional coordinate calibration error is controlled within 0.01 m. Measure the time difference of the signals received by each element through the direct wave signal emitted by the target sound source, and calculate the deviation between the actual distance and the theoretical distance in combination with the sound speed of 1480 m / s to obtain the time-delay residual Δt_i = Δd_i / c.

[0010] Control the stepping motor to drive the target to rotate around the vertical axis, sample once every 0.1°, cover the 360° full-angle backscattered echo acquisition, synchronously collect the backscattered echo signals at different angles, and form the original imaging data.

[0011] Preprocess the original data: including secondary cropping, splicing to form a full-angle long file, performing shift correction based on the direct wave signal, and using a pulse compression algorithm to improve the range accuracy. The pulse compression algorithm is realized by performing a correlation operation between the original signal and an ideal chirp signal.

[0012] Perform image reconstruction: Input the correction parameters and the preprocessed data, and reconstruct the target image through a tomographic sonar imaging algorithm to obtain a three-dimensional imaging result including the target feature size.

[0013] The reconstructed image is processed using a two-dimensional deconvolution super-resolution algorithm. The sidelobes are suppressed through iterative operations, and the 3dB bandwidth of the image point is reduced. When the 3dB bandwidth of the image point < 4.4mm, a relative minimum resolution better than λ / 4 is finally achieved.

[0014] Specifically, the signal parameters of the directional sound source include: sound source level 190dB, pulse width 1.5ms (lake trial) or 0.2ms (pool trial), signal period 100ms or 50ms, and sampling rate 2MHz.

[0015] Specifically, the targets include a square cage target and an elliptical ring target. The square cage target is used for resolution detection and target recognition rate verification, and the elliptical ring target is used for ultimate resolution testing, with a diameter of 1.5cm.

[0016] Specifically, during the three-dimensional coordinate calibration process, the exact positions of the sound source and the array are determined through phase comparison and phase residual analysis, with the error controlled within 0.01m.

[0017] Specifically, the pulse compression algorithm improves the range direction accuracy to below 0.5mm by performing a correlation operation between the original signal and an ideal linear frequency modulation signal.

[0018] Specifically, the number of iterations for the two-dimensional deconvolution iterative operation of the beam-delay diagram based on the system point spread function (PSF) in the two-dimensional deconvolution super-resolution algorithm is 10 - 90 times. When the number of iterations ≥ 10 times, the 3dB bandwidth of the image point ≤ 4.2mm (<λ / 4 = 4.4mm), and finally reaches 0.1mm.

[0019] Specifically, during the image reconstruction process, the central position, unit length, and threshold of the reconstruction area are set to generate a 3D imaging map including the x-y and y-z planes, clearly showing the target feature size.

[0020] Specifically, the test environments include Huangcai Reservoir with a water depth of 16m (lake trial) and Dingjialing Laboratory pool with specific dimensions (pool trial), and the measured sound speed value is 1480m / s.

[0021] Specifically, the top of the vertical array is 0.6m from the water surface, and the center is 3m from the water surface. Attitude correction is performed through 4 cross-shaped high-frequency receivers, and the equivalent noise pressure is less than 40dB.

[0022] Specifically, the imaging result after super-resolution processing determines whether the resolution index is achieved by comparing the numerical relationship between the -3dB width of the image point and λ / 4 (4.4mm), where λ is the acoustic wavelength, calculated from the center frequency of 85kHz, λ = 17.64mm, calculated according to the sound speed of 1480m / s and the formula λ = c / f, where c is the sound speed and f is the center frequency.

[0023] Beneficial effects of the present invention:

[0024] 1. Breaking through the resolution limit: By combining circular synthetic aperture full spatial spectrum acquisition with super-resolution algorithm, a resolution better than λ / 4 (minimum 0.1mm) is achieved, which is 54 times higher than the traditional CBF method. The ability to identify tiny targets is significantly enhanced, and steel columns with a spacing of 5cm (design value 0.05m, measured 0.0519m) and elliptical rings with a diameter of 1.5cm (measured 2.05cm) can be clearly distinguished.

[0025] 2. High-precision calibration technology: The three-dimensional coordinate calibration error is less than 0.01m, and the delay residual compensation eliminates the influence of array posture and position deviation to ensure the geometric accuracy of the reconstructed image. The measured value of the square cage side length is 0.1866m (design value 0.2m), and the measurement error is 3.3%.

[0026] 3. Multi-environment adaptability: It is suitable for different scenarios such as lakes and laboratory pools. By adjusting the signal pulse width (1.5ms / 0.2ms) and sampling rate (2MHz), it takes into account both far-field detection and near-field fine imaging needs.

[0027] 4. Efficient data processing: The pulse compression algorithm improves the signal-to-noise ratio by more than 20dB and shortens the preprocessing time by 50%. The super-resolution algorithm has a fast iterative convergence speed, and the sidelobe suppression ratio is increased from 10dB to 40dB, effectively suppressing noise interference (see Table 1). It can break through the λ / 4 limit in 10 iterations to meet the needs of real-time imaging.

[0028] 5. Three-dimensional imaging capability: Generates 3D images including xy and yz planes, clearly displays the target spatial structure, the measured diameter of the elliptical ring is 2.05cm (design value 1.5cm), and the geometric feature restoration degree is improved by 43%, providing support for complex target morphology analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0030] Figure 1 Schematic diagram of the test configuration on the lake for proof of principle (linear array tomography);

[0031] This figure shows the system layout of the Huangcai Reservoir lake trial, including the test platform, horn sound source, 64 - element high - frequency hydrophone vertical array, and rotatable target. The sound source and the array are fixed at the same underwater depth (3 m from the water surface), the target is about 7.1 m away from the array, and it rotates around the vertical axis controlled by a stepper motor (sampling once every 0.1°). The position parameters of key equipment are marked in the figure: the sound source is 3.6 m from the water surface, the top of the array is 0.6 m from the water surface, and the array aperture is 4.8 m (64 elements, 7.5 cm spacing). Calibration point sound source (Rst) and target area calibration listener (Rta) are installed at the target end to accurately calibrate the relative position of the sound source and the array, ensure the full - space spectrum acquisition ability, and lay the foundation for subsequent resolution tests.

[0032] Figure 2 This is a real - shot picture of the square cage of the target;

[0033] This figure shows a square - cage target for resolution detection. The number and spacing of the four - side columns are different (designed spacing 0.05 m, side length 0.2 m), which can verify the resolution ability of the imaging system for target details. The surface gap structure of the target is used as a feature point to evaluate whether the imaging method can accurately restore the target geometric size. The real - shot picture shows that the target is composed of a metal frame, and a rotating shaft is installed at the bottom and connected to a stepper motor, supporting 360° rotation sampling, and cooperating with the calibration sound source to achieve multi - angle scattered echo acquisition.

[0034] Figure 3 This is a schematic diagram of the test environment and test system scheme (Dingjialing Laboratory pool);

[0035] This figure shows the test configuration of the laboratory pool for the limit resolution test of the elliptical - ring target. The directional sound source is 1 m from the water surface, and the emission parameters are the same as those in the lake trial (center frequency 85 kHz, bandwidth 30 kHz). The target (elliptical ring, diameter 1.5 cm) is fixed to a rotating motor, 0.75 m away from the array (at the same depth as the array center). The size of the water tank is not fully shown, but the core layout is similar to that of the lake trial. By reducing the test distance, the near - field imaging accuracy is improved to verify the processing ability of the super - resolution algorithm for small targets.

[0036] Figure 4 This is a real - shot picture of the water - tank control device and the elliptical - ring target;

[0037] In this figure, the left picture shows the rotation control device for the pool test. The accuracy of the stepper motor reaches 0.1°, ensuring the accuracy of target angle sampling; the right picture is a close - up of the elliptical - ring target, a metal - ring structure with a diameter of 1.5 cm, used to test the imaging ability of the system for curved targets. The real - shot picture shows that the target is installed on a transparent bracket to reduce hydrodynamic interference, and cooperates with the high - frequency hydrophone array to collect high - signal - to - noise - ratio echo signals, providing high - quality data sources for super - resolution processing.

[0038] Figure 5For the sound source calibration results (phase comparison and phase residuals).

[0039] Figure 6 Schematic diagram of the sound source calibration position.

[0040] Figure 7 For the array calibration results (phase comparison and phase residuals).

[0041] Figure 8 Schematic diagram of the array calibration position.

[0042] Figure 9 Schematic diagram of the time delay residual calibration;

[0043] This figure shows the theoretical time delay and actual time delay residual distribution of each channel of the vertical array. By collecting the direct signal of the target sound source, calculating the time delay deviation of each array element (Δt = actual time delay - theoretical time delay), the influence of position residuals is eliminated through the phase compensation formula during reconstruction, improving the imaging focusing accuracy and avoiding image blurring caused by array attitude deviation.

[0044] Figure 10 For the sine diagram after shift correction;

[0045] The horizontal axis of this figure is the rotation angle of the target (at 0.1° intervals), the vertical axis is the channel number, and the bright area is the direct wave signal. After shift correction, the signal positions are aligned, eliminating the time axis misalignment caused by array tilt.

[0046] Figure 11 For the comparison between pulse compression data and original data;

[0047] This figure improves the range resolution from 24.67 mm (theoretical value) to 5.4 mm (measured CBF result) through correlation operation with the ideal LFM signal, effectively suppressing noise and providing high-contrast data for subsequent super-resolution processing.

[0048] Figure 12 For the 3D reconstruction diagram.

[0049] Figure 13 For the 3D image (angle 1).

[0050] Figure 14 For the 3D image (angle 2).

[0051] Figure 15 For the 3D image (x-y plane).

[0052] Figure 16 For the 3D image (y-z plane).

[0053] Figure 17 For the characteristic size schematic.

[0054] Figures 18 - 19Schematic diagram of elliptical ring imaging and characteristic dimensions;

[0055] The width of the elliptical ring in this figure is measured to be 0.1228 m (design value 0.14 m), and the diameter is 2.05 cm (design value 1.5 cm), showing the geometric restoration ability of the system for curved targets and providing support for the recognition of complex targets.

[0056] Figure 20 Imaging result of super-resolution algorithm;

[0057] In this figure, after 10 iterations, the imaging of the CBF and dCv algorithms is compared. The edges of adjacent steel columns in the super-resolution image (dCv) on the right are sharper. The 3dB bandwidth is reduced from 5.4 mm to 4.2 mm, breaking through the λ / 4 (4.4 mm) limit for the first time. The 3dB bandwidth is defined as the horizontal width when the peak intensity drops to 70.7%. The CBF method is 5.4 mm, and the dCv algorithm is reduced to 4.2 mm after 10 iterations.

[0058] Figure 21 Comparison of imaging results of super-resolution algorithm and 3dB bandwidth;

[0059] In this figure, the ultimate resolution result after 90 iterations is shown. The 3dB bandwidth of the image points reaches 0.1 mm, and the sidelobes are basically eliminated, verifying the high-efficiency processing ability of the two-dimensional deconvolution algorithm for continuous spectra. Although the continuity is slightly affected by the energy fluctuation along the z-axis, the lateral resolution is significantly improved, providing the possibility for underwater sub-millimeter detection. Specific implementation manners

[0060] In order to make the technical means, creative features, achieved purposes and effects of the present invention easy to understand, the present invention will be further described below in conjunction with specific implementation manners.

[0061] As Figures 1 - 21 shown, a method for underwater acoustic imaging to achieve the relatively minimum resolution of wavelength order according to the present invention includes the following core steps:

[0062] Build a test system, and construct an imaging system using a directional sound source with a center frequency of 85 kHz (emitting high-frequency linear frequency modulation signals) and a 64-element high-frequency hydrophone vertical array. The sound source and the array are fixed on the test platform, and the target is controlled by a stepping motor to rotate around the vertical axis to achieve full-angle sampling at intervals of 0.1°. In the lake test, the target is 7.1 m away from the array, and in the pool test, it is 0.75 m away, ensuring that the scattered spatial spectrum of the target is covered.

[0063] Three-dimensional coordinate calibration: Sound source calibration. The direct sound signal from the sound source is received by the standard hydrophone at the target, and the signal amplitude and phase are analyzed to determine the three-dimensional coordinates of the sound source, with the error controlled within 0.01 m. Array calibration. The signal is emitted by the target sound source, and the data is collected by the cross hydrophones on the array. The relative position between the array and the target is calculated to obtain the three-dimensional absolute coordinates of the array. Time-delay residual compensation. The direct sound signal from the target sound source is collected, the residual between the actual time delay and the theoretical time delay of each channel is calculated, and phase correction is performed during reconstruction to improve the positioning accuracy.

[0064] Data acquisition and preprocessing. During the rotation of the target, the echo signal is synchronously collected to form a raw data file containing 64 channels and full angles. Through secondary cropping, splicing, and shift correction, the influence of multi-path interference is eliminated; the pulse compression algorithm (correlation operation) is used to improve the range resolution, so that the range accuracy reaches below 0.5 mm.

[0065] Image reconstruction. Based on the tomographic sonar imaging algorithm, the calibrated parameters and preprocessed data are input, the reconstruction area (such as the x-y plane, y-z plane) is set, and the three-dimensional image is generated through the frequency-domain backprojection or time-domain focusing algorithm to clearly display the target feature size (such as the side length of the square cage, the diameter of the elliptical ring).

[0066] Super-resolution processing. The two-dimensional deconvolution (dCv) algorithm is used to iteratively process the reconstructed image to suppress the sidelobe noise. As the number of iterations increases (such as 10 - 90 times), the 3 dB bandwidth of the image point gradually shrinks from 5.4 mm (CBF result) to 0.1 mm, breaking through the λ / 4 (4.4 mm) limit and achieving wavelength-level resolution.

[0067] The test results of the resolution improvement by the two-dimensional deconvolution imaging algorithm are shown in Table 1 below:

[0068]

[0069]

[0070] As can be seen from the above table, the resolution of the focused beamforming (CBF) obtained by the tomographic method is 5.4 mm, close to the quarter-wavelength limit resolution size (the central frequency of the imaging system is 85 kHz, the corresponding wavelength is 17.64 mm, and λ / 4 is about 4.4 mm). Through the super-resolution method - two-dimensional deconvolution (dCv), the resolution can be improved, and the 3 dB bandwidth of the image point continuously shrinks as the number of iterations increases. The relationship between the number of iterations and the resolution is shown in Table 1. When the number of iterations ≥ 10, the resolution breaks through λ / 4, and reaches 0.1 mm after 90 iterations. It can be seen that after 10 iterations, the resolution obtained by the super-resolution processing is 4.2 mm, better than the quarter-wavelength limit resolution size (λ / 4 is about 4.4 mm). The comparison diagram of the results after 10 iterations is attached Figure 20(Results of the CBF method on the left and results of the dCv method on the right).

[0071] Example 1: Setup and parameter configuration of the test system

[0072] Hardware deployment

[0073] Lake test environment: Huangcai Reservoir, water depth 16 m, sound speed 1480 m / s. The directional sound source is fixed 3.6 m underwater and can rotate horizontally. It emits an LFM signal with a center frequency of 85 kHz and a bandwidth of 30 kHz (source level 190 dB, pulse width 1.5 ms, period 100 ms).

[0074] Vertical array: 64 - element high - frequency hydrophone, element spacing 7.5 cm, aperture 4.8 m, the top is 0.6 m from the water surface, and the center is 3 m from the water surface. The array integrates 4 cross - shaped high - frequency receivers for attitude correction, and the equivalent noise pressure < 40 dB.

[0075] Target: Square cage target, designed side length 0.2 m, steel column spacing 0.05 m. Install a calibration point sound source and a target - area listening device, and it can rotate around the vertical axis controlled by a stepper motor, with a sampling interval of 0.1°.

[0076] Pool test configuration

[0077] Dingjialing Laboratory pool, with dimensions adapted to the elliptical - ring target (diameter 1.5 cm). The sound source is 1 m from the water surface, and the emission parameters are adjusted to a pulse width of 0.2 ms, a period of 50 ms, and a sampling rate of 2 MHz. The target is 0.75 m from the array and at the same depth as the array center.

[0078] System synchronization

[0079] The acoustic emission and reception are synchronized using a 48M clock source to ensure signal phase consistency; the stepper motor and the data acquisition card (NI) are trigger - synchronized to achieve a strict correspondence between the angle and the signal.

[0080] Example 2: Three - dimensional coordinate calibration method

[0081] Absolute coordinate calibration of the sound source

[0082] Fix the target and rotate the sound source. Monitor the signal amplitude through a standard hydrophone and fix the sound source angle at the point where the amplitude is maximum (the most directive direction).

[0083] Move the target up and down until the signal amplitude is maximized again to determine the target depth.

[0084] Collect direct - wave signals at different azimuth angles. Calculate the sound - source coordinates (x, y, z) through phase comparison (such as Figure 5 ), and the error is corrected through phase - residual analysis. The final coordinate accuracy is ±0.01 m.

[0085] Absolute coordinate calibration of the array

[0086] The target sound source emits an LFM signal, and the cross hydrophone (4 channels) collects the signal. The relative distance between the array and the target is calculated by the time difference of arrival (TDOA).

[0087] Establish a coordinate system with the center of the target as the origin, and deduce the coordinates of each array element (such as Figure 8 ), and combine the attitude correction data to generate a three-dimensional coordinate array.

[0088] Time delay residual compensation

[0089] Collect the direct signal of the target sound source, and calculate the difference between the theoretical time delay (t = distance / c) of each array element and the actual time delay (such as Figure 9 ).

[0090] When reconstructing, perform time delay compensation on the signals of each channel. The formula is: 8 补偿 \(s_i(t)=s_i\) 原始 (t + Δt i )

[0091] Where Δt_i is the difference between the actual time delay and the theoretical time delay of the i-th channel, which is calculated by the time difference of arrival (TDOA) of the direct wave signal and is used to correct the phase error caused by the array attitude deviation.

[0092] Example 3: Data acquisition and preprocessing process

[0093] Original data acquisition

[0094] The motor drives the target to rotate 360°, and triggers the acquisition once every 0.1°, generating a.mat file with 64 channels (naming rule: CF-1_4_probe number_angle.mat).

[0095] The single acquisition time for the lake test is about 10 minutes, obtaining 3600 angle samples; due to the short distance in the pool test, the sampling time is shortened to 2 minutes.

[0096] Secondary trimming and splicing

[0097] The host computer initially trims to remove invalid data (such as the transient signal at motor startup), and retains a 10s effective data segment.

[0098] Write a program to splice the data of each angle according to the channel to form a long file containing all angles, which is convenient for subsequent generation of the sinogram.

[0099] Shift correction and pulse compression

[0100] Based on the position of the direct wave signal, perform a horizontal shift on the sinogram to eliminate the time axis misalignment caused by the array attitude deviation (such as Figure 10 ).

[0101] Pulse compression formula: 8 pc \(s(t)=8\) 原始 \(s(t)*h\) * \(s(-t)\)

[0102] where \(h(t)\) is the ideal LFM signal. After correlation operation, the range resolution is improved to \(c / (2B)=1480 / (2\times30\times10\) 3 ) = 24.67 mm, and the actual test reaches 5.4 mm (affected by noise).

[0103] Example 4: Image reconstruction and feature extraction

[0104] Tomographic reconstruction parameter settings

[0105] Input the calibrated sound source coordinates, array coordinates, and time delay compensation parameters, and set the reconstruction area: the x range is [-0.3 m, 0.3 m], the y range is [-8 m, -6 m], and the z range is [-4.5 m, -3.5 m], with a unit length of 0.001 m.

[0106] Adopt the frequency domain back projection algorithm to map the echo signals at each angle to the reconstruction grid, and accumulate the energy values to generate a grayscale image.

[0107] Three-dimensional image generation

[0108] Generate the x-y plane projection diagram (such as Figure 15 ), and display the target's lateral dimension; the y-z plane shows the depth distribution (such as Figure 16 ).

[0109] Optimize the display threshold to highlight the target edge. The measured distance between the square cage steel columns is 0.0519 m (the designed value is 0.05 m), with an error of 3.8%.

[0110] Feature size calculation

[0111] Extract the orthogonal profile intensity curve of the peak points, and calculate the -3dB bandwidth (full width at half maximum). The formula is: BW 3dB =\(|x2 - x1|\)

[0112] where \(x1\) and \(x2\) are the coordinates at 0.707 times the peak intensity. The traditional CBF method measures 5.4 mm, and it gradually decreases after super-resolution processing.

[0113] Example 5: Super-resolution processing and resolution verification

[0114] Implementation of two-dimensional deconvolution algorithm

[0115] Define the system point spread function (PSF), and construct the PSF matrix based on the array response model.

[0116] Iterative formula:

[0117] where f is the image estimate, g is the observed data, and h is the PSF, represents convolution, · / represents point division, and sidelobe noise is suppressed through iteration.

[0118] Resolution improvement effect

[0119] After 10 iterations, the resolution drops from 5.4 mm to 4.2 mm (better than λ / 4 = 4.4 mm), as shown in Table 1 of the resolution improvement test results of the two-dimensional deconvolution imaging algorithm; it reaches 0.1 mm after 90 iterations, and the sidelobe suppression ratio is increased by 30 dB.

[0120] Comparing the CBF and dCv imaging results (as shown in Figure 20 ), the target edge is sharper after super-resolution processing, and adjacent steel columns can be clearly separated.

[0121] Index determination

[0122] Calculate the ratio of the 3 dB bandwidth of the actual image point to λ / 4. When the bandwidth < 4.4 mm, it is determined to meet the standard. Both the lake trial and the pool test pass this test, verifying the effectiveness of the method.

[0123] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. An underwater acoustic imaging method for achieving a relatively minimum resolution at the wavelength level, characterized in that It includes the following steps: Build a circumferential synthetic aperture test system, including a directional sound source, a 64-element high-frequency hydrophone vertical array, a rotatable target, and a stepping motor control unit. The directional sound source emits a chirp signal with a center frequency of 85 kHz and a bandwidth of 30 kHz. The element spacing of the vertical array is 7.5 cm, and the aperture is 4.8 m. Perform three-dimensional coordinate calibration on the target: Receive the direct sound signal of the sound source through the standard hydrophone at the target to determine the relative position relationship between the sound source and the target, and obtain the three-dimensional absolute coordinates of the sound source; Transmit a signal from the sound source at the target, and collect the signal by the cross high-frequency hydrophone on the vertical array to calibrate the relative position between the array and the target, and obtain the three-dimensional absolute coordinates of the array; Use the vertical array to collect the direct sound signal of the target sound source, calculate the time delay residuals of each channel and perform compensation. The three-dimensional coordinate calibration error is controlled within 0.01 m. Control the stepping motor to drive the target to rotate around the vertical axis, sample once every 0.1°, cover the 360° full-angle scattering echo acquisition, synchronously collect the backward sound scattering echo signals at different angles, and form the original imaging data. Preprocess the original data: including secondary clipping, splicing to form a full-angle long file, performing shift correction based on the direct wave signal, and using a pulse compression algorithm to improve the range accuracy. The pulse compression algorithm is achieved by performing a correlation operation between the original signal and the ideal chirp signal. Perform image reconstruction: Input the calibration parameters and the preprocessed data, and reconstruct the target image through the tomographic sonar imaging algorithm to obtain a three-dimensional imaging result containing the target feature size. Process the reconstructed image using a two-dimensional deconvolution super-resolution algorithm, suppress the sidelobes through iterative operations, reduce the 3 dB bandwidth of the image point. When the 3 dB bandwidth of the image point < 4.4 mm, finally achieve a relative minimum resolution better than λ / 4.

2. The underwater acoustic imaging method for achieving the relatively minimum resolution of the wavelength order according to claim 1, characterized in that: The signal parameters of the directional sound source include: source level 190 dB, lake test pulse width 1.5 ms, pool test pulse width 0.2 ms, signal period 100 ms or 50 ms, sampling rate 2 MHz.

3. An underwater acoustic imaging method for achieving a relatively minimum resolution at the wavelength level according to claim 1, characterized in that: The target includes a square cage target and an elliptical ring target. The square cage target is used for resolution detection and target recognition rate verification, and the elliptical ring target is used for ultimate resolution testing, with a diameter of 1.5 cm.

4. An underwater acoustic imaging method for achieving a relatively minimum resolution at the wavelength level according to claim 1, characterized in that: During the three-dimensional coordinate calibration process, determine the exact positions of the sound source and the array through phase comparison and phase residual analysis, and the error is controlled within 0.01 m.

5. An underwater acoustic imaging method for achieving a relatively minimum resolution of the wavelength order according to claim 1, characterized in that: The pulse compression algorithm improves the range accuracy to below 0.5 mm by performing a correlation operation between the original signal and the ideal chirp signal.

6. The underwater acoustic imaging method for achieving the relatively minimum resolution of wavelength order according to claim 1, wherein: The number of iterations of the two-dimensional deconvolution super-resolution algorithm for performing two-dimensional deconvolution iterative operations on the beam-delay diagram based on the system point spread function is 10 - 90 times. When the number of iterations ≥ 10 times, the 3 dB bandwidth of the image point ≤ 4.2 mm, and finally reaches 0.1 mm.

7. An underwater acoustic imaging method for achieving a relatively minimum resolution at the wavelength level according to claim 1, characterized in that: During the image reconstruction process, set the center position, unit length, and threshold of the reconstruction area to generate a 3D imaging map including the x-y and y-z planes, clearly showing the target feature size.

8. An underwater acoustic imaging method for achieving a relatively minimum resolution at the wavelength level according to claim 1, characterized in that: The test environment includes a lake test with a water depth of 16 m and a pool test with specific specifications. The measured sound speed value is 1480 m / s.

9. An underwater acoustic imaging method for achieving a relatively minimum resolution at the wavelength level according to claim 1, characterized in that: The top of the vertical array is 0.6 m from the water surface, and the center is 3 m from the water surface. Attitude correction is performed through 4 cross-shaped high-frequency receivers, and the equivalent noise pressure is less than 40 dB.

10. A method for underwater acoustic imaging to achieve a relatively minimum resolution of the wavelength order according to claim 1, characterized in that: The imaging result after super-resolution processing determines whether the resolution index is achieved by comparing the numerical relationship between the -3 dB width of the image point and λ / 4, where λ is the acoustic wavelength, which is calculated from the center frequency of 85 kHz, λ = 17.64 mm, and is calculated according to the sound speed of 1480 m / s and the formula λ = c / f, where c is the sound speed and f is the center frequency.

Images