A high frequency signal processing method of a submarine pipeline SAS monitoring system

Abstract

This invention provides a high-frequency signal processing method for a submarine pipeline SAS monitoring system, relating to the field of signal processing. This method acquires the original echo signal and transforms it to the range frequency domain. It then uses the phase characteristics of the submarine pipeline target in the echo to invert and correct the sound velocity parameters of each layer in a layered medium model, establishing a sound velocity recursive relationship. The method calculates the one-way propagation time corresponding to different range frequencies, obtains the equivalent frequency modulation parameter through numerical difference, and constructs a phase correction value in the range frequency domain to compensate for the nonlinear distortion of the range migration curve caused by sound velocity layering. After range compression, it calculates the actual propagation time sequence at different azimuth times based on the time delay-frequency relationship, and constructs correction coefficients through azimuth frequency domain group delay analysis to compensate for azimuth phase distortion. This invention achieves physical matching compensation for signal distortion caused by layered media, significantly improving the focusing performance and geometric fidelity of submarine pipeline SAS images.

Description

Technical Field

[0001] This invention relates to the field of signal processing, specifically to a high-frequency signal processing method for a submarine pipeline SAS monitoring system. Background Technology

[0002] SAS technology, with its high-resolution imaging capabilities, has been widely used for long-term monitoring and condition assessment of subsea pipelines. By transmitting broadband high-frequency linear frequency modulated signals and coherently processing the echoes, acoustic images of the pipeline and its surrounding environment can be acquired, thereby effectively identifying key information such as the pipeline's suspended, deformed, and buried status.

[0003] However, in real-world marine environments, seawater temperature, salinity, and pressure vary with depth, resulting in significant vertical stratification of sound velocity. This stratified medium causes the sound wave propagation path to bend, especially for high-frequency signals. Different frequency components refract to varying degrees when crossing the interfaces of different water layers, causing a complex nonlinear dependence between the propagation time and frequency of the echo signal. Existing imaging algorithms based on the assumption of a homogeneous medium cannot effectively compensate for this frequency-dependent time delay distortion, thus disrupting the phase matching condition of range-direction matched filtering, causing nonlinear distortion of the range migration curve, and introducing additional errors in the azimuth phase history. Ultimately, this leads to severe imaging quality problems in the pipeline images after synthetic aperture processing, greatly affecting the reliability and accuracy of subsea pipeline monitoring. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: a high-frequency signal processing method for a submarine pipeline SAS monitoring system, comprising:

[0005] S1: Acquire the raw echo signal collected by the synthetic aperture sonar system in the seabed pipeline route area;

[0006] S2: Perform a Fourier transform on the original echo signal along the range direction to transform it to the range frequency domain and the azimuth time domain to obtain the range frequency domain signal;

[0007] S3: Based on the acquired sound velocity profile data of the subsea pipeline route area, a layered medium model containing multiple water layers with constant sound velocities in each layer is established. The one-way propagation time of sound waves from the towed body to the subsea pipeline corresponding to different distance frequencies is calculated for the layered medium model, thus establishing a numerical correspondence between time delay and frequency. Based on this numerical correspondence, an equivalent frequency modulation parameter varying with distance frequency is obtained using the numerical difference method. This parameter characterizes the comprehensive influence of the layered medium on the frequency modulation of the linear frequency modulation signal. Based on this equivalent frequency modulation parameter, a phase correction value is constructed in the distance frequency domain. This phase correction value is used to compensate for the nonlinear distortion of the distance migration curve caused by sound velocity layering. This phase correction value is combined with the distance frequency domain signal to obtain the corrected distance frequency domain signal.

[0008] S4: Perform an inverse Fourier transform on the corrected range frequency domain signal along the range direction to obtain the range migration corrected time domain signal;

[0009] S5: Perform range-direction matched filtering on the time-domain signal after range migration correction to achieve range compression and obtain the range-compressed signal;

[0010] S6: Perform a Fourier transform on the range-compressed signal along the azimuth direction to transform it to the range time domain and azimuth frequency domain, and perform azimuth compression in these domains. Then perform an inverse Fourier transform on the azimuth-compressed signal to output the SAS image of the subsea pipeline.

[0011] As a further technical solution, the establishment of the layered medium model is achieved by inverting and correcting the model parameters through phase characteristic analysis of the echo signal itself. Specifically, complex data of the distance cell where the subsea pipeline is located is extracted from the original echo signal, and the position with the strongest pipeline echo energy along the azimuth direction is selected as the reference azimuth. At the reference azimuth, the extracted distance cell data is Fourier transformed along the distance direction, and its amplitude spectrum and phase spectrum in the distance frequency domain are extracted. Taking the frequency corresponding to the peak position in the amplitude spectrum as the reference frequency, the phase value of the reference frequency in the phase spectrum is set to zero, thereby obtaining the relative phase spectrum after eliminating the fixed phase bias. At this point, the water column is divided into N layers according to depth, with the interface depth z of each layer as the dividing line. i (i=1,…,N-1) are used as model nodes, and the initial value of the sound velocity in each layer is set to the measured value c of the depth corresponding to the sound velocity profile data. i (0); The relative phase spectrum is represented as the sum of the phase jump values ​​caused by the sound velocity jumps at each interface, that is, for frequency fr, we have ,in Let be the instantaneous phase increment caused by the change in sound velocity when the sound wave passes through the i-th interface; under the condition that the sound wave is approximately perpendicularly incident or at a small grazing angle, this instantaneous phase increment is approximately expressed as: , where c i and c i+1 These are the sound velocity measurements for the i-th and (i+1)-th layers, respectively; to eliminate the influence of grazing angle variations, the relative phase spectrum difference between adjacent frequency points is used as the observed value: ,in The k-th distance frequency f represents r,k The relative phase spectrum at that location is as follows:

[0012] This difference is approximately equal to the linear superposition of the sound velocity jump values ​​at each interface within the corresponding frequency range; assuming the frequency sampling interval is uniformly Δf. r ,but

[0013] ;make To represent the unknowns related to the abrupt change in sound speed, we establish a system of linear equations Ax=b, where the elements of matrix A are... Elements of vector b Each equation in this system corresponds to a pair of adjacent frequency points, and each equation includes the contribution of the sound velocity jump values ​​of all layer interfaces.

[0014] For the above system of linear equations, the correlation quantity x of the sound velocity jump at each interface is obtained by solving the singular value decomposition method. i This allows us to obtain the corrected values ​​for the sound velocities at each layer; singular value decomposition can effectively handle ill-conditioned problems that may exist in the equation system, ensuring the numerical stability of the solution; the obtained x i By summing the values ​​in order of depth, the correction values ​​for the sound velocity at each layer can be obtained: Where c1 is the surface sound velocity, which can be obtained directly from sound velocity profile data or set separately to update the sound velocity parameters of each water layer; j is the summation quantity; the updated sound velocity parameters of each layer are arranged in depth order to form a layer-by-layer recursive sound velocity sequence c i In this sequence, the sound velocity value of each layer is obtained by adding the sound velocity value of the previous layer to the sound velocity jump value of the corresponding interface, that is... A recursive relationship for the sound velocity between layers is established; this recursive relationship is directly embedded in the layered medium model and used to calculate the one-way propagation time corresponding to different distance frequencies.

[0015] As a further technical solution, the process of calculating the one-way propagation time corresponding to different distance frequencies using the layered medium model is based on ray acoustics theory and combined with an established recursive sound velocity sequence; the specific steps are as follows: First, the corrected sound velocity sequence c is extracted from the layered medium model. i and the corresponding thickness Δz of each layer i (i=1,…,N); The ray parameter p is used as the variable to be determined. This parameter remains constant along the entire ray path and is defined as p=cosθ. i / c i , where θ i Let θ be the grazing angle of the sound wave in the i-th layer (the angle with the horizontal plane); during calculation, starting from the depth of the towed body, each water layer is traversed sequentially from top to bottom; for a given p, the grazing angle θ in the i-th layer can be calculated. i =arccos(pc i This allows us to obtain the horizontal displacement increment of the sound wave propagation within that layer. and the increase in transmission time

[0016] The total horizontal displacement X(p) = ∑Δx is obtained by summing the horizontal displacement increments of each floor. iThe total propagation time T(p) is obtained by summing the propagation time increments of each layer. i For each distance frequency f r Based on the parameters of the linear frequency modulated signal transmitted by the SAS system, f r Convert to corresponding straight-line distance Where c0 is the reference speed of sound, B is the signal bandwidth, and f c The carrier frequency is used; then, the target's horizontal displacement is calculated by combining the difference ΔH between the towed body depth and the subsea pipeline depth. Solve the ray parameter equation X(p)=R horiz The numerical solution of p is obtained by using the bisection method or Newton's iteration method, and T(p) at this point is taken as the distance frequency f. r The corresponding one-way propagation time τ(f) r By traversing all distance frequency sampling points, a precise correspondence between time delay and frequency value τ(f) can be formed. r ).

[0017] As a further technical solution, the process of obtaining the equivalent frequency modulation parameter through the numerical difference method aims to quantify the comprehensive distortion effect of sound velocity layering on signal frequency modulation; firstly, the layer-by-layer recursive sound velocity sequence c is extracted from the layered medium model. i The ratio of sound velocities in adjacent layers As the refractive intensity coefficient of the i-th interface; in the established time-delay-frequency correspondence τ(f r In ), for each distance frequency point f r,k According to the ray parameter p(f) corresponding to this frequency r,k Determine the penetration depth of the principal energy of the sound wave in the layered medium, i.e., the deepest layer reached by the ray, and then extract a corresponding segment of coefficients from the refractive intensity coefficient sequence (all r values ​​within the range from the surface to that depth). i The cumulative refractive intensity factor is obtained by multiplying the refractive intensity coefficients of this segment. Where M is the number of penetration layers; this factor characterizes the overall distortion of the acoustic path by all water layer interfaces from the tow body to the penetration depth; on the other hand, τ(f) is calculated using the three-point difference method. r ) in f r,k Initial second-order difference value at:

[0018] ; The cumulative refractive intensity factor R k As a weight, it is weighted and fused with the initial second-order difference value to obtain the corrected second-order difference value. The corrected second-order difference value is the equivalent frequency modulation parameter γ at that frequency point. eff (f r,k The computational basis of ), namely It participates in the construction of subsequent phase correction values.

[0019] As a further technical solution, in order to more accurately characterize the comprehensive influence of the layered medium on the frequency modulation frequency of the linear frequency modulated signal, another method for calculating the equivalent frequency modulation parameter is based on the accumulation of propagation sensitivity; specifically, a layer-by-layer recursive sound velocity sequence c is extracted from the layered medium model. i and the thickness Δz of each water layer i And obtain the corresponding distance frequencies f. r,k The ray parameter value p(f) r,k For each distance frequency, according to p(f) r,k Starting from the depth of the towed body, traverse each water layer sequentially from top to bottom, and calculate the sensitivity of each water layer's contribution to the change of sound wave propagation time with frequency when that layer exists alone; this sensitivity is defined as the second derivative of the propagation time with respect to frequency for that layer, and can be obtained using the chain rule: Using implicit differentiation, the horizontal displacement equation can be derived. Through derivation, the contribution sensitivity of the i-th layer can be expressed as: In practical applications, the numerical difference method can be used: perturbation calculation of t near p. i The second-order difference of f is used; the contribution sensitivity of all water layers is accumulated in the order of sound wave propagation to obtain the total cumulative sensitivity value. This total value quantifies the combined modulation intensity of the sound wave across all water layers during its propagation from the towed body to the subsea pipeline; this cumulative sensitivity value is used as the equivalent frequency modulation parameter γ at that distance frequency point. eff (f r,k )=S tot (f r,k This parameter therefore carries the step-by-step modulation information of the signal modulation frequency by the layered medium structure.

[0020] As a further technical solution, the specific process of constructing phase correction values ​​based on equivalent frequency modulation parameters to achieve range-frequency domain phase distortion compensation is as follows: In the equivalent frequency modulation parameter sequence γ eff (f r Extract the equivalent frequency modulation value corresponding to each distance frequency position; for the distance frequency domain signal S(f) r,η For each column of azimuth data in (η is azimuth time), traverse each frequency point sequentially along the distance-frequency axis; at the current frequency point f r At this point, calculate relative to the reference frequency f. ref Take the offset Δf = f from the center of the carrier frequency or frequency band. r -f ref The equivalent frequency modulation value γ corresponding to this frequency point eff (f r ) and the square of the offset Δf2 By combining the results, the phase correction base number is obtained. The sign of the phase correction base is determined by the correction direction, with negative values ​​used to compensate for phase delay. This phase correction base is then used as the independent variable of a complex exponential function to construct the complex correction coefficients. Multiply the correction coefficient by the distance-frequency domain signal at the current frequency point, i.e.

[0021] This operation ensures that the phase adjustment of the signal at each frequency point corresponds precisely to the stepwise refraction effect of each water layer along its path. After performing the above operation point by point along all frequency points, a distance frequency domain signal with phase distortion corrected to physical matching is obtained.

[0022] As a further technical solution, the process of completing azimuth compression in the distance time domain and azimuth frequency domain further includes fine compensation for the azimuth phase distortion caused by the layered medium; specifically, it involves using the time delay-frequency correspondence τ(f r Starting from this point, for each distance unit r corresponding to a certain distance-frequency range, based on the geometric relationship between the towed body's trajectory and the position of the subsea pipeline, combined with the center frequency f corresponding to that distance unit... r,0 The actual one-way propagation time of η at different azimuth times can be calculated; this can be achieved by using τ(f r Mapping to azimuth-dependent propagation time: For a given distance unit, its corresponding actual slant range varies with azimuth. Therefore, it is necessary to recalculate the propagation time based on the geometric distance between the towed body position and the pipeline target at the azimuth time η, combined with sound velocity layering; let the towed body coordinates be (x(η), z... tow The target coordinates of the pipeline are (x...). pipe ,z pipe If ), then the horizontal distance is |x(η)-x pipe |, the vertical distance is |z pipe -z tow For each azimuth time, the corresponding one-way propagation time τ(η) is solved using the layered medium model and ray tracing to construct the actual propagation time sequence τ(η). A Fourier transform is performed on this actual propagation time sequence along the azimuth direction to obtain the phase spectrum Φ(f) of that range cell in the azimuth frequency domain. η )=FFT{τ(η)}; This phase spectrum reflects the azimuth phase history distortion caused by the layered medium; From Φ(f η Extract frequency points f from each direction. η The phase value at the given azimuth frequency is used as the phase difference between adjacent azimuth frequency points, which is then used as a measure of the azimuth group delay, i.e., the group delay. ;wherein τ g (f η ) represents the azimuth frequency f ηThe azimuth group delay at a given location characterizes the cumulative effect of phase variation with frequency; Φ(f η The azimuth frequency domain phase spectrum is obtained by the azimuth-directed Fourier transform of the range-compressed signal; Δf η Given the azimuth frequency sampling interval, satisfying the uniform sampling condition; based on the group delay, construct the azimuth frequency domain complex correction coefficients. The complex correction coefficient is multiplied by the azimuth frequency domain signal to obtain the azimuth phase-corrected signal; finally, the azimuth inverse Fourier transform is performed on the corrected signal to output a high-precision focused SAS image of the subsea pipeline.

[0023] This invention provides a high-frequency signal processing method for a SAS monitoring system for subsea pipelines. It has the following beneficial effects:

[0024] 1. This invention utilizes the phase characteristics of the submarine pipeline target itself in the original echo signal, and combines the initial measurement value of the sound velocity profile to construct the inversion equation set of the layered medium model, solve the sound velocity jump value at each interface, realize the accurate correction of the sound velocity parameters of each water layer in the layered medium model, improve the characterization accuracy of the layered medium model of the real marine acoustic environment, and lay an accurate physical basis for subsequent signal distortion compensation.

[0025] 2. This invention solves the problems of distance migration curve distortion and matched filter mismatch caused by the nonlinear distortion of propagation delay of different frequency components due to sound velocity stratification, and improves the focusing performance and signal energy concentration of distance compression. Attached Figure Description

[0026] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0027] Figure 1 This is a schematic diagram of the process of this invention. Detailed Implementation

[0028] The technical solutions in the embodiments of the present invention will be clearly and completely described below. 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.

[0029] like Figure 1 As shown in the figure, this embodiment of the invention provides a high-frequency signal processing method for a submarine pipeline SAS monitoring system.

[0030] In a submarine pipeline monitoring mission in a real-world marine area, a synthetic aperture sonar (SAS) system was mounted on a towed vehicle positioned 10 meters below the sea surface. The submarine pipeline to be monitored was buried below the seabed at a depth of approximately 50 meters, with the pipeline's center depth at 30 meters. The system emitted high-frequency linear frequency modulated pulses with a carrier frequency of 100 kHz, a signal bandwidth of 20 kHz, and a pulse width of 10 ms. The receiver acquired the echo signals at a sampling rate of 60 kHz, satisfying the Nyquist sampling theorem. In the azimuth direction, the pulse repetition frequency was 200 Hz. A single complete observation acquired 2048 azimuth pulses, with each pulse echo containing 4096 range sampling points, thus forming the original echo data matrix.

[0031] First, the raw echo signals collected by the synthetic aperture sonar system in the seabed pipeline route area are acquired. The signals are then processed to a form suitable for subsequent compensation. Specifically, a fast Fourier transform is performed along the range direction on the echo data of each pulse at each azimuth time, transforming it from the range time domain to the range frequency domain while retaining the azimuth time domain information, resulting in a range frequency domain signal. Each column of this signal corresponds to an azimuth time, and each row corresponds to a range frequency point, with the range frequency coverage ranging from -10kHz to +10kHz in the baseband. In actual processing, to reduce computational load and ensure accuracy, 256 frequency points are selected within the effective bandwidth, with a frequency sampling interval of approximately 156.25Hz.

[0032] Next, a layered medium model that accurately reflects the actual sound velocity layering structure of the sea area needs to be established, because the sound velocity profile obtained by conventional measurement methods often deviates from the actual propagation path, directly affecting the imaging quality. This method utilizes the phase characteristics of the submarine pipeline target itself in the echo to invert and correct the sound velocity parameters of each water layer. Complex data of the distance cell where the submarine pipeline is located is extracted from the original echo. According to prior information, the pipeline is roughly located at the 2000th distance cell, corresponding to a slant distance of about 100 m. The azimuth time of the strongest pipeline echo energy is searched along the azimuth direction as the reference azimuth. At this reference azimuth, the extracted distance cell data is Fourier transformed along the distance direction to obtain the complex spectrum in the range frequency domain, and then its amplitude spectrum and phase spectrum are extracted. The frequency corresponding to the peak position of the amplitude spectrum is taken as the reference frequency, which is exactly near the baseband 0Hz. The phase value of this reference frequency in the phase spectrum is set to 0, thus obtaining the relative phase spectrum after eliminating the fixed phase bias. This relative phase spectrum has a total of 256 points, denoted as . , where k=1,…,256.

[0033] The water column was divided into 10 water layers from a towed depth of 10m to a seabed depth of 50m, with each layer having a thickness of 4m. The depths of the interfaces between these layers were 10m, 14m, 18m, 22m, 26m, 30m, 34m, 38m, 42m, and 46m, with the seabed at a depth of 50m serving as the final interface. Therefore, there were a total of 10 sonic velocity transition interfaces. Initial values ​​for the sonic velocity in each layer were set based on historical hydrological data for this sea area, with the surface sonic velocity at 1500.0m / s, gradually decreasing downwards, with each layer changing by approximately [missing value]. The initial values ​​are set as follows: 1500.0 m / s for the first layer, 1499.2 m / s for the second layer, 1498.4 m / s for the third layer, 1497.6 m / s for the fourth layer, 1496.8 m / s for the fifth layer, 1496.0 m / s for the sixth layer, 1495.2 m / s for the seventh layer, 1494.4 m / s for the eighth layer, 1493.6 m / s for the ninth layer, 1492.8 m / s for the tenth layer, and 1492.0 m / s for the eleventh layer (seabed).

[0034] Under conditions of near-perpendicular incidence or small grazing angle of sound waves, the instantaneous phase increment occurs when the sound wave passes through each interface due to the abrupt change in sound velocity. This increment is proportional to the interface depth, the difference between the reciprocals of the sound velocities of the upper and lower layers, and the signal frequency. Therefore, the entire relative phase spectrum can be considered as the superposition of phase increments from all interfaces. To eliminate the influence of frequency variation in the grazing angle, the difference in the relative phase spectrum between adjacent frequency points is used as the observation. For adjacent frequency points f... r,k and f r,k+1 The difference It is approximately equal to a linear combination of the correlation quantity of the sound velocity jump at each interface multiplied by the interface depth and then multiplied by 4π times the frequency sampling interval. This establishes a system of linear equations containing 255 equations (corresponding to 256 frequency points generating 255 adjacent differences), with the unknowns being the differences of the reciprocals of the sound velocities at 10 interfaces. Each row of the coefficient matrix of the equation system consists of the interface depth multiplied by 4π times the frequency sampling interval, and the observation vector is the 255 adjacent phase differences. Since the number of equations far exceeds the number of unknowns, this system of equations is overdetermined. Considering measurement noise and model approximation, singular value decomposition is used to solve it, with a singular value truncation threshold set at 1% of the maximum singular value. The solution yields... After obtaining the difference in the reciprocals of the sound velocities at each interface, starting from the known surface sound velocity of 1500.0 m / s, the correction values ​​for the sound velocities of each layer are obtained by recursively calculating the reciprocal differences of the sound velocities at each layer. For example, if the difference in the reciprocal differences of the sound velocities at the first interface is 4.45e-7 s / m, then the correction value for the sound velocity at the second layer is 1 / (1 / 1500.0-4.45e-7)≈1499.0 m / s. And so on, until the corrected sound velocity sequence is obtained. The corrected sound velocity sequence is more consistent with the actual propagation path, and the layers are interconnected through recursive relationships, forming an accurate layered medium model. This recursive relationship is directly embedded in the layered medium model and used for subsequent calculations of the one-way propagation time corresponding to different distance frequencies.

[0035] Based on the modified layered medium model, it is necessary to calculate the one-way propagation time of sound waves from the towed body to the subsea pipeline corresponding to different distance frequencies. Based on ray acoustics theory, the ray parameter is used as the variable to be determined. This parameter remains constant along the entire ray path and is determined by the ratio of the grazing angle of the sound wave in each layer to the sound velocity in that layer. Starting from the depth of 10m where the towed body is located, each water layer is traversed sequentially from top to bottom. For a given ray parameter, the increment of horizontal displacement and the increment of propagation time of the sound wave in each layer are calculated. The horizontal displacement of each layer is accumulated to obtain the total horizontal displacement, and the propagation time of each layer is accumulated to obtain the total propagation time. For each distance frequency, the frequency is converted into the corresponding slant range according to the linear frequency modulation signal parameters, and then the target horizontal displacement is calculated by combining the vertical depth difference of 20m between the depth of the towed body (10m) and the depth of the subsea pipeline (30m). For example, for the frequency point of 0Hz baseband, the corresponding slant range is taken as the reference slant range of 102m, then the target horizontal displacement is sqrt(102m). 2 -20 2 =100.0m; then solve the ray parameter equation to make the total horizontal displacement equal to 100.0m; use the bisection method to iterate the solution within the feasible range of ray parameters from 0 to 1 / minimum sound speed, with the accuracy requirement that the horizontal displacement error is less than 1e-6m; after several iterations, the ray parameter value at this frequency point is approximately 6.65e-4s / m, and then the corresponding one-way propagation time is calculated to be 0.1333s; perform the above solution for all 256 frequency points one by one to obtain the numerical correspondence between time delay and frequency.

[0036] Next, it is necessary to quantify the comprehensive impact of the layered medium on the frequency modulation (FM) of the linear frequency modulated (LFM) signal. This impact is manifested as the second derivative of the propagation delay of different frequency components with respect to frequency, i.e., the equivalent FM parameter. In this embodiment, a method based on the accumulation of propagation sensitivity is used to calculate the equivalent FM parameter. Specifically, the layered medium model is used to extract the recursively derived sound velocity sequence and the thickness of each water layer, and the ray parameter values ​​corresponding to each distance frequency are obtained. For each distance frequency, based on the ray parameter corresponding to that frequency, each water layer is traversed sequentially from top to bottom, starting from a tow body depth of 10m. Within each water layer, the contribution of the water layer alone to the change of sound wave propagation time with frequency is calculated. Sensitivity; this sensitivity can be approximated by obtaining the second-order difference of propagation time with respect to frequency after making a small perturbation to the ray parameters; for example, for a 0Hz frequency point, its ray parameter p = 6.65e-4s / m, a small perturbation of ±1e-8s / m is applied near p, the propagation time of each water layer is recalculated, and then the second-order partial derivative of the propagation time of each layer with respect to frequency is obtained; the contribution sensitivity of all water layers passed through is accumulated in the order of sound wave propagation to obtain the total cumulative sensitivity value; this total value is the equivalent frequency modulation parameter at that frequency point; for a 0Hz frequency point, the calculated equivalent frequency modulation parameter is approximately 1.23e-5s / Hz.

[0037] After obtaining the equivalent frequency modulation parameters, a phase correction value can be constructed in the range frequency domain to compensate for the nonlinear distortion of the range migration curve caused by sound velocity stratification. Using the signal carrier frequency of 100kHz (i.e., baseband 0Hz) as the reference frequency, the offset relative to the reference frequency is calculated for each range frequency point. For the k-th frequency point, the offset Δf... k =f r,k-0 The equivalent frequency modulation value γ corresponding to this frequency point eff (f r The phase correction base is obtained by combining the square of the offset Δf2 with the offset. Take the negative of the base as the independent variable of the complex exponential function to construct the complex correction coefficient. Multiply the correction coefficient by the distance-frequency domain signal at the current frequency point, i.e. For example, at a frequency of 0Hz, the offset is 0, the correction coefficient is 1, and the phase remains unchanged; at a frequency of +5kHz, the offset is 5000Hz. If the equivalent modulation frequency of this point is 1.25e-5s / Hz, then the phase correction base is π×1.25e-5×5000≈0.98rad, and the corresponding correction coefficient is exp(-j0.98). Multiplying this by the data at this frequency point achieves phase compensation; after operating on all frequency points one by one, the corrected distance-frequency domain signal is obtained.

[0038] The corrected range-frequency domain signal is subjected to an inverse Fourier transform along the range direction to restore the signal to the range-time domain. At this point, range migration correction is complete, and the range-migration-corrected time-domain signal is obtained. Next, range-direction matched filtering is performed: the range-time domain signal is again subjected to a Fourier transform along the range direction and multiplied by the frequency response of the matched filter. This response is determined by the modulation frequency of the transmitted signal, which is 2e6Hz / s, and its frequency domain form is exp(jπf r / γ); After multiplication, an inverse Fourier transform is performed to achieve range compression, resulting in a range-compressed signal; Tests show that the main lobe width of the range-compressed pulse is approximately 0.075m, and the peak sidelobe level is below -13dB, which is consistent with the theoretical value, indicating that the range distortion has been effectively corrected.

[0039] After range compression, the signal remains in both the range and azimuth time domains, requiring further azimuth processing to achieve synthetic aperture focusing. Azimuth processing must consider the azimuth phase history distortion caused by the layered medium. Therefore, based on the previously established correspondence between time delay and frequency values, for each range unit, the actual one-way propagation time at different azimuth moments is calculated according to the geometric relationship between the towed body's trajectory and the subsea pipeline's position, combined with the range frequency range corresponding to that range unit, thus forming an actual propagation time sequence. Specifically, for each azimuth moment, the instantaneous position of the towed body and the fixed position of the pipeline are known. The towed body moves in a straight line at a speed of approximately 2 m / s, with the horizontal position of the pipeline center set to 0. The horizontal position of the towed body changes with the azimuth moment; for example, at azimuth moment η=0, the towed body is located at x=-100m, and at the synthetic aperture center moment η=50s, the towed body is located at x=0. For a given range unit, its center frequency is f. r,0 The corresponding reference slant distance is R0. Using a layered medium model and ray tracing, the ray parameters that make the horizontal displacement equal to the current horizontal distance between the towed body and the pipe and the vertical displacement equal to 20m are solved, thus obtaining the actual one-way propagation time τ(η) at that azimuth moment. This calculation is repeated for all azimuth moments to obtain the actual propagation time sequence of that distance unit, totaling 2048 points. A Fourier transform is performed on this sequence along the azimuth direction to obtain the phase spectrum Φ(f) in the azimuth frequency domain. η Since the azimuth sampling rate is 200Hz, the azimuth frequency range is -100Hz to 100Hz, and the frequency sampling interval is approximately 0.0977Hz; the phase values ​​of each azimuth frequency point are extracted from the phase spectrum, and the azimuth group delay is calculated using the center difference between adjacent frequency points.

[0040] For example, at an azimuth frequency of 0 Hz, the group delay is approximately 0.1333 s; at +50 Hz, the group delay is approximately 0.1335 s; the azimuth frequency domain complex correction coefficients are constructed based on the group delay. ; azimuth frequency domain signal Src (t,f η Multiply by H at each frequency point az (f η The system compensates for the azimuth phase distortion caused by the layered medium. Finally, the corrected azimuth frequency domain signal is subjected to inverse Fourier transform along the azimuth direction to output the final SAS image of the subsea pipeline.

[0041] Through the above steps, this embodiment makes full use of the sound velocity layering information carried by the echo itself, accurately inverts the sound velocity of each water layer, establishes an accurate propagation model, and performs physical matching compensation for the signal distortion caused by the layered medium in the range and azimuth directions respectively.

[0042] 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 high-frequency signal processing method for a submarine pipeline SAS monitoring system, characterized in that, include: S1: Acquire the raw echo signal collected by the synthetic aperture sonar system in the seabed pipeline route area; S2: Perform a Fourier transform on the original echo signal along the range direction to transform it to the range frequency domain and the azimuth time domain to obtain the range frequency domain signal; S3: Based on the obtained sound velocity profile data of the subsea pipeline route area, establish a layered medium model containing multiple water layers with constant sound velocity in each layer; calculate the one-way propagation time of sound waves from the towed body to the subsea pipeline for different distance frequencies for the layered medium model, thereby forming a correspondence between time delay and frequency values. Based on the correspondence between the time delay and frequency values, the equivalent frequency modulation parameter that varies with distance frequency is obtained through the numerical difference method. This parameter characterizes the comprehensive influence of the layered medium on the frequency modulation of the linear frequency modulation signal. According to the equivalent frequency modulation parameter, a phase correction value is constructed in the distance frequency domain. This phase correction value is used to compensate for the nonlinear distortion of the distance migration curve caused by sound velocity layering. The phase correction value is combined with the range frequency domain signal to obtain the corrected range frequency domain signal; S4: Perform an inverse Fourier transform on the corrected range frequency domain signal along the range direction to obtain the range migration corrected time domain signal; S5: Perform range-direction matched filtering on the time-domain signal after range migration correction to achieve range compression and obtain the range-compressed signal; S6: Perform a Fourier transform on the range-compressed signal along the azimuth direction to transform it to the range time domain and azimuth frequency domain, and perform azimuth compression in these domains. Then perform an inverse Fourier transform on the azimuth-compressed signal to output the SAS image of the subsea pipeline.

2. The high-frequency signal processing method for a submarine pipeline SAS monitoring system according to claim 1, characterized in that: The layered medium model includes: extracting complex data of the distance cell where the submarine pipeline is located from the original echo signal, and selecting the azimuth position with the strongest pipeline echo energy as the reference azimuth; performing a Fourier transform on the extracted distance cell data along the distance direction at the reference azimuth, and extracting its amplitude spectrum and phase spectrum in the distance frequency domain; using the frequency corresponding to the peak position in the amplitude spectrum as the reference, setting the phase value of the reference frequency in the phase spectrum to zero to obtain the relative phase spectrum; dividing the water column into N layers according to depth, using the interface depth of each layer as the model node, and setting the initial sound velocity of each layer to the measured value of the sound velocity profile data at the corresponding depth; representing the relative phase spectrum as the sum of the phase jump values ​​caused by the sound velocity jump at each layer interface, that is, the value of the relative phase spectrum at each frequency point is equal to the sum of the instantaneous phase increments caused by the sound velocity change when the sound wave passes through each layer interface, and the instantaneous phase increment is determined by the ratio of the sound velocities of the upper and lower layers of the interface and the interface depth; using the difference in the relative phase spectrum of adjacent frequency points as the observation value, and establishing a system of linear equations about the sound velocity jump values ​​at each layer interface.

3. The high-frequency signal processing method for a submarine pipeline SAS monitoring system according to claim 2, characterized in that: In the linear equation system, each equation corresponds to a pair of adjacent frequency points, and each equation includes the contribution of the sound velocity jump values ​​at all layer interfaces. The equation system is solved by singular value decomposition to obtain the sound velocity jump values ​​at each layer interface. The sound velocity jump values ​​at each layer interface are accumulated in depth order to obtain the corrected sound velocity values ​​for each layer, and the sound velocity parameters of each water layer are updated with these corrected values. The updated sound velocity parameters of each layer are arranged in depth order to form a layer-by-layer recursive sound velocity sequence. The sound velocity value of each layer in this sequence is obtained by adding the sound velocity value of the previous layer to the sound velocity jump value of the corresponding interface, thus establishing a recursive relationship between the sound velocities of the layers. This recursive relationship is directly embedded in the layered medium model and used for subsequent calculation of the one-way propagation time corresponding to different distance frequencies.

4. The high-frequency signal processing method for a submarine pipeline SAS monitoring system according to claim 3, characterized in that: The process of calculating the one-way propagation time corresponding to different distance frequencies using a layered medium model includes: extracting a corrected sound velocity sequence from the layered medium model, where the sound velocity value of each layer is obtained by summing the sound velocity value of the previous layer and the sound velocity jump value of the corresponding interface, and the layers are interconnected through this recursive relationship; using a ray parameter as the variable to be determined, which remains constant along the entire ray path, its value is determined by the ratio of the grazing angle of the sound wave in any water layer to the sound velocity of that layer; starting from the depth where the towed body is located, traversing each water layer from top to bottom, determining the grazing angle of the sound wave in the current layer based on the sound velocity value and the ray parameter, and then using the thickness of that layer to calculate... The horizontal displacement increment and propagation time increment generated by the propagation of the sound wave within the layer are calculated. The horizontal displacement increments of each layer are summed to obtain the total horizontal displacement corresponding to the current ray parameter, and the propagation time increments of each layer are summed to obtain the corresponding total propagation time. For each distance frequency, the frequency is converted into the corresponding straight-line distance according to the linear frequency modulated signal parameters transmitted by the SAS system, and the target horizontal displacement is calculated by combining the difference between the towed body depth and the depth of the subsea pipeline. The ray parameter equation is solved to make the total horizontal displacement equal to the target horizontal displacement, and the total propagation time at this time is taken as the one-way propagation time of that distance frequency. All distance frequencies are traversed to form a numerical correspondence between time delay and frequency.

5. The high-frequency signal processing method for a submarine pipeline SAS monitoring system according to claim 1, characterized in that: The process of obtaining the equivalent frequency modulation parameter varying with distance using the numerical difference method includes: extracting a layer-by-layer recursive sound velocity sequence from the layered medium model, using the ratio of the sound velocities of adjacent layers in the sound velocity sequence as the refractive intensity coefficient of that layer interface; in the established numerical correspondence between time delay and frequency, for each distance frequency point, estimating the penetration depth of the main sound energy in the layered medium according to the correspondence between frequency and sound velocity, and extracting a corresponding coefficient segment from the refractive intensity coefficient sequence based on this depth; performing a cumulative multiplication operation on this segment of refractive intensity coefficients to obtain a cumulative refractive intensity factor characterizing the comprehensive distortion of the sound wave path by all water layer interfaces from the tow body to the penetration depth; using this cumulative refractive intensity factor as a weight, weighting and fusing it with the initial second-order difference value at that frequency point obtained through three-point difference to obtain a corrected second-order difference value; using this corrected second-order difference value as the basis for calculating the equivalent frequency modulation parameter at that frequency point, and participating in the construction of subsequent phase correction values.

6. The high-frequency signal processing method for a submarine pipeline SAS monitoring system according to claim 5, characterized in that: The equivalent frequency modulation parameter characterizes the comprehensive influence of the layered medium on the frequency modulation of a linear frequency modulated signal. A recursive sound velocity sequence and the thickness of each water layer are extracted from the layered medium model, and ray parameter values ​​corresponding to each distance frequency are obtained. For each distance frequency, based on the ray parameter corresponding to that frequency, each water layer is traversed sequentially from top to bottom, starting from the depth where the towed body is located. Within each water layer, based on the sound velocity value, thickness, and the grazing angle of the sound wave determined by the ray parameter, the sensitivity of the water layer's contribution to the change in sound wave propagation time with frequency is calculated. This sensitivity characterizes the local influence intensity of the medium layer on the frequency modulation. The contribution sensitivity of all traversed water layers is accumulated in the order of sound wave propagation to obtain the cumulative sensitivity value, which quantifies the comprehensive modulation intensity of the sound wave from the towed body to the subsea pipeline. This cumulative sensitivity value is used as the equivalent frequency modulation parameter at that distance frequency point.

7. The high-frequency signal processing method for a submarine pipeline SAS monitoring system according to claim 6, characterized in that: The equivalent frequency modulation (CCM) value corresponding to each distance frequency position is extracted from the equivalent frequency modulation parameter sequence. This equivalent CCM value itself is the sum of the sensitivity contributions of each water layer through which the sound wave propagates from the towed body to the subsea pipeline, accumulated in the order of propagation. Therefore, it carries the step-by-step modulation information of the signal frequency modulation by the layered medium structure. For each column of azimuth data in the distance frequency domain signal, each frequency position is traversed sequentially along the distance frequency axis. At the current frequency position, the equivalent CCM value corresponding to that frequency point is calculated by dividing the square of the offset of that frequency point relative to the reference frequency. The combined result is used as the phase correction base for the current frequency point. This phase correction base implicitly contains the cumulative contribution of all water layers on the acoustic path corresponding to this frequency point to the phase distortion. The phase correction base is used as the independent variable of a complex exponential function to construct a complex correction coefficient. This complex correction coefficient is multiplied by the range frequency domain signal at the current frequency point so that the phase adjustment of the signal at this frequency point corresponds precisely to the stepwise refraction effect of each water layer on its path. After performing the above operation point by point along all frequencies, the range frequency domain signal with physically matched phase distortion correction is obtained.

8. The high-frequency signal processing method for a submarine pipeline SAS monitoring system according to claim 1, characterized in that: S6 includes: from the correspondence between time delay and frequency values, for each distance unit, based on the geometric relationship between the towed body's trajectory and the subsea pipeline's position, and combined with the distance frequency range corresponding to that distance unit, calculating the actual one-way propagation time at different azimuth times to form an actual propagation time sequence; performing a Fourier transform on the actual propagation time sequence along the azimuth direction to obtain the phase spectrum of that distance unit in the azimuth frequency domain, which reflects the historical azimuth phase distortion caused by the layered medium; extracting the phase values ​​at each azimuth frequency point from the phase spectrum, and using the phase difference between adjacent azimuth frequency points as a measure of the azimuth group delay, thereby obtaining the variation law of the azimuth phase with frequency; based on this variation law, constructing a complex correction coefficient for each azimuth frequency point in the azimuth frequency domain, the phase of which corresponds to the group delay value at that azimuth frequency point, used to compensate for the azimuth phase distortion caused by the layered medium; multiplying the complex correction coefficient with the azimuth frequency domain signal to obtain the azimuth phase-corrected signal; and then performing an inverse azimuth Fourier transform on the corrected signal to output the subsea pipeline SAS image.

Images