A ship radiated noise enhancement method and system, electronic equipment and storage medium

Abstract

The application discloses a ship radiation noise enhancement method and system, electronic equipment and a storage medium, and relates to the technical field of radiation noise enhancement. The array receiving signal is sequentially subjected to time delay beam forming and time sliding window segmentation to obtain a plurality of time direction signal vectors, and a plurality of to-be-processed signal vectors are selected from the time direction signal vectors. For a FrFT domain real signal determined after processing of any to-be-processed signal vector by using a plurality of different fractional Fourier transform algorithms, fitting goodness tests of Alpha stable distribution and asymmetric Laplace distribution are respectively performed to obtain a first D value and a second D value, and a target order and a target statistical distribution are determined. Any time direction signal vector is processed by using a fractional Fourier transform algorithm with the target order to determine a target FrFT domain real signal. According to the target statistical distribution and all target FrFT domain real signals, a time bearing history graph of the array receiving signal is determined, and the enhancement of the ship radiation noise is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of radiation noise enhancement, in particular to a ship radiation noise enhancement method, system, electronic equipment and storage medium. BACKGROUND

[0002] Sound waves are currently the only means known to be able to propagate over long distances in the ocean, and underwater acoustic engineering technology based on sonar has a very important role and significance for China's marine industry. According to different working methods, sonar systems can be divided into active sonar systems and passive sonar systems. For array signal processing under passive sonar, ship radiation noise enhancement and spatio-temporal information extraction have always been a research hotspot.

[0003] The time-bearings history plot is a very common and effective target trajectory or spatio-temporal information representation ship radiation noise enhancement method. For the traditional time-bearings history plot, the general steps of the algorithm process are: first, the array received signal is spatially scanned by spatial filtering to obtain received signals with different bearing directivity angles under a certain directivity angle step; the received signals with different bearing directivity are segmented under a certain time window and step; the energy of each segment signal is calculated to obtain the time-bearings history plot.

[0004] The traditional time-bearings history plot algorithm is simple and easy to implement, but its shortcomings are also obvious: when the actual passive sonar received signal has a low signal-to-noise ratio, the performance of the traditional time-bearings history plot will decay; when the array received signal contains strong pulse interference, its performance will be greatly degraded or even invalid, and the target trajectory information cannot be effectively extracted; the history plot only involves energy characteristics. SUMMARY

[0005] The purpose of the present application is to provide a ship radiation noise enhancement method, system, electronic equipment and storage medium, which realizes the enhancement of ship radiation noise.

[0006] To achieve the above purpose, the present application provides the following solutions:

[0007] A ship radiation noise enhancement method, comprising:

[0008] Obtaining an array received signal received by a passive sonar; the array received signal contains information of ship radiation;

[0009] Based on J1 different preset directivity angles, time delay beamforming is performed on the array received signal to obtain a direction signal matrix; the direction signal matrix includes J1 direction signal vectors with different preset directivity angles; J1>1;

[0010] For any direction signal vector:

[0011] time-sliding segmenting the direction signal vectors to obtain J2time direction signal vectors with preset pointing angles and different time period centers;

[0012] selecting m max time direction signal vectors from the J2time direction signal vectors, thereby obtaining m max time direction signal vectors to be processed; 1≤m max ≤J2;

[0013] for any of the time direction signal vectors to be processed:

[0014] performing fractional Fourier transform on the time direction signal vectors to be processed respectively by using p fractional Fourier transform algorithms with different preset orders, thereby determining p FrFT domain real signals;

[0015] for any of the FrFT domain real signals:

[0016] performing goodness-of-fit test of Alpha stable distribution on the FrFT domain real signals to obtain D value of the FrFT domain real signals under Alpha stable distribution, denoted as first D value;

[0017] performing goodness-of-fit test of asymmetric Laplace distribution on the FrFT domain real signals to obtain D value of the FrFT domain real signals under asymmetric Laplace distribution, denoted as second D value;

[0018] summing up final D values of the first D values of all FrFT domain real signals under the same preset order to obtain p first total D values;

[0019] summing up the second D values of all FrFT domain real signals under the same preset order to obtain p second total D values;

[0020] determining target order and target statistical distribution based on the p first total D values and the p second total D values; the target statistical distribution is Alpha stable distribution or asymmetric Laplace distribution;

[0021] for any of the time direction signal vectors:

[0022] performing fractional Fourier transform on the time direction signal vectors by using fractional Fourier transform algorithm with the target order, thereby determining target FrFT domain real signal;

[0023] determine a time-bearings range profile of the array receiving signal according to the target statistical distribution and all the target FrFT domain real signals, and realize enhancement of the ship radiated noise; the time-bearings range profile is a time-bearings range profile under an Alpha stable distribution or a time-bearings range profile under an asymmetric Laplace distribution; the time-bearings range profile under the Alpha stable distribution includes a characteristic-time-bearings range profile and a scale-time-bearings range profile, and the time-bearings range profile under the asymmetric Laplace distribution includes a shape-time-bearings range profile and a standard deviation-time-bearings range profile.

[0024] Optionally, the p FrFT domain real signals are determined by performing fractional Fourier transform on the to-be-processed signal vector respectively by using p fractional Fourier transform algorithms with different preset orders.

[0025] The p FrFT domain real signals are determined by performing fractional Fourier transform on the to-be-processed signal vector respectively by using p fractional Fourier transform algorithms with different preset orders.

[0026] The p FrFT domain real signals are determined by taking real parts of the FrFT domain signals.

[0027] Optionally, a goodness-of-fit test of the Alpha stable distribution is performed on the FrFT domain real signals to obtain a D value of the FrFT domain real signals under the Alpha stable distribution, denoted as a first D value.

[0028] The parameter combination under the Alpha stable distribution is determined based on the FrFT domain real signals by using a sample characteristic function method; the parameter combination under the Alpha stable distribution includes a characteristic index, a symmetry parameter, a scale coefficient and a location parameter.

[0029] The probability density function under the Alpha stable distribution is determined based on the parameter combination under the Alpha stable distribution by using a direct numerical integration method.

[0030] The numerical values of the FrFT domain real signals are arranged in ascending order, and the FrFT domain real signals are evenly divided into preset interval numbers of intervals.

[0031] The theoretical frequency under the Alpha stable distribution of each interval is determined based on the probability density function under the Alpha stable distribution.

[0032] For any current interval:

[0033] It is determined whether the theoretical frequency under the Alpha stable distribution of the current interval is less than a preset frequency.

[0034] If yes, merging the interval adjacent to the current interval and the current interval into one interval, updating the interval corresponding to the FrFT domain real signal, and returning to "determining theoretical frequency of each interval based on probability density function under Alpha stable distribution";

[0035] If no, calculating the actual frequency of the current interval;

[0036] Based on the theoretical frequency and the actual frequency of all intervals under Alpha stable distribution, calculating the D value of the FrFT domain real signal under Alpha stable distribution, and recording as a first D value.

[0037] Optionally, performing goodness-of-fit test of the FrFT domain real signal under asymmetric Laplace distribution to obtain the D value of the FrFT domain real signal under asymmetric Laplace distribution, and recording as a second D value, specifically comprising:

[0038] Using a parameter estimation method of asymmetric Laplace distribution, determining a parameter combination under asymmetric Laplace distribution based on the FrFT domain real signal; the parameter combination under asymmetric Laplace distribution comprises a shape parameter, a standard deviation and a location parameter;

[0039] Based on the definition of asymmetric Laplace distribution and the parameter combination under asymmetric Laplace distribution, determining a probability density function under asymmetric Laplace distribution;

[0040] Arranging the numerical values of the FrFT domain real signal in ascending order, and dividing the FrFT domain real signal into a preset number of intervals uniformly;

[0041] Based on the probability density function under asymmetric Laplace distribution, determining the theoretical frequency of each interval under asymmetric Laplace distribution;

[0042] For any current interval:

[0043] Determining whether the theoretical frequency of the current interval under asymmetric Laplace distribution is less than a preset frequency;

[0044] If yes, merging the interval adjacent to the current interval and the current interval into one interval, updating the interval corresponding to the FrFT domain real signal, and returning to "determining theoretical frequency of each interval based on probability density function under asymmetric Laplace distribution";

[0045] If no, calculating the actual frequency of the current interval;

[0046] Based on the theoretical frequency and the actual frequency of all intervals under asymmetric Laplace distribution, calculating the D value of the FrFT domain real signal under asymmetric Laplace distribution, and recording as a second D value.

[0047] Optionally, the target order and the target statistical distribution are determined based on the p first total D values and the p second total D values, specifically including:

[0048] The preset order corresponding to the maximum value of the p first total D values and the p second total D values is determined as the target order;

[0049] The Alpha stable distribution or the asymmetric Laplace distribution corresponding to the maximum value of the p first total D values and the p second total D values is determined as the statistical distribution.

[0050] Optionally, the time direction signal vector is subjected to fractional Fourier transform by using a fractional Fourier transform algorithm with the target order, so as to determine a target FrFT domain real signal, specifically including:

[0051] The time direction signal vector is subjected to fractional Fourier transform by using a fractional Fourier transform algorithm with the target order, so as to obtain a target FrFT domain signal;

[0052] The real part of the target FrFT domain signal is taken, so as to obtain the target FrFT domain real signal.

[0053] Optionally, the time direction history map of the array receiving signal is determined according to the target statistical distribution and all the target FrFT domain real signals, specifically including:

[0054] When the target statistical distribution is an Alpha stable distribution:

[0055] For any target FrFT domain real signal:

[0056] A target parameter combination under the Alpha stable distribution is determined based on the target FrFT domain real signal by using a sample characteristic function method; the parameter combination under the Alpha stable distribution includes a target characteristic index, a target symmetry parameter, a target scale coefficient and a target position parameter;

[0057] All the target characteristic indexes are represented in a gray scale map or a pseudo-color map, so as to obtain a characteristic-time direction history map;

[0058] All the target scale coefficients are represented in a gray scale map or a pseudo-color map, so as to obtain a scale-time direction history map;

[0059] When the target statistical distribution is an asymmetric Laplace distribution:

[0060] For any target FrFT domain real signal:

[0061] Using the parameter estimation method of the asymmetric Laplace distribution, based on the target FrFT domain real signal, the target parameter combination under the asymmetric Laplace distribution is determined; the target parameter combination under the asymmetric Laplace distribution includes: target shape parameter, target standard deviation, and target position parameter;

[0062] Represent all the target shape parameters as grayscale or pseudocolor images to obtain the shape-time orientation history map;

[0063] Represent all the target standard deviations as grayscale or pseudocolor images to obtain the standard deviation-time azimuth history map.

[0064] A ship radiated noise enhancement system includes:

[0065] An array receiving signal acquisition module is used to acquire array receiving signals received by a passive sonar; the array receiving signals contain information about ship radiation.

[0066] A time-delay beamforming module is used to perform time-delay beamforming on the array received signal based on J1 different preset pointing angles to obtain a directional signal matrix; the directional signal matrix includes J1 directional signal vectors with different preset pointing angles; J1 > 1;

[0067] The signal vector determination module is used for determining the signal vector to be processed in any of the stated directions:

[0068] The direction signal vector is segmented by a time sliding window to obtain J2 time direction signal vectors with preset pointing angles and different time interval centers;

[0069] Select m from the J2 time direction signal vectors max Each time-direction signal vector is used to obtain m max There are n signal vectors to be processed; 1 ≤ m max ≤J2;

[0070] The FrFT domain real signal determination module is used for any of the aforementioned signal vectors to be processed:

[0071] Using p fractional Fourier transform algorithms with different preset orders, fractional Fourier transforms are performed on the signal vector to be processed, thereby determining p real signals in the FrFT domain.

[0072] The total D value determination module is used for any real signal in the FrFT domain:

[0073] Perform a goodness-of-fit test on the real signal in the FrFT domain under the Alpha stable distribution to obtain the D value under the Alpha stable distribution of the real signal in the FrFT domain, which is denoted as the first D value;

[0074] performing a goodness-of-fit test on the FrFT domain real signals according to the asymmetric Laplace distribution to obtain a D value of the FrFT domain real signals under the asymmetric Laplace distribution, denoted as a second D value;

[0075] summing up the final D values of the first D values of all FrFT domain real signals at the same preset order to obtain p first total D values;

[0076] summing up the second D values of all FrFT domain real signals at the same preset order to obtain p second total D values;

[0077] a target order and target statistical distribution total D value determination module configured to determine a target order and a target statistical distribution based on the p first total D values and the p second total D values; the target statistical distribution is an Alpha stable distribution or an asymmetric Laplace distribution;

[0078] a target FrFT domain real signal determination module configured to perform, for any time direction signal vector, a fractional Fourier transform on the time direction signal vector by using a fractional Fourier transform algorithm with the target order, so as to determine a target FrFT domain real signal;

[0079] a target FrFT domain real signal determination module configured to perform, for any time direction signal vector, a fractional Fourier transform on the time direction signal vector by using a fractional Fourier transform algorithm with the target order, so as to determine a target FrFT domain real signal;

[0080] a time bearing history map determination module configured to determine a time bearing history map of the array receiving signal according to the target statistical distribution and all the target FrFT domain real signals, so as to enhance the ship radiated noise; the time bearing history map is a time bearing history map under an Alpha stable distribution or a time bearing history map under an asymmetric Laplace distribution; the time bearing history map under the Alpha stable distribution includes a characteristic-time bearing history map and a scale-time bearing history map, and the time bearing history map under the asymmetric Laplace distribution includes a shape-time bearing history map and a standard deviation-time bearing history map.

[0081] An electronic device includes:

[0082] one or more processors;

[0083] a memory device having one or more programs stored thereon;

[0084] when the one or more programs are executed by the one or more processors, the one or more processors implement the ship radiated noise enhancement method as described above.

[0085] A storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the ship radiated noise enhancement method as described above.

[0086] According to the specific embodiments of the present application, the following technical effects are disclosed:

[0087] The application discloses a ship radiation noise enhancement method and system, an electronic device and a storage medium, and the statistical characteristics in a fractional order domain are described and a suitable target statistical distribution is selected after time delay beam forming and time sliding window segmentation, that is, the signal energy distribution is changed through FrFT to suppress the statistical characteristics of noise and enhance the statistical signal, a new feature of non-energy is obtained through parameter estimation of the statistical distribution, and a new time bearing history diagram is obtained, so that the enhancement of the target radiation noise is realized. BRIEF DESCRIPTION OF DRAWINGS

[0088] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0089] Figure 1 A ship radiation noise enhancement method flowchart is provided for the embodiment 1 of the present application.

[0090] Figure 2 A rotation diagram of the fractional Fourier transform on the time-frequency plane is shown. DETAILED DESCRIPTION

[0091] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0092] The purpose of the present application is to provide a ship radiation noise enhancement method, system, electronic device and storage medium, which aims to enhance the ship radiation noise.

[0093] In order to make the above-mentioned purposes, characteristics and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0094] Embodiment 1

[0095] Figure 1 A ship radiation noise enhancement method flowchart is provided for the embodiment 1 of the present application. As shown in the figure, the ship radiation noise enhancement method in the embodiment includes: Figure 1

[0096] ​Step 101: obtaining an array received signal received by a passive sonar; the array received signal contains information of ship radiations.

[0097] Step 102: based on J1 different preset pointing angles, performing time delay beam forming on the array received signal to obtain a directional signal matrix.

[0098] The directional signal matrix includes J1 directional signal vectors with different preset pointing angles; J1>1.

[0099] Specifically, in the implementation of step 102, the array received signal received by the passive sonar with a sampling frequency of f s is subjected to spatial domain filtering, in this case, by means of time delay beam forming, and the beam forming pointing angle is set to step θ min , θ min may be set according to actual conditions, for example, θ min is set to Δθ / 5, and Δθ is the main lobe width of the beam forming; the variation range of the directivity angle is set to 0-θ max ; when the receiving array participating in the beam forming is a linear array, the beam forming has the characteristics of left and right side lobe ambiguity, and θ max is set to 180°; when the receiving array participating in the beam forming is a non-linear array, θ max is set to 360°.

[0100] The array received signal is taken as input, and different pointing angles (i.e., preset pointing angles) θ j1 =j1θ min , j1=1, 2, …J1, representing the floor operation, a plurality of signal vectors with corresponding directivities (i.e., directional signal vectors) are obtained by time delay beam forming, wherein the j1th directional signal vector is represented as The length of is M points.

[0101] Step 103: determining a plurality of to-be-processed signal vectors corresponding to each directional signal vector.

[0102] Wherein, for any directional signal vector:

[0103] The directional signal vector is subjected to time sliding window segmentation to obtain J2 time directional signal vectors with preset pointing angles and different time period centers.

[0104] Specifically, for All samples are segmented using a time sliding window of a certain width and step size. To ensure statistical stability and good time resolution of the signal sequence, based on experience, the length T of the time sliding window can be set to 3000 sampling points. While ensuring the stability of the signal sequence, the length T can be adjusted according to the actual required time resolution (the higher the time resolution, the shorter the length T). The step size Δt... T Setting it to T / 4 affects the smoothness of the final result. The smaller the step size, the smoother the final time-location history plot. It can be adjusted flexibly according to actual needs.

[0105] by For example, As input, after time-window segmentation, the pointing angle is obtained as follows: The center of the time period is signal vector (i.e., having the j1th preset pointing angle) and the center τ of the j2nd time period j2 (time direction signal vector), j2=1,2,…J2, t1 = j3 / f s j3 = 1, 2, ..., T. For ease of representation, we will start from here. Abbreviated as All time-direction signal vectors are composed of signals with different time-time centers. and preset pointing angle Time direction signal matrix X θτ X θτ The expression is:

[0106] Select m from J2 time direction signal vectors max Each time-direction signal vector is used to obtain m max There are n signal vectors to be processed; 1 ≤ m max ≤J2.

[0107] Specifically, using the j1th preset pointing angle For example, from Corresponding time direction signal vector Select the time period center as τ n signal vector (that is, m was uniformly sampled for each directional angle) max (group signal), m = 1, 2, ..., m max Generally m max Let's take 10. In practical applications, J2 is large enough. m maxThe step size can be appropriately increased to increase the accuracy of selecting the model (i.e., the target statistical distribution) and the optimal order (i.e., the target order) in the subsequent step, and the selection method can also be changed according to actual conditions to ensure that it is arbitrary directivity The non-subjective random selection of the direction signal vector corresponding to the different time periods can be performed.

[0108] Step 104: Determine the p FrFT domain real signals corresponding to each to-be-processed signal vector.

[0109] For any to-be-processed signal vector:

[0110] The p FrFT domain real signals are determined by performing, respectively, fractional Fourier transform on the to-be-processed signal vector using p fractional Fourier transform algorithms with different preset orders.

[0111] As an optional implementation, the p FrFT domain real signals are determined by performing, respectively, fractional Fourier transform on the to-be-processed signal vector using p fractional Fourier transform algorithms with different preset orders, and the specific implementation includes the following steps.

[0112] The p FrFT domain signals with preset directivity angles and time period centers are obtained by performing, respectively, fractional Fourier transform on the to-be-processed signal vector using p fractional Fourier transform algorithms with different preset orders.

[0113] Specifically, taking the to-be-processed signal vector as an example, the selected is subjected to p different preset order fractional Fourier transform (FrFT) processing, and p can be adjusted according to actual conditions, for example, the range of the order p is selected as [0.5, 1.5], the step size Δt F is 0.1, that is, p = 0.5 + 0.1j4, j4 = 0, 1, 2, …, J4, the step size Δt F can be reduced to obtain a more accurate optimal p value (i.e., the target order) in the subsequent step.

[0114] As shown in FIG. 1, when performing FrFT on a general one-dimensional signal x(t), the FrFT algorithm is: Figure 2

[0115]

[0116] wherein the kernel function K p (u, t) is defined as:

[0117]

[0118] ​

[0119] wherein t is a time domain parameter of the signal x(t), u is a parameter of the signal x(t) in the fractional order domain, δ(·) is an impulse function, n F is an integer, i is a unit imaginary number, α=p1π / 2, α is a rotation angle of the FrFT, p1 is an order of the FrFT, and the corresponding signal x(t) can be referred to as a Fourier transform at an angle of α or a p1-order Fourier transform of the signal x(t), A α is an exponential term coefficient in the kernel function when α≠π.

[0120] The to-be-processed signal vector is taken as an input one-dimensional signal x(t), and the FrFT algorithm is used at different orders p to obtain the corresponding FrFT domain signals at the time period center τ under the pointing angle n .

[0121] The real parts of the FrFT domain signals are taken to obtain p FrFT domain real signals.

[0122] Specifically, the real part of is taken to obtain the FrFT domain real signal

[0123] Step 105: determining a target order and a target statistical distribution.

[0124] Step 105 specifically includes:

[0125] For any FrFT domain real signal:

[0126] Step 1051: performing a goodness-of-fit test of the Alpha stable distribution on the FrFT domain real signal to obtain a D value of the FrFT domain real signal under the Alpha stable distribution, denoted as a first D value.

[0127] As an optional implementation, step 1051 specifically includes:

[0128] Using the sample characteristic function method, a parameter combination under the Alpha stable distribution is determined based on the FrFT domain real signal; the parameter combination under the Alpha stable distribution includes a characteristic exponent, a symmetry parameter, a scale coefficient, and a location parameter.

[0129] Specifically, using the sample characteristic function method, the is taken as input to obtain the corresponding characteristic exponent α a , symmetry parameter β a , scale coefficient σ a , and location parameter μ a.

[0130] The probability density function under the Alpha stable distribution is determined based on the parameter combination under the Alpha stable distribution by using the direct numerical integration method.

[0131] Specifically, the parameter combination α a , β a , σ a and μ a are taken as inputs, and the probability density function f Alpha (x; α a , β a , σ a , μ a ) under the corresponding Alpha stable distribution is calculated by using the direct numerical integration method, which is simply denoted as f Alpha (x; α a ).

[0132] The numerical values of the real FrFT domain signal are arranged in ascending order, and the real FrFT domain signal is uniformly divided into a preset number of intervals.

[0133] Specifically, the numerical values in the real FrFT domain signal are arranged in ascending order, and are uniformly divided into 200 intervals with the same interval size. At this time, r = 1, c1 = 1, 2, …, 200, wherein d1 is the minimum value of , and d 201 is the maximum value of .

[0134] Based on the probability density function under the Alpha stable distribution, the theoretical frequency under the Alpha stable distribution of each interval is determined.

[0135] Specifically, the calculation formula of the theoretical frequency under the Alpha stable distribution of each interval is as follows:

[0136] Wherein, T is the vector length of ,

[0137] For any current interval:

[0138] It is judged whether the theoretical frequency under the Alpha stable distribution of the current interval is less than the preset frequency.

[0139] If yes, the interval adjacent to the current interval and the current interval are merged into one interval, the interval corresponding to the FrFT domain real signal is updated, and the method of "determining the theoretical frequency of each interval based on the probability density function under the Alpha stable distribution" is returned.

[0140] Specifically, if there is in a certain interval, the corresponding interval and the adjacent interval on both sides are merged, and for the first and last intervals, only the adjacent interval on one side needs to be merged. Then, r is re-assigned as r = r + 1, and the merged interval is recorded as c r = 1, 2, …, Q, c r The maximum value Q is determined by the interval merging condition, and the method of "determining the theoretical frequency of each interval based on the probability density function under the Alpha stable distribution" is returned.

[0141] If no, the actual frequency of the current interval is calculated.

[0142] Specifically, if there is in any interval, the actual frequency of the current interval is calculated

[0143] Based on the theoretical frequency and the actual frequency of all intervals under the Alpha stable distribution, the D value of the FrFT domain real signal under the Alpha stable distribution is calculated, and is recorded as the first D value.

[0144] Specifically, the first D value D a is calculated by the following formula:

[0145] Where max{·} represents the maximum value operation.

[0146] Step 1052: Perform the goodness-of-fit test of the asymmetric Laplace distribution on the FrFT domain real signal to obtain the D value of the FrFT domain real signal under the asymmetric Laplace distribution, which is recorded as the second D value.

[0147] As an optional implementation, step 1052 specifically includes:

[0148] Using the parameter estimation method of the asymmetric Laplace distribution, determine the parameter combination under the asymmetric Laplace distribution based on the FrFT domain real signal; the parameter combination under the asymmetric Laplace distribution includes: shape parameter, standard deviation and location parameter.

[0149] Specifically, using the parameter estimation method of the asymmetric Laplace distribution, take as input to obtain the corresponding shape parameter λ L , standard deviation σ L and location parameter μL .

[0150] Based on the definition of the asymmetric Laplace distribution and the parameter combination under the asymmetric Laplace distribution, the probability density function under the asymmetric Laplace distribution is determined.

[0151] Specifically, λ L σ L and μ L As input to the definition of the asymmetric Laplace distribution, we obtain the probability density function f under the corresponding asymmetric Laplace distribution. L (x;μ L ,σ L ,λ L ), abbreviated as f L (x;λ L ).

[0152] Arrange the values ​​of the real signals in the FrFT domain in ascending order, and divide the real signals in the FrFT domain evenly into a preset number of intervals.

[0153] Specifically, for real signals in the FrFT domain The values ​​in the array are arranged in ascending order and evenly divided into 200 intervals of equal size. Let the size of the interval be . At this time, r = 1, c1 = 1, 2, ..., 200, where d1 is The minimum value, d 201 for The maximum value.

[0154] Based on the probability density function under the asymmetric Laplace distribution, the theoretical frequency under the asymmetric Laplace distribution in each interval is determined.

[0155] Specifically, the theoretical frequency of each interval under the asymmetric Laplace distribution. The calculation formula is:

[0156] Where T is The length of the vector.

[0157] For any current interval:

[0158] Determine whether the theoretical frequency under the asymmetric Laplace distribution in the current interval is less than the preset frequency.

[0159] If so, the intervals adjacent to the current interval and the current interval will be merged into one interval, the interval corresponding to the real signal in the FrFT domain will be updated, and "theoretical frequency of each interval will be determined based on the probability density function under the asymmetric Laplace distribution" will be returned.

[0160] Specifically, if there exists a certain interval... Then, for the merging of the corresponding interval with its two adjacent intervals, for the intervals at the beginning and end, only the adjacent intervals on one side need to be merged. Afterwards, r is reassigned so that r = r + 1, and the merged interval is redefined as... c r =1,2,…,Q,c r The maximum value Q is determined by the interval merging situation, and returns "the theoretical frequency of each interval based on the probability density function under the asymmetric Laplace distribution".

[0161] If not, calculate the actual frequency of the current interval.

[0162] Specifically, if there are in any interval Then calculate the actual frequency of the current interval.

[0163] Based on the theoretical and actual frequencies under the asymmetric Laplace distribution of all intervals, calculate the D value of the real signal in the FrFT domain under the asymmetric Laplace distribution, and denot it as the second D value.

[0164] Specifically, the second D value D L The calculation formula is:

[0165] The final D values ​​of the first D values ​​of all real signals in the FrFT domain at the same preset order are summed to obtain p first total D values.

[0166] Specifically, under the same order p and Alpha stable distribution, for different pointing angles... and time τ n The first D value of the real signal in the FrFT domain Summation, denoted as The formula for summation is:

[0167]

[0168] The second D values ​​of all real signals in the FrFT domain at the same preset order are summed to obtain p second total D values.

[0169] Specifically, under the same order p and asymmetric Laplace distribution, for different pointing angles... and time τ n The second D value of the real signal in the FrFT domain Summation, denoted as The formula for summation is:

[0170]

[0171] Step 1053: Based on p first total D values ​​and p second total D values, determine the target order and target statistical distribution; the target statistical distribution is an Alpha stable distribution or an asymmetric Laplace distribution.

[0172] As an optional implementation, step 1053 specifically includes:

[0173] The preset order corresponding to the maximum value among the p first total D values ​​and p second total D values ​​is determined as the target order.

[0174] The alpha stable distribution or asymmetric Laplace distribution corresponding to the maximum value among the p first total D values ​​and p second total D values ​​is determined as the statistical distribution.

[0175] Specifically, comparing different orders of p and The size is used to obtain the maximum value. Let the corresponding order p be the optimal order p. best (i.e., the target order), and denote the corresponding statistical distribution as the optimal statistical distribution (i.e., the target statistical distribution).

[0176] Step 106: Determine the target FrFT domain real signal corresponding to each time direction signal vector.

[0177] For any time direction signal vector:

[0178] By using a fractional Fourier transform algorithm with a target order, a fractional Fourier transform is performed on the time-direction signal vector to determine the target real signal in the FrFT domain.

[0179] As an optional implementation, a fractional Fourier transform algorithm with a target order is used to perform a fractional Fourier transform on the time-direction signal vector to determine the target real signal in the FrFT domain, specifically including:

[0180] By using a fractional Fourier transform algorithm with a target order, a fractional Fourier transform is performed on the time-direction signal vector to obtain the target FrFT domain signal.

[0181] The real part of the target FrFT domain signal is taken to obtain the target FrFT domain real signal.

[0182] Specifically, for matrix X θτ Each vector in In order p best Next, As input, perform the FrFT algorithm, and calculate the real part of the result, denoted as .

[0183] Step 107: according to the target statistical distribution and all target FrFT domain real signals, determining the time bearing history map of the array receiving signal, and realizing the enhancement of the ship radiated noise.

[0184] Wherein, the time bearing history map is a time bearing history map under an Alpha stable distribution or a time bearing history map under an asymmetric Laplace distribution; the time bearing history map under the Alpha stable distribution includes a characteristic-time bearing history map and a scale-time bearing history map, and the time bearing history map under the asymmetric Laplace distribution includes a shape-time bearing history map and a standard deviation-time bearing history map.

[0185] As an optional implementation, step 107 specifically includes:

[0186] When the target statistical distribution is an Alpha stable distribution:

[0187] For any target FrFT domain real signal.

[0188] By using a sample characteristic function method, the target parameter combination under the Alpha stable distribution is determined based on the target FrFT domain real signal; the parameter combination under the Alpha stable distribution includes a target characteristic index, a target symmetric parameter, a target scale coefficient and a target position parameter.

[0189] All target characteristic indexes are represented in a gray scale map or a pseudo-color map to obtain a characteristic-time bearing history map.

[0190] All target scale coefficients are represented in a gray scale map or a pseudo-color map to obtain a scale-time bearing history map.

[0191] Specifically, when the target statistical distribution is an Alpha stable distribution, by using Alpha stable distribution target characteristic index is solved as input. Target symmetric parameter Target scale coefficient And target position parameter And The values of and contain little information in underwater acoustic signal processing, and are discarded here.

[0192] Let The parameter matrix α a And σ a are respectively output in the form of a gray scale map or a pseudo-color map as the characteristic-time bearing history map and the scale-time bearing history map under the Alpha stable distribution.

[0193] When the target statistical distribution is an asymmetric Laplace distribution:

[0194] For any target real signal in the FrFT domain:

[0195] Using the parameter estimation method of asymmetric Laplace distribution, the target parameter combination under asymmetric Laplace distribution is determined based on the target FrFT domain real signal. The target parameter combination under asymmetric Laplace distribution includes: target shape parameter, target standard deviation and target position parameter.

[0196] Represent all target shape parameters as grayscale or pseudocolor images to obtain a shape-time orientation history map.

[0197] Represent all target standard deviations as grayscale or pseudocolor images to obtain a standard deviation-time azimuth history plot.

[0198] Specifically, when the target statistical distribution is an asymmetric Laplace distribution, the asymmetric Laplace distribution parameter solution method is used to... As input, solve for the target shape parameters under the asymmetric Laplace distribution. Target standard deviation and target position parameters For underwater acoustic signals It contains very little information and is therefore omitted here.

[0199] remember and

[0200] λ L and σ L Output the shape parameter-time azimuth history plot and standard deviation-time azimuth history plot under an asymmetric Laplace distribution in grayscale or pseudocolor format.

[0201] Example 2

[0202] The ship radiated noise enhancement system in this embodiment includes:

[0203] The array receiving signal acquisition module is used to acquire the array receiving signal received by the passive sonar; the array receiving signal contains information about the ship's radiation.

[0204] The time-delay beamforming module is used to perform time-delay beamforming on the array received signal based on J1 different preset pointing angles to obtain a direction signal matrix; the direction signal matrix includes J1 direction signal vectors with different preset pointing angles; J1>1.

[0205] The signal vector determination module is used to determine the signal vector in any direction:

[0206] The direction signal vector is segmented by a time sliding window to obtain J2 time direction signal vectors with preset pointing angles and centers of different time periods.

[0207] selecting m max time direction signal vectors from the J2 time direction signal vectors, thereby obtaining m max time direction signal vectors; 1≤m max ≤J2.

[0208] a FrFT domain real signal determination module, configured to, for any to-be-processed signal vector:

[0209] performing fractional Fourier transform on the to-be-processed signal vector by using p fractional Fourier transform algorithms with different preset orders respectively, thereby determining p FrFT domain real signals.

[0210] a total D value determination module, configured to, for any FrFT domain real signal:

[0211] performing goodness-of-fit test of Alpha stable distribution on the FrFT domain real signal, thereby obtaining a D value of the FrFT domain real signal under the Alpha stable distribution, denoted as a first D value.

[0212] performing goodness-of-fit test of asymmetric Laplace distribution on the FrFT domain real signal, thereby obtaining a D value of the FrFT domain real signal under the asymmetric Laplace distribution, denoted as a second D value.

[0213] summing up final D values of the first D values of all FrFT domain real signals at the same preset order, thereby obtaining p first total D values.

[0214] summing up the second D values of all FrFT domain real signals at the same preset order, thereby obtaining p second total D values.

[0215] a target order and target statistical distribution total D value determination module, configured to determine a target order and a target statistical distribution based on the p first total D values and the p second total D values; the target statistical distribution is Alpha stable distribution or asymmetric Laplace distribution.

[0216] a target FrFT domain real signal determination module, configured to, for any time direction signal vector:

[0217] performing fractional Fourier transform on the time direction signal vector by using a fractional Fourier transform algorithm with the target order, thereby determining a target FrFT domain real signal.

[0218] The time direction history graph determination module is configured to determine a time direction history graph of the array receiving signal according to the target statistical distribution and all target FrFT domain real signals, and to realize enhancement of the ship radiated noise; the time direction history graph is a time direction history graph under an Alpha stable distribution or a time direction history graph under an asymmetric Laplace distribution; the time direction history graph under the Alpha stable distribution comprises a characteristic-time direction history graph and a scale-time direction history graph, and the time direction history graph under the asymmetric Laplace distribution comprises a shape-time direction history graph and a standard deviation-time direction history graph.

[0219] Embodiment 3

[0220] An electronic device includes:

[0221] One or more processors.

[0222] A memory device having stored thereon one or more programs.

[0223] The one or more programs, when executed by the one or more processors, cause the one or more processors to implement the ship radiated noise enhancement method as in embodiment 1.

[0224] Embodiment 4

[0225] A storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the ship radiated noise enhancement method as in embodiment 1.

[0226] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same or similar parts between the various embodiments can be mutually referred to. For the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method part.

[0227] The principles and implementation manners of the present application are described by using specific examples in the present application, and the above embodiment description is only used to help understand the method of the present application and its core idea; at the same time, for the general technical personnel in the art, according to the idea of the present application, the specific implementation manner and application range will be changed. In conclusion, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method of enhancing ship radiated noise, characterized by, The method comprises: acquiring an array receiving signal received by a passive sonar; the array receiving signal containing information of ship radiation; performing time-delay beamforming on the array receiving signal based on J1 different preset pointing angles to obtain a direction signal matrix; the direction signal matrix comprising J1 direction signal vectors with different preset pointing angles; J1>1; for any direction signal vector: performing time sliding window segmentation on the direction signal vector to obtain J2 time direction signal vectors with preset pointing angles and different time period centers; selecting m max time direction signal vectors from the J2 max time direction signal vectors, thereby obtaining m max time direction signal vectors; 1 for any to-be-processed signal vector: performing fractional Fourier transform on the to-be-processed signal vector by using p fractional Fourier transform algorithms with different preset orders respectively, so as to determine p FrFT domain real signals; for any FrFT domain real signal: performing goodness-of-fit test of Alpha stable distribution on the FrFT domain real signal to obtain a D value of the FrFT domain real signal under Alpha stable distribution, denoted as a first D value; performing goodness-of-fit test of asymmetric Laplace distribution on the FrFT domain real signal to obtain a D value of the FrFT domain real signal under asymmetric Laplace distribution, denoted as a second D value; summing up final D values of the first D values of all FrFT domain real signals with the same preset order to obtain p first total D values; summing up the second D values of all FrFT domain real signals with the same preset order to obtain p second total D values; determining a target order and a target statistical distribution based on the p first total D values and the p second total D values; the target statistical distribution is Alpha stable distribution or asymmetric Laplace distribution; for any time direction signal vector: performing fractional Fourier transform on the time direction signal vector by using a fractional Fourier transform algorithm with the target order, so as to determine a target FrFT domain real signal; determining a time-bearings history graph of the array receiving signal according to the target statistical distribution and all the target FrFT domain real signals, so as to enhance ship radiation noise; the time-bearings history graph is a time-bearings history graph under Alpha stable distribution or a time-bearings history graph under asymmetric Laplace distribution; the time-bearings history graph under Alpha stable distribution comprises a characteristic-time-bearings history graph and a scale-time-bearings history graph, and the time-bearings history graph under asymmetric Laplace distribution comprises a shape-time-bearings history graph and a standard deviation-time-bearings history graph.

2. The ship radiated noise enhancement method of claim 1, wherein, performing fractional Fourier transform on the to-be-processed signal vector by using p fractional Fourier transform algorithms with different preset orders respectively, so as to determine p FrFT domain real signals, specifically comprising: performing fractional Fourier transform on the to-be-processed signal vector by using p fractional Fourier transform algorithms with different preset orders respectively, to obtain p FrFT domain signals with preset pointing angles and time period centers; taking real parts of the FrFT domain signals to obtain the p FrFT domain real signals.

3. The ship radiated noise enhancement method of claim 1, wherein, Performing goodness-of-fit test on the FrFT domain real signal in an Alpha stable distribution to obtain a D value of the FrFT domain real signal in the Alpha stable distribution, denoted as a first D value, specifically comprising: Determining a parameter combination in the Alpha stable distribution based on the FrFT domain real signal by using a sample characteristic function method; the parameter combination in the Alpha stable distribution comprises a characteristic index, a symmetry parameter, a scale coefficient, and a position parameter; Determining a probability density function in the Alpha stable distribution based on the parameter combination in the Alpha stable distribution by using a direct numerical integration method; Arranging the numerical values of the FrFT domain real signal in ascending order, and dividing the FrFT domain real signal into a preset number of intervals uniformly; Determining a theoretical frequency in the Alpha stable distribution of each interval based on the probability density function in the Alpha stable distribution; For any current interval: Determining whether the theoretical frequency in the Alpha stable distribution of the current interval is less than a preset frequency; If yes, merging the interval adjacent to the current interval and the current interval into one interval, updating the interval corresponding to the FrFT domain real signal, and returning to "Determining a theoretical frequency of each interval based on the probability density function in the Alpha stable distribution"; If no, calculating an actual frequency of the current interval; Based on the theoretical frequencies and the actual frequencies in the Alpha stable distribution of all intervals, calculating a D value of the FrFT domain real signal in the Alpha stable distribution, and denoted as a first D value.

4. The ship radiated noise enhancement method of claim 1, wherein, Performing goodness-of-fit test on the FrFT domain real signal in an asymmetric Laplace distribution to obtain a D value of the FrFT domain real signal in the asymmetric Laplace distribution, denoted as a second D value, specifically comprising: Determining a parameter combination in the asymmetric Laplace distribution based on the FrFT domain real signal by using a parameter estimation method of the asymmetric Laplace distribution; the parameter combination in the asymmetric Laplace distribution comprises a shape parameter, a standard deviation, and a position parameter; Determining a probability density function in the asymmetric Laplace distribution based on the definition of the asymmetric Laplace distribution and the parameter combination in the asymmetric Laplace distribution; Arranging the numerical values of the FrFT domain real signal in ascending order, and dividing the FrFT domain real signal into a preset number of intervals uniformly; Determining a theoretical frequency in the asymmetric Laplace distribution of each interval based on the probability density function in the asymmetric Laplace distribution; For any current interval: Determining whether the theoretical frequency in the asymmetric Laplace distribution of the current interval is less than a preset frequency; If yes, merging the interval adjacent to the current interval and the current interval into one interval, updating the interval corresponding to the FrFT domain real signal, and returning to "Determining a theoretical frequency of each interval based on the probability density function in the asymmetric Laplace distribution"; If no, calculating an actual frequency of the current interval; Based on the theoretical frequency and the actual frequency of the asymmetric Laplace distribution of all intervals, a D value of the FrFT domain real signal under the asymmetric Laplace distribution is calculated, and is recorded as a second D value.

5. The method of claim 1, wherein, Based on the p first total D values and the p second total D values, a target order and a target statistical distribution are determined, specifically including: A preset order corresponding to the maximum of the p first total D values and the p second total D values is determined as the target order; An Alpha stable distribution or an asymmetric Laplace distribution corresponding to the maximum of the p first total D values and the p second total D values is determined as the statistical distribution.

6. The ship radiated noise enhancement method of claim 1, wherein, A fractional Fourier transform algorithm with the target order is used to perform fractional Fourier transform on the time direction signal vector, so as to determine a target FrFT domain real signal, specifically including: A fractional Fourier transform algorithm with the target order is used to perform fractional Fourier transform on the time direction signal vector, so as to obtain a target FrFT domain signal; The real part of the target FrFT domain signal is taken to obtain the target FrFT domain real signal.

7. The method of claim 1, wherein, Based on the target statistical distribution and all the target FrFT domain real signals, a time bearing history graph of the array receiving signal is determined, specifically including: When the target statistical distribution is an Alpha stable distribution: For any target FrFT domain real signal: A sample characteristic function method is used to determine a target parameter combination under the Alpha stable distribution based on the target FrFT domain real signal; the parameter combination under the Alpha stable distribution includes a target characteristic index, a target symmetry parameter, a target scale coefficient and a target position parameter; All the target characteristic indexes are represented in a gray scale graph or a pseudo-color graph to obtain a characteristic-time bearing history graph; All the target scale coefficients are represented in a gray scale graph or a pseudo-color graph to obtain a scale-time bearing history graph; When the target statistical distribution is an asymmetric Laplace distribution: For any target FrFT domain real signal: A parameter estimation method of the asymmetric Laplace distribution is used to determine a target parameter combination under the asymmetric Laplace distribution based on the target FrFT domain real signal; the target parameter combination under the asymmetric Laplace distribution includes a target shape parameter, a target standard deviation and a target position parameter; All the target shape parameters are represented in a gray scale graph or a pseudo-color graph to obtain a shape-time bearing history graph; All the target standard deviations are represented in a gray scale graph or a pseudo-color graph to obtain a standard deviation-time bearing history graph.

8. A ship radiated noise enhancement system, characterized by, The system includes: An array receiving signal acquisition module is configured to acquire an array receiving signal received by a passive sonar; the array receiving signal contains information of ship radiation; A time delay beam forming module is configured to perform time delay beam forming on the array receiving signal based on J1 different preset pointing angles to obtain a direction signal matrix; the direction signal matrix includes J1 direction signal vectors with different preset pointing angles; J1>1; A to-be-processed signal vector determination module is configured to, for any direction signal vector: time-sliding segmenting the direction signal vector to obtain J2 time direction signal vectors with preset pointing angles and different time segment centers; selecting m max time direction signal vectors from the J2 max time direction signal vectors, thereby obtaining m max time direction signal vectors; 1≤m a FrFT domain real signal determination module configured to determine, for any of the to-be-processed signal vectors: performing FrFT on the to-be-processed signal vectors respectively by using p FrFT algorithms with different preset orders, thereby determining p FrFT domain real signals; a total D value determination module configured to determine, for any of the FrFT domain real signals: performing goodness-of-fit test of Alpha stable distribution on the FrFT domain real signal to obtain a D value of the FrFT domain real signal under the Alpha stable distribution, denoted as a first D value; performing goodness-of-fit test of asymmetric Laplace distribution on the FrFT domain real signal to obtain a D value of the FrFT domain real signal under the asymmetric Laplace distribution, denoted as a second D value; summing up the final D values of the first D values of all the FrFT domain real signals with the same preset order to obtain p first total D values; summing up the second D values of all the FrFT domain real signals with the same preset order to obtain p second total D values; a target order and target statistical distribution total D value determination module configured to determine a target order and a target statistical distribution based on the p first total D values and the p second total D values; the target statistical distribution is Alpha stable distribution or asymmetric Laplace distribution; a target FrFT domain real signal determination module configured to determine, for any of the time direction signal vectors: performing FrFT on the time direction signal vector by using the FrFT algorithm with the target order, thereby determining a target FrFT domain real signal; a time bearing history map determination module configured to determine a time bearing history map of the array receiving signal according to the target statistical distribution and all the target FrFT domain real signals, thereby achieving enhancement of the ship radiated noise; the time bearing history map is a time bearing history map under Alpha stable distribution or a time bearing history map under asymmetric Laplace distribution; the time bearing history map under Alpha stable distribution includes a characteristic-time bearing history map and a scale-time bearing history map, and the time bearing history map under asymmetric Laplace distribution includes a shape-time bearing history map and a standard deviation-time bearing history map.

9. An electronic device, comprising: comprise: one or more processors; a storage device having one or more programs stored thereon; when the one or more programs are executed by the one or more processors, the one or more processors implement the ship radiated noise enhancement method according to any one of claims 1 to 7.

10. A storage medium, characterized by a computer program is stored thereon, wherein the computer program is executed by a processor to implement the ship radiated noise enhancement method according to any one of claims 1 to 7.

Images