Multi-physical field joint target detection method and system based on underwater low-frequency seismic wave field and sound field signals

Abstract

The invention discloses a multi-physical-field joint target detection method based on underwater low-frequency seismic wave field and sound field signals. The method comprises the following steps: acquiring underwater low-frequency seismic signals x1 (t) and sound field signals x2 (t) through a sensor; performing cross-correlation spectrum calculation on the seismic wave signal and the sound field signal in the fixed time window t0 to obtain a frequency spectrum Xm; removing a direct current component in the frequency spectrum Xm to obtain a direct current component-removed FFT sequence XFFT; the XFFT data of the direct-current component FFT sequence in the third step are preprocessed; the detection threshold value Xt of # imgabs0 # is calculated; comparing # imgabs1 # with Xt, and judging a suspected target line spectrum frequency point according to a difference value T of # imgabs1 # and Xt: performing time integration on the difference value T, and judging that a suspected target appears if the maximum frequency Q of the suspected line spectrum frequency point in a time window is greater than a set value R; if the suspected target appears within a plurality of continuous seconds, judging that a target signal appears; suspected targets can be eliminated and the targets can be quickly locked under complex environment interference, and a basis is provided for underwater weapon development and underwater multi-physics field target recognition.

Description

Technical Field

[0001] The present invention belongs to the field of target detection methods, and particularly relates to a multi-physical field joint target detection method based on underwater low-frequency seismic wave fields and acoustic field signals. Background Art

[0002] The physical field of a ship refers to different physical fields generated in the surrounding space during the ship's navigation, including electric fields, magnetic fields, acoustic fields, hydrostatic pressure fields, and seismic fields, etc. These physical fields are the spatial distributions around the ship caused by the ship's generation or the change of physical field parameters in the surrounding environment induced by the ship's movement. Their variation laws can well reflect the characteristics of target signals and are important information sources for underwater target detection and identification.

[0003] With the continuous improvement of the demand for ocean exploration, underwater target detection technology has become one of the key technologies in the fields of ocean science research, underwater safety protection, and ocean resource development. Underwater detection technology has developed rapidly in the past few decades. There are various applications of underwater target recognition technology, including fish school detection, seabed detection, underwater vehicle detection, etc. For ship mechanical waves, in previous studies, researchers have proposed many methods for extracting underwater target features. These methods can generally be divided into three aspects: time domain, frequency domain, and time-frequency domain. Time-frequency analysis combines the advantages of time domain and frequency domain analysis and is more suitable for the analysis of non-stationary signals, and has received extensive research. Currently, common signal analysis methods include Hilbert transform, wavelet transform, modal decomposition, entropy value calculation, etc. With the development of deep learning, feature extraction algorithms based on deep learning have also been widely applied. Most deep learning methods tend to use time-frequency signals for analysis, including support vector machine (SVM), Hidden Markov Model (HMM), and Convolutional Neural Network (CNN), etc. However, deep learning requires a large amount of data support. From the current research status, it is found that there is less research on underwater signal data, resulting in a lower system generalization ability and an inability to accurately capture target signals. In summary, the above signal analysis only focuses on a single underwater physical field, and the difficulty of detecting stealth targets in the complex underwater background environment by relying on the information of a single underwater physical field is increasing day by day. Summary of the Invention

[0004] The present invention aims to solve the problem in the prior art that only a single underwater physical field is concerned, and it is difficult to detect stealth targets in the complex underwater background environment by relying on the information of a single underwater physical field.

[0005] The present invention adopts the following scheme: A multi-physical field joint target detection method based on underwater low-frequency seismic wave fields and acoustic field signals, comprising the following steps:

[0006] Step 1: Collect the underwater low-frequency seismic signal x1(t) and the acoustic field signal x2(t) through sensors;

[0007] Step 2: Calculate the cross-correlation spectrum of the seismic wave signal and the acoustic field signal within a fixed time window t0 to obtain the frequency spectrum X m ;

[0008] Step 3: Remove the DC component in the frequency spectrum X m to obtain the DC-removed FFT sequence X FFT ,

[0009] Step 4: Preprocess the data of the DC component FFT sequence X FFT in Step 3;

[0010] Step 5: Calculate the detection threshold value X of t , X t being the adaptive threshold value;

[0011] Step 6: Compare with X t in size, and judge the suspected target line spectrum frequency points according to their difference T:

[0012] Step 7: Perform time integration on T = [T0, T1, …, T N to improve the detection reliability, and the integration feature P is as follows:

[0013]

[0014] In the formula, [P0, P1, …, P N are the integration features of each frequency point respectively; M is the integration duration, and Y Nt is the suspected target quantity of the Nth frequency point at time t; if the maximum frequency Q of the suspected line spectrum frequency points appearing within the time window is greater than the set value R, it is judged that the suspected target appears; if the suspected target appears continuously for several seconds, it is judged that the target signal appears.

[0015] Furthermore: In Step 1, the sensors include a seismic wave sensor and a sound sensor. The seismic wave sensor and the sound sensor are placed in the middle of the waterway, and the MR6000 data acquisition instrument is used on the shore to collect signals in real time.

[0016] Furthermore: Step 4 specifically includes: Perform energy normalization on X FFT to improve the dynamic observation range. When the ship sails from far to near through the detection point, the energy difference can reach dozens to hundreds of times. Within the first 20 s of the detection signal, since the ship is far from the detector, the energy range changes very little, and the normalized sequence As shown in Equation (4):

[0017]

[0018] Furthermore, the specific method of Step Six is as follows:

[0019]

[0020] In the formula, X i is the i-th element in the FFT sequence X FFT after removing the DC component, and X ti is the i-th element in the adaptive threshold sequence X t ; T = [T0, T1, …, T N is the difference between the target signal and the threshold, and T N represents the difference between the detection quantity at the N-th frequency point and the threshold; if T N ≥0, then the detection identification quantity Y i = 1, and this frequency point is a suspected target line spectrum frequency point; otherwise Y i = 0, and this frequency point is not a suspected target line spectrum frequency point.

[0021] Furthermore, in the cross-correlation spectrum calculation of Step Two, it also includes: the selection of the time window t0 will affect the result; by changing the time window size and conducting multiple experiments, comparing the differences in the spectrum Xm, the optimal time window is determined.

[0022] Furthermore, the calculation method of X t in Step Five is as follows: X t is the sum of the adaptive floating threshold and the fixed threshold:

[0023]

[0024] In the formula, is the median filtering operation, and the adaptive threshold is calculated according to the ocean environment noise spectrum; Δ = [Δ0, Δ1, …, Δ N is the fixed threshold vector, which can be obtained through simulation calculation. Assuming that the ocean environment background noise is Gaussian noise, the initial Δ i = 0; i = 0, 1, …, N,

[0025] In order to improve the adaptability to different sea areas and reduce the false alarm probability, the fixed threshold Δ can be set to the ocean background noise of the current actual environment; according to Steps One to Five, the variance σ of the Gaussian noise signal X t is statistically calculated, and the fixed threshold Δ = [Δ0, Δ1, …, Δ N is set according to the 3σ criterion, where Δ i = 3σ; the threshold weight is adjusted according to the real-time ocean environment parameters; compared with the traditional signal detection algorithm, during the detection process, X FFTis changing in real time, so X t is changing, that is, X t is an adaptive floating threshold.

[0026] Furthermore, during the sensor acquisition process of the first step, a real-time calibration method is also included: using a calibration source with a known frequency to dynamically adjust the sensor gain and phase offset, and the calibrated signal xcalibrated(t) = α(t)·x raw (t) + β(t), where α(t) and β(t) are optimized online by the least squares method, and where x raw is the original acquisition data of x1(t) or x2(t).

[0027] As another aspect of the present invention, it also relates to a multi-physical field joint target detection system based on underwater low-frequency seismic wave fields and acoustic field signals, including:

[0028] A signal acquisition unit: used to collect underwater low-frequency seismic signal x1(t) and acoustic field signal x2(t) through sensors;

[0029] A cross-spectrum calculation unit: used to perform cross-spectrum calculation on the seismic wave signal and the acoustic field signal within a fixed time window t0 to obtain the spectrum X m

[0030] A DC component removal unit: used to remove the DC component in the spectrum X m to obtain the DC component removed FFT sequence X FFT ;

[0031] A data preprocessing unit: used to preprocess the DC component FFT sequence X FFT data in step three;

[0032] An adaptive threshold calculation unit: calculates the detection threshold X t as the adaptive threshold, which is the sum of the adaptive floating threshold and the fixed threshold:

[0033]

[0034] In the formula, is the median filtering operation, and the adaptive threshold can be calculated according to the ocean environment noise spectrum; Δ = [Δ0, Δ1,..., Δ N is the fixed threshold vector, which can be obtained through simulation calculation. Assuming that the ocean environment background noise is Gaussian noise, the initial Δ i = 0; i = 0, 1,..., N, the variance σ of the Gaussian noise signal X t is statistically calculated, and the fixed threshold Δ = [Δ0, Δ1,..., Δ N is set according to the 3σ criterion, where Δi = 3σ; relative to the traditional signal detection algorithm, X is FFT changing in real time during the detection process, so X t is changing, that is, X t is the adaptive threshold;

[0035] Target line spectrum frequency point judgment unit: used to compare with X t in size, and judge the suspected target line spectrum frequency point according to the difference value T;

[0036] Target signal appearance judgment unit: used to perform time integration on T = [T0, T1, …, T N , to further improve the reliability of detection, and the integration feature P is as follows:

[0037]

[0038] In the formula, [P0, P1, …, P N are the integration features of each frequency point respectively; M is the integration duration, and Y Nt is the suspected target quantity of the Nth frequency point at time t; if the maximum frequency Q of the suspected line spectrum frequency point appearance within the time window is greater than the set value R, it is judged that the suspected target appears; if the suspected target appears continuously for several seconds, it is judged that the target signal appears.

[0039] As another aspect of the present invention, it also relates to a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to perform the above-mentioned multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals.

[0040] Generally speaking, compared with the prior art through the above technical solutions conceived by the present invention, the following beneficial effects can be achieved:

[0041] (1) The multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals of the present invention uses the cross-spectrum line spectrum detection method to detect target ships. Compared with single-signal detection, the unknown background interference is significantly reduced, and it can eliminate suspected targets and quickly lock the target under complex environmental interference, providing a basis for the development of underwater weapons and the identification of underwater multi-physical field targets.

[0042] (2) The multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals of the present invention combines the strong correlation between the seismic wave field and the acoustic field to improve the accuracy of target detection. Description of the Drawings

[0043] Figure 1 It is a schematic diagram of the overall process of a preferred embodiment of the invention;

[0044] Figure 2 The line spectrum diagram of the seismic wave field with a target speed of 3 kn for the preferred embodiment of the invention;

[0045] Figure 3 The line spectrum diagram of the acoustic field with a target speed of 3 kn for the preferred embodiment of the invention;

[0046] Figure 4 The cross-correlation line spectrum of the seismic field and the acoustic field corresponding to a target speed of 3 kn for the preferred embodiment of the invention;

[0047] Figure 5 The line spectrum of the seismic wave field with a target speed of 8 kn for the preferred embodiment of the invention;

[0048] Figure 6 The line spectrum diagram of the acoustic field with a target speed of 8 kn for the preferred embodiment of the invention;

[0049] Figure 7 The cross-correlation line spectrum of the seismic field and the acoustic field corresponding to a target speed of 8 kn for the preferred embodiment of the invention. Detailed implementation manners

[0050] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0051] Please refer to Figure 1 , the present invention relates to a multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals, and includes the following steps:

[0052] Step 1: Collect underwater low-frequency seismic signal x1(t) and acoustic field signal x2(t) through sensors; the sensors include a seismic wave sensor (force balance accelerometer) and an acoustic sensor (piezoelectric sensor), and the seismic wave sensor and the acoustic sensor are placed in the middle of the waterway, and the signal is collected in real time by the MR6000 data collector on the shore.

[0053] In some other preferred embodiments, before the cross-correlation spectrum calculation in Step 2, x1(t) and x2(t) can also be subjected to wavelet noise reduction to retain the low-frequency target line spectrum. For example, the Morlet wavelet basis can be selected to match the low-frequency vibration characteristics of the ship. This can improve the signal-to-noise ratio and reduce the false alarm rate in Step 5.

[0054] In some other preferred embodiments, considering that underwater sensors are vulnerable to temperature and pressure changes, resulting in signal drift (such as the acquisition errors of x1(t) and x2(t) in Step 1)), during the sensor acquisition process in Step 1, a real-time calibration method is further included: using a calibration source with a known frequency to dynamically adjust the sensor gain and phase offset, and the calibrated signal xcalibrated(t) = α(t).x raw (t) + β(t), where α(t) and β(t) are optimized online by the least squares method, and where x raw (t) is the original acquisition data of x1(t) or x2(t).

[0055] Step 2: Calculate the cross-correlation spectrum of the seismic wave signal and the sound field signal within a fixed time window t0 to obtain the spectrum X m ;

[0056] In some preferred embodiments, in the cross-correlation spectrum calculation, it further includes: considering that the selection of the time window t0 will affect the result; conducting multiple experiments by changing the time window size, comparing the differences in the spectra Xm, and determining the optimal time window.

[0057] Step 3: Remove the DC component in the spectrum X m to obtain the DC component removed FFT sequence X FFT , as shown in Equation (3):

[0058] X FFT = [X0, X1, …, X N (3)

[0059] In the formula, N is the number of frequency points, and the frequency corresponding to X i is F i , i = 0, 1, …, N;

[0060] Step 4: Preprocess the DC component removed FFT sequence X FFT data in Step 3; specifically: perform energy normalization on X FFT to improve the dynamic observation range (when the ship sails from far to near through the detection point, the energy difference can reach dozens to hundreds of times, while the energy change is not significant within an observation time of about 20 s), and obtain the normalized sequence as shown in Equation (4):

[0061]

[0062] Step 5: Calculate the detection threshold value X of t is an adaptive threshold value, which is the sum of an adaptive floating threshold value and a fixed threshold value:

[0063]

[0064] In the formula, is the median filtering operation, and the adaptive threshold can be calculated according to the ocean environment noise spectrum; Δ = [Δ0, Δ1, …, Δ N is the fixed threshold vector, which can be obtained through simulation calculation. Assuming that the ocean environment background noise is Gaussian noise and the initial Δ i = 0; i = 0, 1, …, N, the variance σ of the Gaussian noise signal X t is statistically calculated, and the fixed threshold Δ = [Δ0, Δ1, …, Δ N is set according to the 3σ criterion, where Δ i = 3σ; compared with the traditional signal detection algorithm, X FFT is changing in real time during the detection process, so X t is changing, that is, X t is the adaptive threshold;

[0065] In some other preferred embodiments, to improve the adaptability to different sea areas and reduce the false alarm rate, the threshold formula can be further extended to where E env is the environmental parameter, and α, β, γ are dynamic weight coefficients, and the threshold weight is adjusted according to the real-time ocean environmental parameters (such as temperature, salinity, flow velocity).

[0066] Step Six: Comparison and X t in size, and judge the suspected target line spectrum frequency point according to their difference T: The specific method is:

[0067] In the formula, X i is the i-th element in the FFT sequence X FFT after removing the DC component, X ti is the i-th element in the adaptive threshold value sequence X t , T = [T0, T1, …, T N is the difference between the target signal and the threshold value, and T N represents the difference between the detection amount at the N-th frequency point and the threshold; if T N ≥ 0, then there is a detection identification quantity Y i = 1, and this frequency point is a suspected target line spectrum frequency point; otherwise Y i = 0, and this frequency point is not a suspected target line spectrum frequency point.

[0068] Step Seven: Integrate T = [T0, T1, …, T N over time to further improve the reliability of the detection. The integration feature P is as follows:

[0069]

[0070] In the formula, [P0, P1, …, P N are the integral features of each frequency point; M is the integral duration, and Y Nt is the suspected target quantity of the Nth frequency point at time t; if the maximum frequency Q of the suspected line spectrum frequency point appearing within the time window is greater than the set value R, it is determined that a suspected target appears; if a suspected target appears continuously for several seconds, it is determined that a target signal appears.

[0071] In some other preferred embodiments, in step seven, clustering analysis can also be performed on the suspected target frequency points to screen out a frequency point group that conforms to the ship vibration harmonic distribution (fundamental frequency + multiple frequency); if the energy proportion of the frequency point group after clustering exceeds the threshold, it is determined as a target signal.

[0072] As another aspect of the present invention, it also relates to a multi-physical field joint target detection system based on underwater low-frequency seismic wave fields and acoustic field signals, including:

[0073] Signal acquisition unit: used to collect underwater low-frequency seismic signal x1(t) and acoustic field signal x2(t) through sensors;

[0074] Cross-spectrum calculation unit: used to perform cross-spectrum calculation on the seismic wave signal and the acoustic field signal within a fixed time window t0 to obtain the frequency spectrum X m

[0075] DC component removal unit: used to remove the DC component in the frequency spectrum X m to obtain the FFT sequence X without DC component FFT ;

[0076] Data preprocessing unit: used to preprocess the FFT sequence X of the DC component in step three FFT data;

[0077] Adaptive threshold calculation unit: calculate the detection threshold X t as the adaptive threshold, which is the sum of the adaptive floating threshold and the fixed threshold:

[0078]

[0079] In the formula, is the median filtering operation, and the adaptive threshold can be calculated according to the ocean environment noise spectrum; Δ = [Δ0, Δ1, …, Δ N is the fixed threshold vector, which can be obtained through simulation calculation. Assuming that the ocean environment background noise is Gaussian noise and the initial Δ i = 0; i = 0, 1, …, N, the variance σ of the Gaussian noise signal X t is statistically calculated, and the fixed threshold Δ = [Δ0, Δ1, …, Δ is set according to the 3σ criterionN , where Δ i = 3σ; relative to the traditional signal detection algorithm, X FFT is changing in real time during the detection process, so X t is changing, that is, X t is the adaptive threshold;

[0080] Target line spectrum frequency point judgment unit: used to compare with X t in size, and judge the suspected target line spectrum frequency point according to the difference T;

[0081] Target signal appearance judgment unit: used to perform time integration on T = [T0, T1,..., T N , to further improve the reliability of detection. The integration feature P is as follows:

[0082]

[0083] In the formula, [P0, P1,..., P N are the integration features of each frequency point respectively; M is the integration duration, and Y Nt is the suspected target quantity of the Nth frequency point at time t; if the maximum frequency Q of the suspected line spectrum frequency point appearance within the time window is greater than the set value R, it is judged that the suspected target appears; if the suspected target appears continuously for several seconds, it is judged that the target signal appears.

[0084] As another aspect of the present invention, it also relates to a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to perform the above multi-physical field joint target detection method based on underwater low-frequency seismic wave fields and acoustic field signals.

[0085] Effect verification:

[0086] To verify the effectiveness of the above method, the seismic wave field and acoustic field time-domain signals when the target ship A passes at 3 kn and 8 kn are respectively detected. The line spectrum detection of single signals and the cross-correlation line spectrum detection are respectively performed using the time-domain signals. The results are as Figure 2 , Figure 3 shown. It can be found that when the target ship passes at a speed of 3 kn, Figure 2 in the seismic wave field signal, there is an obvious target line spectrum at the frequency point of 27.4 Hz, and the duration is about 60 s, which conforms to the target passing time, and its second harmonic frequency signal of 55.02 Hz is also very obvious. In addition, there are other frequency signal line spectra. And Figure 3 in the acoustic field line spectrum, the signal line spectrum at the frequency point of 27.5 Hz is also very obvious, but the line spectra of other interference signals are also very obvious. Therefore, single signals cannot accurately judge suspicious targets without knowing the characteristics of the target signal;

[0087] In order to capture the target more accurately, the cross-spectrum of the seismic wave field and the acoustic field signal is calculated and then line spectrum detection is carried out. The results are as follows Figure 4 shown. At the frequency point of 27.5 Hz, the target signal is obvious, the duration is about 60 s, and the unknown interference sources other than the target signal are significantly reduced, so that the target can be determined quickly and effectively.

[0088] In order to further verify the effectiveness of the algorithm under the condition of high-speed movement of the target, the line spectrum is calculated by using the seismic wave field and the time-domain signal of the acoustic field with the ship speed of 8 kn of the target to be measured. As shown in Figure 4 、 Figure 5 shown. Compared with the low-speed movement state of the target, the suspected line spectra are significantly reduced under the high-speed movement condition. Figure 4 In the seismic wave field signal in , there is an obvious target line spectrum at the frequency point of 47.78 Hz, but the duration of the target signal is reduced, indicating that there may be a problem of short target detection distance for a single signal; Figure 5 In the acoustic field signal in , there is also an obvious target line spectrum at the frequency point of 47.78 Hz. In addition, there are obvious line spectra at other frequency points, which may lead to a high false alarm rate in target detection. Therefore, the cross-spectrum of the seismic wave field and the acoustic field signal is calculated and then line spectrum detection is carried out. The results are as follows Figure 7 shown. It can be seen from Figure 7 that after cross-spectrum calculation and line spectrum detection, the suspected target line spectra are significantly reduced. There is an obvious line spectrum feature at the frequency point of 47.78 Hz of the target signal, and the signal duration is about 60 s, which is in line with the target passing time. At the same time, compared with the line spectrum detection of a single signal, the unknown background interference is significantly reduced. Through the above experimental verification, the method of the present invention can effectively and quickly determine the target and reduce the false alarm rate of target detection.

[0089] The present invention proposes a real-time cross-spectrum target detection method based on the correlation between underwater low-frequency seismic wave fields and acoustic fields. According to the co-source characteristics of signals, the cross-spectrum line spectrum detection method is used to detect target ships. Compared with single-signal detection, the unknown background interference is significantly reduced, and suspected targets can be excluded and the target can be quickly locked under complex environmental interference, which provides a basis for the development of underwater weapons and the identification of targets in underwater multi-physical fields.

[0090] It is easy for those skilled in the art to understand that the above is only a preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals, characterized in that, It includes the following steps: Step 1: Collect the underwater low-frequency seismic signal x1(t) and the acoustic field signal x2(t) through sensors; Step 2: Calculate the cross-correlation spectrum of the seismic wave signal and the sound field signal within a fixed time window t0 to obtain the spectrum X m ; Step 3: Remove the DC component from spectrum X m to obtain the DC-removed FFT sequence X FFT , Step 4: Preprocess the DC component FFT sequence X FFT data in Step 3; Step Five: Calculate the detection threshold value X t of t , where X is an adaptive threshold value; Step 6: Compare with X t in terms of magnitude, and based on the difference value T, determine the suspected target line spectrum frequency points: Step 7: Integrate the time of T = [T0, T1, …, T N to improve the reliability of detection. The integrated feature P is as follows: wherein, [P0, P1, …, P N are the integral features of each frequency point; M is the integral duration, and Y Nt is the suspected target quantity of the Nth frequency point at time t; If the maximum frequency Q of the suspected line spectrum frequency points appearing within the time window is greater than the set value R, it is determined that a suspected target appears; If suspected targets appear continuously for several seconds, it is determined that a target signal appears.

2. The multi-physical-field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals according to claim 1, characterized in that: In the said Step 1, the sensors include a seismic wave sensor and an acoustic sensor. The seismic wave sensor and the acoustic sensor are placed in the middle of the waterway, and the MR6000 data acquisition instrument is used on the shore to collect the signals in real time.

3. The multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals according to claim 1, characterized in that: Step 4 specifically includes: For X FFT Perform energy normalization to improve the dynamic observation range. When the ship sails from far to near and passes through the detection point, the energy difference can reach dozens to hundreds of times. Within the first 20 s of the detection signal, since the ship is far from the detector, the change in its energy range is very small, and a normalized sequence is shown in the following formula (4):

4. The multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals according to claim 1, characterized in that: The specific method of the said Step 6 is: where X i is the i-th element in the FFT sequence X FFT after removing the DC component, X ti is the i-th element in the adaptive threshold sequence X t , T = [T0, T1, …, T N is the difference between the target signal and the threshold, T N represents the difference between the detection quantity at the N-th frequency point and the threshold; if T N ≥0, then there is a detection identification quantity Y i = 1, and this frequency point is a suspected target line spectrum frequency point; otherwise Y i = 0, and this frequency point is not a suspected target line spectrum frequency point.

5. The multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals according to claim 1, wherein: In the calculation of the cross-correlation spectrum in the said Step 2, it also includes: The selection of the time window t0 will affect the result; Through multiple experiments by changing the size of the time window and comparing the differences in the spectrum Xm, the optimal time window is determined.

6. The multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals according to claim 1, wherein: The X in step five t is calculated as follows: X t is the sum of the adaptive floating threshold and the fixed threshold: In the formula, is the median filtering operation, and the adaptive threshold is calculated according to the ocean environment noise spectrum; Δ = [Δ0, Δ1, …, Δ N is the fixed threshold vector, which can be obtained through simulation calculation. Assuming that the ocean environment background noise is Gaussian noise and the initial Δ i = 0; i = 0, 1, …, N, To improve the adaptability in different sea areas and reduce the false alarm probability, the fixed threshold Δ is set to the ocean background noise of the current actual environment; the variance σ of the Gaussian noise signal X is statistically calculated according to steps one to five, and the fixed threshold Δ = [Δ0, Δ1, …, Δ t is set according to the 3σ criterion, where Δ N = 3σ; the threshold weight is adjusted according to the real-time ocean environment parameters; compared with the traditional signal detection algorithm, X i is changing in real time during the detection process, so X FFT is changing, that is, X t is an adaptive floating threshold. t ​ 7. The multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals according to any one of claims 1-6, wherein: During the acquisition process of the first-step sensor, a real-time calibration method is also included: using a calibration source with a known frequency to dynamically adjust the sensor gain and phase offset, and the calibrated signal xcalibrated(t) = α(t).x raw (t) + β(t), where α(t) and β(t) are coefficients optimized online by the least squares method, and where x raw (t) is the original acquisition data of x1(t) or x2(t).

8. A multi-physical field joint target detection system based on underwater low-frequency seismic wave fields and acoustic field signals, characterized in that It includes: A signal acquisition unit: used to collect the underwater low-frequency seismic signal x1(t) and the acoustic field signal x2(t) through sensors; Cross-spectrum calculation unit: used to perform cross-spectrum calculation on the seismic wave signal and the sound field signal within a fixed time window t0 to obtain the spectrum X m DC component removal unit: used to remove the DC component in spectrum X m to obtain the DC-removed FFT sequence X FFT ; Data preprocessing unit: used to preprocess the DC component FFT sequence X in step three FFT data; Adaptive threshold calculation unit: Calculate the detection threshold X t as the adaptive threshold, which is the sum of the adaptive floating threshold and the fixed threshold: In the formula, is the median filtering operation, and the adaptive threshold can be calculated according to the ocean environment noise spectrum; Δ = [Δ0, Δ1, …, Δ N is the fixed threshold vector, which can be obtained through simulation calculation. Assuming that the ocean environment background noise is Gaussian noise and the initial Δ i = 0; i = 0, 1, …, N, the variance σ of the Gaussian noise signal X t is statistically calculated, and the fixed threshold Δ = [Δ0, Δ1, …, Δ N is set according to the 3σ criterion, where Δ i = 3σ; compared with the traditional signal detection algorithm, X FFT is changing in real time during the detection process, so X t is changing, that is, X t is the adaptive threshold; Target line spectrum frequency point judgment unit: used to compare with X t in size, and judge the suspected target line spectrum frequency point according to the difference value T; Target signal appearance judgment unit: used to perform time integration on T = [T0, T1, …, T N , to further improve the reliability of detection. The integration feature P is as follows: wherein, [P0, P1, …, P N are the integral features of each frequency point; M is the integral duration, and Y Nt is the suspected target quantity of the Nth frequency point at time t; If the maximum frequency Q of the suspected line spectrum frequency points appearing within the time window is greater than the set value R, it is determined that a suspected target appears; If suspected targets appear continuously for several seconds, it is determined that a target signal appears.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program executes the multi-physical field joint target detection method based on underwater low-frequency seismic wave field and acoustic field signals according to any one of claims 1-7 by the processor.

Images