Underwater explosive sound signal interception method and device

Abstract

The invention relates to the field of underwater sound signal processing, and discloses an underwater explosive sound signal interception method and device, and the method comprises the steps: determining a signal rough reading starting moment and a signal rough reading ending moment based on the reference starting moment of an explosive sound signal obtained through a bomb dropping moment recording table, a set error duration, and an explosive sound signal reference duration, reading data from the underwater explosive sound signal file; performing time-frequency transformation on the read data based on a first window function, and calculating signal energy of each first window; based on the signal energy maximum value, the ratio of the explosion starting window signal energy to the signal energy maximum value and a set first threshold value, the explosion sound signal starting moment is determined; and intercepting the underwater explosive sound signal based on the determined starting moment of the underwater explosive sound signal. According to the method, starting and ending time of the explosive sound signal is automatically searched through energy comparison based on an analysis design threshold value of a time-frequency energy spectrum, manpower and time cost is saved, intercepted signals are standardized, and noise interference is low.

Description

Technical Field

[0001] This invention relates to the field of underwater acoustic signal processing, and in particular to a method and apparatus for intercepting underwater explosion acoustic signals. Background Technology

[0002] With the continuous development of underwater acoustic engineering and marine technology, the demand for marine test data is increasing. Underwater explosion sound sources have advantages such as high power, wide bandwidth, and omnidirectionality, and are widely used in fields such as underwater acoustic propagation loss acquisition, underwater acoustic channel measurement, and marine environment inversion. Currently, the types and quantities of explosion sound sources and receiving equipment used in marine tests are numerous, and a large amount of underwater explosion sound signal data needs to be effectively extracted before processing and analysis. Since underwater explosion sound signals are difficult to identify through matched filtering and other methods, conventional extraction of underwater explosion sound signals mostly relies on manual identification, which is costly in terms of human resources and time. In addition, to reduce noise interference and achieve data standardization, signals extracted based on manual identification have significant errors, seriously reducing data quality. Therefore, it is particularly necessary to design an underwater explosion sound signal interception method to achieve rapid and accurate extraction of underwater explosion sound signals. Summary of the Invention

[0003] The purpose of this invention is to overcome the defects of the prior art and to propose a method for intercepting underwater explosion sound signals. This invention also discloses an underwater explosion sound signal interception device.

[0004] This invention provides a method for intercepting underwater explosion acoustic signals, comprising:

[0005] Step 1: Based on the reference start time of the explosion sound signal obtained from the bombing time record table, the set error duration, and the reference duration of the explosion sound signal, determine the start time and end time of the coarse reading of the signal;

[0006] Step 2: Read data from the underwater explosion sound signal file according to the determined start and end times of the signal coarse reading;

[0007] Step 3: Perform time-frequency transformation on the read data based on the first window function, and calculate the signal energy of each first window, where the first window corresponds to the first window function;

[0008] Step 4: Based on the relationship between the maximum signal energy of the first window, the ratio of the signal energy of the explosion start window to the maximum signal energy of the first window, and the set first threshold, determine the start time of the explosion sound signal, wherein the explosion start window is the first window where the explosion start time is located;

[0009] Step 5: Based on the determined start time of the underwater explosion sound signal, extract the underwater explosion sound signal from the underwater explosion sound signal file.

[0010] In an improved method for intercepting underwater explosion acoustic signals, after step 4 and before step 5, the method further includes:

[0011] Based on the reference duration of the explosion sound signal and the determined start time of the underwater explosion sound signal, calculate the signal reference average energy based on the first window;

[0012] The data is transformed by time and frequency based on the second window function, and the average energy of the signal in each second window is calculated. The second window corresponds to the second window function, and the moving step size of the second window is larger than that of the first window.

[0013] Calculate the ratio of the average signal energy of each second window to the average signal energy referenced based on the first window, and take the second window to which the average signal energy with a ratio less than a set second threshold belongs as the second window where the explosion sound signal ends.

[0014] Calculate the average energy of the signal in each of the first windows within the second window at the moment the explosion sound signal ends;

[0015] The ratio of the average energy of the signal in each first window within the second window at the time the explosion sound signal ends to the average energy of the signal reference based on the first window is calculated, and the time the explosion sound signal ends is determined by the first window to which the average energy of the signal with a ratio less than a set second threshold belongs.

[0016] Step 5 specifically includes:

[0017] Based on the determined start and end times of the underwater explosion sound signal, the underwater explosion sound signal is extracted from the underwater explosion sound signal file.

[0018] In an improved method for intercepting underwater explosion acoustic signals, step 1 includes: calculating the signal coarse reading start time R according to the following formula. begin and signal coarse reading end time R end :

[0019] R begin =TR begin +TR mod -TR n2

[0020] R end =TR begin +TR mod +TR duration

[0021] Among them, TR begin TR is the reference start time for the obtained explosion sound signal. mod TR is the set error duration. modThe actual start time T of the explosion sound signal begin With TR begin The estimate of the difference, TR n2 TR is the time when the noise ends. duration This is the longest possible duration for the explosion sound signal.

[0022] In an improved method for intercepting underwater explosion acoustic signals, step 3 includes:

[0023] Step 3-1: Perform a high-pass filter on the read data x[n] to remove noise and obtain x′[n];

[0024] Step 3-2: Perform a short-time discrete Fourier transform on x′[n] to obtain the transformed time spectrum X(m,k) as shown in the following equation;

[0025]

[0026] Where w[n] is the first window function, m is the time index of the first window, R is the step size of the first window, j represents the imaginary part, k represents the frequency index, M is the number of points in the discrete Fourier transform, and n is the index of the number of data points.

[0027] Step 3-3: Calculate the signal energy E within the required signal bandwidth of each first window using the following formula. B :

[0028] E B =∑E(f l f h )

[0029] E = q|X| 2 ,

[0030] Among them, f l f is the lowest frequency required for the signal bandwidth. h The highest frequency of the required signal bandwidth is E, the energy spectral density of the read signal is E, q is the transform coefficient, X is the time spectrum X(M,k), fs is the sampling rate, and L is the length of the first window function.

[0031] In an improved method for intercepting underwater explosion acoustic signals, step 4 includes:

[0032] Step 4-1: Calculate the average noise energy over a period of time before the reference start time;

[0033] Step 4-2: Before the maximum value is generated, find the extreme point that satisfies the condition that the ratio of the net energy of the explosion start window signal to the maximum value of the net signal energy of the first window is greater than the first threshold. Take the time window number of the signal energy of the first window that satisfies the extreme point as the time window number containing the start time of the explosion sound signal. Here, the net energy of the explosion start window signal is the difference between the signal energy of the explosion start window and the average noise energy, and the maximum value of the net signal energy of the first window is the difference between the maximum value of the signal energy of the first window and the average noise energy.

[0034] Step 4-3: Determine the start time of the explosion sound signal based on the reserved time before and after the signal and the time window number of the start time of the explosion sound signal.

[0035] In an improved method for intercepting underwater explosion acoustic signals, step 4-1 includes: calculating the reference start time TR. begin A while ago [TR] n1 TR n2 The average noise energy E within the range noise :

[0036]

[0037] Among them, TR n1 <TR n2 L n =(TR) n2 -TR n1 *fs / R represents the total number of time windows used to calculate the average noise energy, where fs is the sampling rate and TR is the sampling rate. n1 TR is the start time for noise truncation. n2 The noise truncation ends at the specified time, and i represents the first window number within the noise segment.

[0038] In an improved method for intercepting underwater explosion acoustic signals, step 4-2 includes:

[0039] Find the signal energy E of the first window B The maximum value E within max And return the time window index x containing the moment of maximum value. max ;

[0040] Based on the time-spectral characteristics of the underwater explosion sound signal, a first threshold W is set. max Find the extreme point that satisfies the following conditions before the maximum value is reached:

[0041] E B (x s )-E noise >W max *(E max -E noise ),x s≤x max

[0042] Where, x s It is the time window number of the extreme point that satisfies the above conditions, x s The final selected time window number is the one that contains the start time of the explosion sound signal.

[0043] In an improved method for intercepting underwater explosion acoustic signals, the step of calculating the reference average energy of the signal based on a first window, according to the reference duration of the explosion acoustic signal and the determined start time of the underwater explosion acoustic signal, includes: calculating the reference average energy E of the signal using the following formula. ref :

[0044]

[0045] Among them, L ref =TR block *fs / R represents the total number of time windows for calculating the reference average energy of the signal, TR block This is the shortest possible duration for the explosion sound signal.

[0046] In an improved method for intercepting underwater explosion acoustic signals, the ratio of the average signal energy of each second window to the average signal energy of a reference signal based on the first window is calculated. The second window containing the average signal energy whose ratio is less than a set second threshold is designated as the second window where the explosion acoustic signal ends. This includes:

[0047] The index x′ of the second window containing the end time of the explosion sound signal is determined using the following formula. end :

[0048] E B (x′ end )-E noise <W end *(E ref -E noise )

[0049] Among them, W end The second threshold is set;

[0050] Calculate the ratio of the average signal energy of each first window within the second window containing the end time of the explosion sound signal to the average signal energy referenced based on the first window. Determine the end time of the explosion sound signal using the first window containing the average signal energy whose ratio is less than a set second threshold, including:

[0051] At x′ end Within the corresponding second window, the average energy of the signal is calculated using the first window. The end window number x, calculated using the first window as the time window scale, is then obtained using the following formula. end ;

[0052] E B (x end )-E noise <W end *(E ref -E noise )

[0053] Where, x s +L s ≤x end ≤x s +L l L s =TR block *fs / R represents the number of time windows for capturing the shortest duration of the signal, L l =TR duration *fs / R represents the number of time windows that capture the longest duration of the intercepted signal;

[0054] Use formula T end =(x end *R) / fs+T pre Calculate the end time T of the underwater explosion acoustic signal end .

[0055] An underwater explosion acoustic signal interception device based on the above-mentioned underwater explosion acoustic signal interception method includes:

[0056] The coarse reading timing determination module is used to determine the start and end times of coarse reading of the signal based on the reference start time of the explosion sound signal obtained from the bombing time record table, the set error duration, and the reference duration of the explosion sound signal.

[0057] The data reading module is used to read data from the underwater explosion sound signal file according to the determined signal coarse reading start time and signal coarse reading end time;

[0058] The window energy calculation module is used to perform time-frequency transformation on the read data based on the first window function and calculate the signal energy of each first window, wherein the first window corresponds to the first window function;

[0059] The explosion sound signal start and end time determination module is used to determine the start time of the explosion sound signal based on the relationship between the maximum signal energy of the first window, the ratio of the signal energy of the explosion start window to the maximum signal energy of the first window, and a set first threshold, wherein the explosion start window is the first window where the explosion start time is located.

[0060] The extraction module is used to extract underwater explosion sound signals from the underwater explosion sound signal file based on a determined start time of the underwater explosion sound signal.

[0061] Compared with the prior art, the advantages of the present invention are:

[0062] 1. Different types of hydrophones receive signals asynchronously; this can be addressed by increasing the error time TR. mod Furthermore, the signal reference duration TR is increased. duration The method is to eliminate errors.

[0063] 2. The reserved time T before and after the signal can be selected by adjusting the first threshold used for starting time lookup and the second threshold used for ending time lookup. pre This minimizes interference from noise and other factors, reduces errors in manual identification and extraction methods, and enhances data quality.

[0064] 3. It can achieve batch, standardized, automatic, and accurate interception of underwater explosion sound signals, greatly saving manpower and time costs. Attached Figure Description

[0065] Figure 1 This is a flowchart of the underwater explosion sound signal interception method provided in Embodiment 1 of the present invention;

[0066] Figure 2 This is a flowchart of the underwater explosion sound signal interception method provided in Embodiment 2 of the present invention;

[0067] Figure 3 This is a flowchart of the underwater explosion sound signal interception method provided in Embodiment 3 of the present invention;

[0068] Figure 4 This is a time-frequency distribution diagram of a coarse reading of an explosion sound signal provided in Embodiment 4 of the present invention;

[0069] Figure 5 This is the time-frequency distribution diagram of the precisely captured signal provided in Embodiment 4 of the present invention;

[0070] Figure 6(a) is a diagram of the first explosion sound signal captured under a signal-to-noise ratio of approximately 0 dB, provided in Embodiments 3 and 4 of the present invention.

[0071] Figure 6(b) is a second explosion sound signal diagram captured under a signal-to-noise ratio of approximately 0 dB, provided in Embodiment 4 of the present invention.

[0072] Figure 6(c) is a diagram of the third explosion sound signal captured under a signal-to-noise ratio of approximately 0 dB, provided in Embodiment 4 of the present invention. Detailed Implementation

[0073] This invention proposes a method for automatically extracting underwater explosion sound signals based on time-frequency energy. This method utilizes time-frequency transformation to calculate the time-frequency spectrum of the original signal, and designs a threshold based on the analysis of the time-frequency energy spectrum to automatically find the start and end times of the explosion sound signal by comparing energy levels. Compared to underwater explosion sound signal extraction methods that rely on manual identification, this invention has advantages such as saving manpower and time costs, standardizing the extracted signal, and reducing noise interference.

[0074] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0075] Example 1

[0076] like Figure 1 As shown, it is a flowchart of the underwater explosion sound signal interception method provided in Embodiment 1 of the present invention, which includes the following steps:

[0077] Step 101: Based on the reference start time of the explosion sound signal obtained from the bombing time record table, the set error duration, and the reference duration of the explosion sound signal, determine the start time and end time of the coarse reading of the signal.

[0078] Here, the reference start time of each explosion sound signal can be obtained based on the bomb drop record table recorded during the experiment.

[0079] In step 101, the error duration and the reference duration of the explosion sound signal are set for signal reading, so that for asynchronous reception of different types of hydrophones, the error can be eliminated by increasing the error duration and the reference duration of the signal.

[0080] Step 102: Read data from the underwater explosion sound signal file according to the determined signal coarse reading start time and signal coarse reading end time.

[0081] Among them, the file that stores the single explosion sound signal recorded by a single underwater acoustic device is called the underwater explosion sound signal file.

[0082] Step 103: Perform time-frequency transformation on the read data based on the first window function, and calculate the signal energy of each first window.

[0083] The time-frequency transform here can be a discrete short-time Fourier transform (SFT). The energy spectral density of the signal can be obtained by assigning values ​​to the time spectrum of the SFT. Integrating the energy density spectrum over the required signal bandwidth yields the signal energy for the first window. The first window function is a time window, which can be a rectangular time window, corresponding to the first window.

[0084] Step 104: Based on the relationship between the maximum signal energy of the first window, the ratio of the signal energy of the explosion start window to the maximum signal energy of the first window, and the set first threshold, determine the start time of the explosion sound signal, wherein the explosion start window is the first window where the explosion start time is located.

[0085] Specifically, step 104 may include the following steps 104-1 to 104-3:

[0086] Step 104-1: Calculate the average noise energy over a period of time before the reference start time.

[0087] To calculate the exact moment of the explosion's onset more accurately, we calculate the average noise energy unaffected by the explosion's acoustic signal, preparing for the net energy calculation in the following steps.

[0088] Step 104-2: Before the maximum value is generated, find the extreme point that satisfies the condition that the ratio of the net energy of the explosion start window signal to the maximum value of the net signal energy of the first window is greater than the first threshold. Use the time window number of the signal energy of the first window that satisfies the extreme point as the time window number that includes the start time of the explosion sound signal.

[0089] Among them, the net energy of the explosion start window signal is the difference between the signal energy of the explosion start window and the average noise energy, and the maximum net signal energy of the first window is the difference between the maximum signal energy of the first window and the average noise energy.

[0090] Step 104-3: Determine the start time of the explosion sound signal based on the reserved time before and after the signal and the time window number of the start time of the explosion sound signal.

[0091] Step 105: Based on the determined start time of the underwater explosion sound signal, extract the underwater explosion sound signal from the underwater explosion sound signal file.

[0092] Here, by accurately determining the start time of the underwater explosion sound signal, the underwater explosion sound signal can be intercepted based on the reference duration of the explosion sound signal, realizing automatic signal extraction. Compared with manual extraction, this improves extraction efficiency and quality.

[0093] This embodiment 1 utilizes the time-frequency characteristics of the data, sets a first threshold, and accurately determines the start time of the underwater explosion sound signal by comparing the signal energy of the explosion start time window with the maximum signal energy of the first window, thus achieving relatively accurate interception of the underwater explosion sound signal. To more accurately intercept the underwater explosion sound signal, embodiment 2 of this invention provides an accurate determination of the end time of the underwater explosion sound signal, which will be described below using the scheme of embodiment 2.

[0094] Example 2

[0095] like Figure 2 As shown, it is a flowchart of the underwater explosion sound signal interception method provided in Embodiment 2 of the present invention, which includes the following steps:

[0096] Step 201: Based on the reference start time of the explosion sound signal obtained from the bombing time record table, the set error duration, and the reference duration of the explosion sound signal, determine the start time and end time of the coarse reading of the signal.

[0097] Step 202: Read data from the underwater explosion sound signal file according to the determined signal coarse reading start time and signal coarse reading end time.

[0098] Step 203: Perform time-frequency transformation on the read data based on the first window function, and calculate the signal energy of each first window.

[0099] Step 204: Based on the relationship between the maximum signal energy of the first window, the ratio of the signal energy of the explosion start window to the maximum signal energy of the first window, and the set first threshold, determine the start time of the explosion sound signal, wherein the explosion start window is the first window where the explosion start time is located.

[0100] Step 205: Calculate the reference average energy of the signal based on the first window, according to the reference duration of the explosion sound signal and the determined start time of the underwater explosion sound signal.

[0101] Step 206: Perform time-frequency transformation on the read data based on the second window function, and calculate the average signal energy of each second window, wherein the moving step size of the second window is greater than the moving step size of the first window.

[0102] Here, to improve the accuracy of finding the end time of the explosion sound signal, a second window function is added. The movement step size of the second window is larger than that of the first window. The second window function is also a time window, which can be a rectangular time window, corresponding to the second window. The first and second window functions can be window functions that are identical except for the movement step size.

[0103] Step 207: Calculate the ratio of the average signal energy of each second window to the average signal energy referenced based on the first window, and take the second window to which the average signal energy with a ratio less than the set second threshold belongs as the second window where the explosion sound signal ends.

[0104] Step 208: Calculate the average energy of the signal in each of the first windows within the second window at the time the explosion sound signal ends.

[0105] Step 209: Calculate the ratio of the average signal energy of each first window within the second window at the time the explosion sound signal ends to the average signal energy referenced based on the first window. Determine the time the underwater explosion sound signal ends by using the first window to which the average signal energy with a ratio less than a set second threshold belongs.

[0106] Step 210: Based on the determined start and end times of the underwater explosion sound signal, extract the underwater explosion sound signal from the underwater explosion sound signal file. In the scheme of Embodiment 2 of the present invention, by utilizing the time-spectrum characteristics of the data and setting a second threshold, the end time of the underwater explosion sound signal is accurately determined by comparing the average energy of the window signal at the end time of the explosion sound signal with the reference average energy of the entire explosion signal. Combined with the already accurately determined start time of the underwater explosion sound signal, more accurate extraction of the underwater explosion sound signal can be achieved.

[0107] The present invention will be described in detail below through Example 3.

[0108] Example 3

[0109] The underwater explosion sound signal interception method of Embodiment 3 of the present invention includes the following steps:

[0110] Step 1) Based on the bombing record table recorded during the experiment, obtain the reference start time TR for each explosion sound signal. begin Based on the time-frequency characteristics of underwater explosion sound signals, the error duration TR is set. mod Calculate the noise time period [TR] n1 TR n2 ](0 <TR n1 <TR n2 Set the maximum duration TR of the explosion sound signal as a reference. duration and shortest duration TR block .

[0111] TR mod ~(T) begin -Tr begin

[0112] Among them, T begin It is the actual start time of the explosion sound signal.

[0113] Step 2) For each single explosion sound signal recorded by a single underwater acoustic device, store the corresponding file for that signal, and combine it with the parameters set in Step 1) from R begin Time to R end The original data x[n] is coarsely read out at any time, where n is the index of the data point, n = 1, 2, ..., N, and N is the total number of data points.

[0114] R begin=TR begin +TR mod -TR n2

[0115] R end =TR begin +TR mod +TR duration

[0116] Among them, R begin R is the start time of coarse signal reading. end This is the end time of the coarse signal reading.

[0117] The read x[n] is first subjected to a high-pass filter to remove the background noise of the underwater acoustic receiver, resulting in x′[n]. Then, a discrete short-time Fourier transform is applied to obtain the transformed time-frequency spectrum X(m,k):

[0118]

[0119] Where X(m,k) represents the discrete short-time Fourier transform result at time index m and frequency index k, x′[n] is the high-pass filtered signal data, w[n-mR] is the first window function, m is the time index of the first window, R is the moving step size of the first window, j represents the imaginary part, k represents the frequency index, M is the number of points of the discrete Fourier transform, and n is the index of the sampling point.

[0120] Step 3) Based on the time spectrum X(m,k), add a transformation coefficient q to compensate for the influence of the window function, and obtain the energy spectral density E:

[0121] E = q|X| 2 ,

[0122] Where E represents the energy spectral density of the read signal, q is the transform coefficient, X is the time spectrum X(m,k) obtained in step 2, fs is the sampling rate, w(n) is the first window function in step 2), and L is the length of the first window function.

[0123] Based on the energy spectral density E, according to the required signal bandwidth B(f) l =0,f h =5000Hz), for the signal energy E within the required bandwidth B ;

[0124] E B =∑E(f l =0,f h =5000Hz)

[0125] Among them, f l f is the lowest frequency required for the signal bandwidth.h The highest frequency for the required signal bandwidth.

[0126] Step 4) Calculate the reference start time TR begin A while ago [TR] n1 TR n2 ](TR n1 <TR n2 The average noise energy E within the range noise This is to help determine the start and end points of the signal in subsequent steps;

[0127]

[0128] Among them, L n =(TR) n2 -TR n1 *fs / R represents the total number of time windows used to calculate the average noise energy, where fs is the sampling rate, R is the time window shift step mentioned in step 2), and E B Let i represent the signal energy within the required signal bandwidth obtained in step 3), and let i represent the first window number within the noise segment.

[0129] Step 5) Find the signal energy E B The maximum value E within max And return the time window index x containing the moment of maximum value. max Based on the time-spectrum characteristics of the underwater explosion sound signal, a first threshold W is set. max Find the extreme values ​​that satisfy the following conditions before the maximum value is generated:

[0130] E B (x s )-E noise >W max *(E max -E noise ),x s ≤x max

[0131] Among them, E B x represents the signal energy within the desired signal bandwidth obtained in step 3). s It is the time window number of the extreme point that satisfies the above conditions, E noise This is the average noise energy obtained in step 4). Finally, x is chosen. s As the sequence number of the time window containing the start time of the explosion sound signal.

[0132] Step 6) Refer to the shortest duration TR of the explosion sound signal. block And the time window number x, which includes the start time of the explosion sound signal, obtained from step 5). s Calculate the reference average energy E of the signal. ref.

[0133]

[0134] Among them, L ref =TR block *fs / R represents the total number of time windows used to calculate the reference average energy of the signal, E B The signal energy within the required signal bandwidth obtained in step 3).

[0135] Step 7) Increase the time window's moving step size by 10 times the moving step size R of the window function w[n-mR] in Step 2), resulting in a new window w′[nm′R′], also known as the second window (where R′=10R). Calculate the average energy of the signal within each of the second windows searched, and set a second threshold W. end Combined with the signal reference average energy E obtained in step 6), ref Compare with average noise energy.

[0136] E B (x′ end )-E noise <W end *(E ref -E noise )

[0137] Where, x′ end The final selected second window number contains the time of the end of the explosion sound signal.

[0138] At x′ end Within the second window, the above process is repeated using the first window to search again, obtaining the result relative to the starting point x. s End window number x under the same time window scale end (x s +L s ≤x end ≤x s +L l ).

[0139] E B (x end )-E noise <W end (E ref -E noise )

[0140] Where, x s +L s ≤x end ≤x s +L l L s =TR block *fs / R represents the number of time windows for capturing the shortest duration of the signal, Ll =TR duration *fs / R represents the number of time windows capturing the longest duration of the intercepted signal. Where x s L is the time window sequence number containing the start time of the explosion sound signal, obtained in step 5). s =TR block *fs / R represents the number of time windows for capturing the shortest duration of the signal, L l =TR duration TR represents the number of time windows that capture the longest duration of the signal. block It is the shortest duration set in step 1), TR duration It is the longest duration set in step 1), fs is the sampling rate, and R is the step size of the first window mentioned in step 2).

[0141] Here, the start and end times of the explosion are both marked with the first window function index. Adding a second window function when searching for the end time of the explosion sound signal is to avoid errors caused by the excessively small scale of the first window function. As shown in Figure 6(a), if the second window function is not used and the shortest signal duration is not set, and only energy is used for discrimination, the explosion end point determined by the program will be within the time gap between the arrival of the first path and the arrival of the multipath after the start time. This gap refers to the 2nd to 6th second in Figure 6(a).

[0142] Step 8) Based on the time-frequency characteristics of the underwater explosion sound signal, set the pre-reserved time T before and after the signal. pre According to the time window number x that contains the start time of the explosion sound signal obtained in step 5), s And the time window number x that contains the end time of the explosion sound obtained in step 7). end This allows for the extraction of a single underwater explosion sound signal from a single file.

[0143] T begin =((x) s -1)*R+1) / fs-T pre

[0144] T end =(x end *R) / fs+T pre

[0145] Where fs is the sampling rate and R is the time window moving step size mentioned in step 2).

[0146] For batch-imported data, the processing procedure can be as follows: Figure 3 As shown, based on the set parameters, such as the error duration TR... mod Calculate the noise time period [TR] n1 TR n2](0 <TR n1 <TR n2 Set the maximum duration TR of the explosion sound signal as a reference. duration Shortest duration TR block Using the above-mentioned method for capturing underwater explosion sound signals, the starting and ending time points of the underwater explosion sound signals are determined by the following parameters: moving step size R, transformation coefficient q, window function step size L, and signal bandwidth B. This process is repeated until all signals have been processed.

[0147] Example 4

[0148] This embodiment 4 illustrates the solution of the present invention through a specific example.

[0149] During an underwater acoustic test conducted in an actual sea area, a certain underwater acoustic device received a series of underwater explosion sound signals, saving a total of 32 recording files, each containing data. m (n), n=1,2,…,230400000, m=1,2,…,32. Sampling frequency f of the underwater acoustic receiving device. s =64kHz, meaning each recording file is 1 hour long;

[0150] Randomly select any underwater explosion sound signal. According to the bomb drop record table recorded during the experiment, the recording time is 15:02:12 Beijing time on that day. Then refer to the start time TR. begin =2*60+12=132s; Based on the manual recording error and the asynchrony between different underwater acoustic receiving devices, the error duration T is set. mod =20s; the noise calculation time period is set to [10,20], i.e., T n1 =10s,T n2 =20s; Reference maximum duration TR duration =50s, referencing the shortest duration T block =10s.

[0151] Read the file storing the explosion sound signal, obtain x[n], and then use a length of 2 14 And the superposition amount is 2 13 Using the Hanning window function, perform a short-time Fourier transform with 2^6 points. 10 Calculate the time-frequency distribution of the explosion sound signal, and the result is as follows: Figure 4 As shown. Figure 4 The data indicates that this type of hydrophone has background noise below 100Hz. In subsequent processing, a high-pass filter can be applied first to reduce errors.

[0152] After high-pass filtering x[n] to a frequency above 100Hz, x′[n] is obtained. The aforementioned window is then applied to x′[n], followed by a short-time Fourier transform to obtain the time-frequency spectrum X(m,k). Transform coefficients are then added. The energy spectral density E of the signal is obtained. Then, the signal energy E is obtained according to the required signal bandwidth B. B .

[0153] At the reference start time TR begin =2*60+12=132s Calculate the average noise energy E within the first 10-20 seconds noise Set the threshold W based on experience. max =0.5, W end =0.1, according to steps 5)-7) in Example 3, find the start and end times of the explosion sound signal respectively. Set the pre-reserved time T before and after the signal. pre =2s, the final intercepted signal time-frequency distribution diagram is as follows Figure 5 As shown in the example, the captured explosion sound signal has no obvious interference and a signal-to-noise ratio of approximately 2dB, achieving ideal results with automatic capture. If the automatically captured signal is unsatisfactory, the threshold and reference duration can be fine-tuned for automatic capture again.

[0154] Based on the determined high-pass filter frequency lower limit, signal bandwidth, threshold, and signal reference duration, batch automatic interception of explosion sound signals can be achieved. Figures 6(a), 6(b), and 6(c) show the first, second, and third explosion sound signals automatically intercepted when the signal-to-noise ratio is approximately 0 dB, respectively. The results show that this method can effectively achieve automatic and accurate interception of underwater explosion sound signals under high signal-to-noise ratio conditions, greatly saving the high manpower and time costs required by conventional manual identification methods. It also overcomes the problem of time asynchrony when different underwater acoustic devices are used in conjunction. This provides convenience for subsequent underwater acoustic detection and research based on underwater explosion sound signals.

[0155] Example 5

[0156] Based on the methods in the above embodiments, Embodiment 5 of the present invention provides an underwater explosion sound signal interception device, comprising: a coarse reading time determination module, a data reading module, a window energy calculation module, an explosion sound signal start and end time determination module, and an interception module; wherein:

[0157] The coarse reading time determination module is used to determine the start time and end time of coarse reading of the signal based on the reference start time of the explosion sound signal obtained from the bombing time record table, the set error duration, and the reference duration of the explosion sound signal.

[0158] The data reading module is used to read data from the underwater explosion sound signal file according to the determined signal coarse reading start time and signal coarse reading end time;

[0159] The window energy calculation module is used to perform time-frequency transformation on the read data based on the first window function and the second window function, and calculate the signal energy of each first window and the second window.

[0160] The explosion sound signal start and end time determination module is used to determine the start time of the explosion sound signal based on the relationship between the maximum signal energy of the first window, the ratio of the signal energy of the explosion start window to the maximum signal energy of the first window, and a set first threshold, wherein the explosion start window is the first window where the explosion start time is located.

[0161] The interception module is used to intercept the underwater explosion sound signal from the underwater explosion sound signal file based on the determined start time of the underwater explosion sound signal.

[0162] In an improved embodiment, the explosion sound signal start and end time determination module is further configured to: calculate the signal reference average energy based on a first window according to the reference duration of the explosion sound signal and the determined start time of the underwater explosion sound signal; perform time-frequency transformation on the read data based on a second window function to calculate the signal average energy of each second window, wherein the movement step size of the second window is greater than that of the first window; calculate the ratio of the signal average energy of each second window to the signal reference average energy based on the first window, and use the second window to which the signal average energy with a ratio less than a set second threshold belongs as the second window where the explosion sound signal ends; calculate the signal average energy of each first window within the second window where the explosion sound signal ends; calculate the ratio of the signal average energy of each first window within the second window where the explosion sound signal ends to the signal reference average energy based on the first window, and determine the explosion sound signal end time using the first window to which the signal average energy with a ratio less than a set second threshold belongs.

[0163] The interception module is specifically used to intercept the underwater explosion sound signal from the underwater explosion sound signal file based on the determined start time of the underwater explosion sound signal and the start time of the explosion sound signal.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of the present invention do not depart from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intercepting underwater explosion acoustic signals, comprising: Step 1: Based on the reference start time of the explosion sound signal obtained from the bombing time record table, the set error duration, and the reference duration of the explosion sound signal, determine the start time and end time of the coarse reading of the signal; Step 2: Read data from the underwater explosion sound signal file according to the determined start and end times of the signal coarse reading; Step 3: Perform time-frequency transformation on the read data based on the first window function, and calculate the signal energy of each first window, where the first window corresponds to the first window function; Step 4: Based on the relationship between the maximum signal energy of the first window, the ratio of the signal energy of the explosion start window to the maximum signal energy of the first window, and the set first threshold, determine the start time of the explosion sound signal, wherein the explosion start window is the first window where the explosion start time is located; Step 5: Based on the determined start time of the underwater explosion sound signal, extract the underwater explosion sound signal from the underwater explosion sound signal file.

2. The underwater explosion acoustic signal interception method as described in claim 1, characterized in that, After step 4 and before step 5, the following is also included: Based on the reference duration of the explosion sound signal and the determined start time of the underwater explosion sound signal, calculate the signal reference average energy based on the first window; The data is transformed by time and frequency based on the second window function, and the average energy of the signal in each second window is calculated. The second window corresponds to the second window function, and the moving step size of the second window is larger than that of the first window. Calculate the ratio of the average signal energy of each second window to the average signal energy referenced based on the first window, and take the second window to which the average signal energy with a ratio less than a set second threshold belongs as the second window where the explosion sound signal ends. Calculate the average energy of the signal in each of the first windows within the second window at the moment the explosion sound signal ends; The ratio of the average energy of the signal in each first window within the second window at the time the explosion sound signal ends to the average energy of the signal reference based on the first window is calculated, and the time the explosion sound signal ends is determined by the first window to which the average energy of the signal with a ratio less than a set second threshold belongs. Step 5 specifically includes: Based on the determined start and end times of the underwater explosion sound signal, the underwater explosion sound signal is extracted from the underwater explosion sound signal file.

3. The underwater explosion acoustic signal interception method as described in claim 2, characterized in that, Step 1 includes: calculating the signal coarse reading start time R according to the following formula. begin and signal coarse reading end time R end : R begin =TR begin +TR mod -TR n2 R end =TR begin +TR mod +TR duration Among them, T begin TR is the reference start time for the obtained explosion sound signal. mod TR is the set error duration. mod The actual start time T of the explosion sound signal begin With TR begin The estimate of the difference, TR n2 TR is the time when the noise ends. duration This is the longest possible duration for the explosion sound signal.

4. The underwater explosion acoustic signal interception method as described in claim 3, characterized in that, Step 3 includes: Step 3-1: Perform a high-pass filter on the read data x[n] to remove noise and obtain x. ′ [n]; Step 3-2: For x ′ Performing a short-time discrete Fourier transform on [n] yields the transformed time-frequency spectrum X(m,k) as shown in the following equation; Where w[n] is the first window function, m is the time index of the first window, R is the step size of the first window, j represents the imaginary part, k represents the frequency index, M is the number of points in the discrete Fourier transform, and n is the index of the number of data points. Step 3-3: Calculate the signal energy E within the required signal bandwidth of each first window using the following formula. B : E B =∑E(f l ,f h ) E=q|X| 2 , Among them, f l f is the lowest frequency required for the signal bandwidth. h The highest frequency of the required signal bandwidth is E, the energy spectral density of the read signal is E, q is the transform coefficient, X is the time spectrum X(m,k), fs is the sampling rate, and L is the length of the first window function.

5. The underwater explosion acoustic signal interception method as described in claim 4, characterized in that, Step 4 includes: Step 4-1: Calculate the average noise energy over a period of time before the reference start time; Step 4-2: Before the maximum value is generated, find the extreme point that satisfies the condition that the ratio of the net energy of the explosion start window signal to the maximum value of the net signal energy of the first window is greater than the first threshold. Take the time window number of the signal energy of the first window that satisfies the extreme point as the time window number containing the start time of the explosion sound signal. Here, the net energy of the explosion start window signal is the difference between the signal energy of the explosion start window and the average noise energy, and the maximum value of the net signal energy of the first window is the difference between the maximum value of the signal energy of the first window and the average noise energy. Step 4-3: Determine the start time of the explosion sound signal based on the reserved time before and after the signal and the time window number of the start time of the explosion sound signal.

6. The underwater explosion acoustic signal interception method as described in claim 5, characterized in that, Step 4-1 includes: calculating the reference start time TR. begin A while ago [TR] n1 TR n2 The average noise energy E within the range noise : Among them, TR n1 <TR n2 L n =(TR) n2 -TR n1 *fs / R represents the total number of time windows used to calculate the average noise energy, where fs is the sampling rate and TR is the sampling rate. n1 TR is the start time for noise truncation. n2 The noise truncation ends at the specified time, and i represents the first window number within the noise segment.

7. The underwater explosion acoustic signal interception method as described in claim 6, characterized in that, Step 4-2 includes: Find the signal energy E of the first window B The maximum value E within max And return the time window index x containing the moment of maximum value. max ; Based on the time-spectral characteristics of the underwater explosion sound signal, a first threshold W is set. max Find the extreme point that satisfies the following conditions before the maximum value is reached: E B (x s )-E noise >W max *(E max -E noise ),x s ≤x max Where, x s It is the time window number of the extreme point that satisfies the above conditions, and it is the time window number finally selected as the time window number containing the start time of the explosion sound signal.

8. The underwater explosion acoustic signal interception method as described in claim 7, characterized in that, The step of calculating the reference average energy of the signal based on the reference duration of the explosion sound signal and the determined start time of the underwater explosion sound signal includes: calculating the reference average energy E of the signal using the following formula. ref : Among them, L ref =TR block *fs / R represents the total number of time windows for calculating the reference average energy of the signal, TR block This is the shortest possible duration for the explosion sound signal.

9. The underwater explosion acoustic signal interception method as described in claim 8, characterized in that, Calculate the ratio of the average signal energy in each second window to the average signal energy referenced in the first window. The second window containing the average signal energy whose ratio is less than a set second threshold is designated as the second window where the explosion sound signal ends. The index x′ of the second window containing the end time of the explosion sound signal is determined using the following formula. end : AND B (x′ end )-AND noise <W end *(AND ref -AND noise ) Among them, W end The second threshold is set; Calculate the ratio of the average signal energy of each first window within the second window containing the end time of the explosion sound signal to the average signal energy referenced based on the first window. Determine the end time of the explosion sound signal using the first window containing the average signal energy whose ratio is less than a set second threshold, including: At x′ end Within the corresponding second window, the average energy of the signal is calculated using the first window. The end window number x, calculated using the first window as the time window scale, is then obtained using the following formula. end ; AND B (x end )-AND noise <W end *(AND ref -AND noise ) Where, x s +L s ≤x end ≤x s +L l L s =TR block *fs / R represents the number of time windows for capturing the shortest duration of the signal, L l =TR duration *fs / R represents the number of time windows that capture the longest duration of the intercepted signal; Use formula T end =(x end *R) / fs+T pre Calculate the end time T of the underwater explosion acoustic signal end .

10. An underwater explosion acoustic signal interception device based on the underwater explosion acoustic signal interception method according to claim 1, characterized in that, include: The coarse reading timing determination module is used to determine the start and end times of coarse reading of the signal based on the reference start time of the explosion sound signal obtained from the bombing time record table, the set error duration, and the reference duration of the explosion sound signal. The data reading module is used to read data from the underwater explosion sound signal file according to the determined signal coarse reading start time and signal coarse reading end time; The window energy calculation module is used to perform time-frequency transformation on the read data based on the first window function and calculate the signal energy of each first window, wherein the first window corresponds to the first window function; The explosion sound signal start and end time determination module is used to determine the start time of the explosion sound signal based on the relationship between the maximum signal energy of the first window, the ratio of the signal energy of the explosion start window to the maximum signal energy of the first window, and a set first threshold, wherein the explosion start window is the first window where the explosion start time is located. The extraction module is used to extract underwater explosion sound signals from the underwater explosion sound signal file based on a determined start time of the underwater explosion sound signal.

Images