Automatic intercepting method for underwater explosion sound signal based on long-short time average energy ratio

Abstract

The application discloses a kind of underwater explosion sound signal automatic interception methods based on long-short time average energy ratio, comprising: selecting a underwater explosion sound signal file from the experimental measuring line record result of underwater acoustic equipment, time-frequency analysis is carried out to obtain time-frequency spectrum;According to the time-frequency spectrum characteristics of explosion sound, set the upper and lower limits of filter frequency, determine signal interception length and signal pre-reserved length;According to the long-short time average energy ratio of the filter upper and lower limits calculated after data, long-short time parameter and explosion sound signal trigger threshold are determined in combination;According to the long-short time parameter and filter frequency upper and lower limits specified, move long-short time time window, carry out batch processing to the record data of whole measuring line, calculate the long-short time average energy ratio of each data point, and compare with explosion sound signal trigger threshold, obtain trigger determination result;According to trigger determination result and the signal interception length and signal pre-reserved length determined, carry out unified interception to underwater explosion sound signal.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of underwater acoustic physics, and particularly relates to an underwater explosion sound signal automatic interception method based on long-short time average energy ratio. BACKGROUND

[0002] Due to the advantages of high source level, wide frequency band, short pulse time and convenient use, underwater explosion type sound sources are widely used in marine acoustic investigation and underwater acoustic detection experiments such as underwater acoustic propagation loss measurement, ocean reverberation measurement and seabed acoustic low-frequency characteristic inversion measurement. In the current marine acoustic investigation and underwater acoustic detection experiments, in order to achieve more comprehensive and in-depth investigation and detection purposes, the number of underwater acoustic receiving devices, the distance of the measuring line and the number of sound bombs are increasing. However, the joint use of a large number of sound bombs and multiple underwater acoustic devices produces extremely huge underwater explosion sound signal recording data, greatly increasing the difficulty of underwater acoustic recording effective signal interception and subsequent analysis. Relying only on conventional manual identification and signal interception results in great waste of human resources and huge time cost. At the same time, the joint use of multiple underwater acoustic devices often cannot achieve time synchronization acquisition, further increasing the difficulty of signal interception, and often introducing additional errors for subsequent analysis due to different starting times of signal interception. Therefore, based on the characteristics of the ocean channel and the ocean environment background noise, according to the time-frequency characteristics and high source level characteristics of the underwater explosion sound signal, it is of great practical significance to establish an underwater explosion sound signal automatic interception method. SUMMARY

[0003] The purpose of the present application is to overcome the defects of the prior art and provide an underwater explosion sound signal automatic interception method based on long-short time average energy ratio.

[0004] In order to achieve the above purpose, the present application provides an underwater explosion sound signal automatic interception method based on long-short time average energy ratio, which comprises the following steps:

[0005] Step 1) selecting an underwater explosion sound signal file from the experimental measuring line recording results of the underwater acoustic device, and performing time-frequency analysis on the recorded data to obtain a time-frequency spectrum;

[0006] Step 2) setting the upper and lower limits of the filtering frequency according to the time-frequency spectrum characteristics of the explosion sound signal, and determining the signal interception time length and the signal pre-reserved time length;

[0007] Step 3) calculating the long-short time average energy ratio of the filtered data according to the set upper and lower limits of the filtering;

[0008] Step 4) determining the long-short time parameters and the explosion sound signal triggering threshold in combination with the long-short time average energy ratio;

[0009] Step 5) According to the specified long-short time parameters and the upper and lower limits of the filtering frequency, moving the long-short time window, batch processing the recorded data of the entire measuring line, calculating the long-short time average energy ratio of each data point of each file in the measuring line, and comparing the long-short time average energy ratio with the explosion sound signal trigger threshold to obtain the trigger determination result of each file.

[0010] Step 6) According to the trigger determination result of each file and the determined signal interception time length and signal pre-reserved time length, the underwater explosion sound signal of each file is intercepted.

[0011] As an improvement of the above method, the step 1) specifically comprises:

[0012] Select an underwater explosion sound signal file from the experimental measuring line record result of the underwater acoustic equipment, and the recorded data is x(n), n is the data point number, n = 1, 2, … N, N is the total number of data points;

[0013] The short-time Fourier transform is used to obtain the transformed time-frequency spectrum P STFT (t, f) is:

[0014]

[0015] Wherein, s(τ) represents the recorded data in the time domain, w(τ-t) is a time window function with a translation t length; t represents the time variable in the time-frequency spectrum, f represents the frequency variable in the time-frequency spectrum, τ represents the time point corresponding to the recorded data, and i represents the imaginary part.

[0016] As an improvement of the above method, the step 2) specifically comprises:

[0017] According to the time-frequency spectrum characteristics of the underwater explosion sound signal, the lower limit of the filtering frequency is determined as f L , the upper limit is f H , the interception time length is T S , and the signal pre-reserved time length is T ca .

[0018] As an improvement of the above method, the step 3) specifically comprises:

[0019] According to the selected long-short time characteristic function F(n), the preliminarily set short time window length T S and the long time window length T L , the short time average energy S TA (n) and the long time average energy L TA (n) of each data point n after filtering are calculated:

[0020]

[0021] Wherein, LS L is the short-time window point number s T s L s L L L is the long-time window point number L T L L s L s f is the sampling frequency, and j represents each data point within the time window range

[0022] The long-short-time average energy ratio N SL (n) is obtained according to the following formula:

[0023]

[0024] As an improvement of the above method, the long-short-time feature function F(n) satisfies the following formula:

[0025] F(n) = K1|x(n)| + K2|x(n) - x(n-1)|

[0026] Wherein, |x(n)| represents the amplitude feature of the signal to be processed, |x(n) - x(n-1)| represents the amplitude change trend and frequency domain feature of the signal, and K1 and K2 are weight values set according to the signal sampling frequency and the inherent noise properties of the device.

[0027] As an improvement of the above method, the short-time window length T S and the long-time window length T L are selected based on the time domain characteristics of underwater explosion sound signals and the characteristics of ocean environmental noise, and are preliminarily determined.

[0028] As an improvement of the above method, the step 4) specifically comprises:

[0029] According to the long-short-time average energy ratio N SL (n), the effectiveness of the feature function F(n) and the size of the explosion sound signal triggering threshold T h are determined, and the short-time window length T S and the long-time window length T L are further adjusted.

[0030] As an improvement of the above method, the step 5) specifically comprises:

[0031] According to the specified long-short-time parameters and the upper and lower limits of the filtering frequency, the long-short-time time window is moved, and each record file of the experimental line is traversed, wherein the data of the mth record file is x m(n), n is the number of data points, n = 1, 2,..., N, N is the total number of data points of each record file, m is the file serial number saved according to the device setting parameters in the experimental measuring line, m = 1, 2,..., M, M is the total number of files recorded on the experimental measuring line;

[0032] Calculate x m (n) is the long-short time average energy ratio N SL-m (n) of the data point n of the file, and compare the long-short time average energy ratio N SL-m (n) with the set explosion sound signal trigger threshold T h , and obtain the trigger judgment result J cen-m (n) of the mth record file.

[0033]

[0034] As an improvement of the above method, the step 6) specifically comprises:

[0035] According to the signal length estimation value T ch , the trigger judgment J cen-m (n) = 1 is further processed, and for the data length T ch * f s , the first data point n1 and other trigger data points n satisfying the trigger condition are obtained, and the following formula is satisfied:

[0036]

[0037] Wherein, n1 is the automatic interception trigger data point of the underwater explosion sound signal, and the automatic interception trigger time t n = n1 / f s .

[0038] According to the interception duration T S and the signal pre-reserved duration T ca , the underwater explosion sound signal in the data x m (n) of the record file is intercepted and saved.

[0039] Compared with the prior art, the advantages of the present application are:

[0040] The application establishes an automatic interception method of underwater explosion sound signals based on long-short time average energy ratio according to the time-frequency spectrum distribution of underwater explosion sound signals and the characteristics of ocean channel, selects appropriate long-short time window parameters, characteristic functions and threshold values, fully considers the difference between the time-frequency characteristics of underwater explosion source and ocean ambient noise, and uses the rapidness and robustness of long-short time average energy calculation method, greatly improves the interception time and interception accuracy of underwater explosion sound signals, avoids the additional errors caused by manual interception and different equipment time non-synchronization, and provides strong support for rapid analysis of marine environmental investigation and underwater acoustic detection data. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a flow chart of the automatic interception method of underwater explosion sound signals based on long-short time average energy ratio of the application;

[0042] Fig. 2(a) is an explosion sound signal recorded by a water acoustic device in a simulation example;

[0043] Fig. 2(b) is the time-frequency spectrum distribution corresponding to Fig. 2(a);

[0044] Fig. 3(a) is the long-short time average energy of underwater explosion sound signals in a simulation example;

[0045] Fig. 3(b) is the trigger determination corresponding to Fig. 3(a);

[0046] Figure 4 is the intercepted waveform of underwater explosion sound signals in a simulation example. DETAILED DESCRIPTION

[0047] The signal interception method is an automatic interception method of underwater explosion sound signals established on the basis of the time-frequency characteristics of underwater explosion sound signals and long-short time average energy ratio, which mainly includes the following steps: setting the upper and lower limits of filter frequency, signal length estimation value, long-short time window length and long-short time characteristic function according to the time-frequency spectrum characteristics of explosion sound; performing filter processing on the data to be processed, preliminarily calculating the long-short time average energy ratio according to the selected long-short time parameters, determining the effectiveness of the characteristic function and the trigger threshold of the explosion sound signal, and appropriately adjusting the long-short time window length according to the actual situation; moving the long-short time window according to the specified long-short time parameters and filter frequency upper and lower limits, and repeatedly processing the recorded data in batches to intercept and save each explosion sound signal.

[0048] The technical solutions of the application will be described in detail below with reference to the drawings and examples.

[0049] Example 1

[0050] As shown in Figure 1 , the application proposes an automatic interception method of underwater explosion sound signals based on long-short time average energy ratio, which includes the following steps:

[0051] Step one: record the data x(n), n=1, 2, …N as data points, and the sampling frequency is f s When analyzing the time-frequency characteristics of underwater explosion signals by using time-frequency transform, considering the size of data calculation and the advantage of calculation efficiency, the most basic short-time Fourier transform (STFT) method is adopted:

[0052]

[0053] Wherein, P STFT (t, f) is the transformed time-frequency spectrum, w(τ) is the selected time window function; according to the underwater explosion sound signal, the lower limit of the filter frequency f L and the upper limit f H , the signal length estimate T ch , the signal front intercept reserved time T ca , the short-time window length T S and the long-time window length T L , the long-short time characteristic function F(n) is selected;

[0054] Step two: after filtering the data x(n) to be processed according to the selected frequency upper and lower limits, according to the selected long-short time parameters T S , T L , F(n), the short-time average energy S TA (n), the long-time average energy L TA (n) and the long-short time average energy ratio N SL (n) at each data point n are calculated:

[0055]

[0056] Wherein, L s = T s *f s and L L = T L *f s are the short-time window point number and the long-time window point number calculated according to the window length T S and T L ; according to the long-short time average energy ratio N SL (n), the effectiveness of the characteristic function F(n) and the size of the explosion sound signal triggering threshold T h are determined, and the long-short time window length is adjusted appropriately according to the actual situation;

[0057] Step three: according to the specified long-short time parameters and the filter frequency upper and lower limits, the recorded data are processed in batches in a cycle, and for each recorded data x(n), the long-short time average energy ratio N SL(n) and the long-short time average energy ratio N SL (n) and the set threshold T h are compared to obtain a trigger determination result. Specifically:

[0058] The equipment records x m (n) data points, m is the file serial number saved according to the equipment's own setting parameters in the experimental measuring line, m = 1, 2,..., M, M is the total number of files recorded on the experimental measuring line, n is the number of data points, n = 1, 2,..., N, N is the total number of data points of each recorded file;

[0059] The M recorded files on the experimental measuring line are processed in batches in sequence, and the long-short time average energy ratio N m (n) of each data point x SL-m (n) is calculated. SL-m (n) and the set explosion sound signal trigger threshold T h are compared to obtain a trigger determination result:

[0060]

[0061] At the same time, according to the signal length estimate T ch , the trigger determination J cen-m (n) = 1 is further processed, and for data length T ch * f s , the first data point n1 and other trigger data points n that meet the trigger condition in the internal satisfy:

[0062]

[0063] Where n1 is the first data point of the underwater explosion sound signal automatic interception trigger, and the automatic interception trigger time t n = n1 / f s , according to the interception duration T S and the signal pre-reserved duration T ca , the underwater explosion sound signal in the file data x m (n) is intercepted and saved.

[0064] According to the signal time-frequency characteristics, a suitable characteristic function is selected, which can effectively reflect the changes of signal amplitude and energy, and contains the amplitude change trend and frequency domain characteristics of the signal.

[0065] According to the properties of ocean waveguide, the short-time window length T S and the long-time window length T L are appropriately selected.

[0066] According to the ocean waveguide properties and signal time-frequency characteristics, appropriate thresholds are selected in signal triggering and automatic interception.

[0067] Simulation example:

[0068] For actual sea area underwater acoustic experiment, a series of underwater explosion sound signals are received by the underwater acoustic equipment, and 40 record files are saved, each record file data is x m (n), n = 1, 2,...N, N = 900000, m = 1, 2,...M, M = 40, the sampling frequency of the underwater acoustic equipment is f s = 250Hz, that is, the length of each record file is 1 hour;

[0069] According to the data point value, a Hanning window function with a length of 2 12 and an overlap amount of 2 11 is used to perform short-time Fourier transform on random record file data, the short-time Fourier transform point number is 2 10 , the time-frequency distribution of the record data is calculated, and the results are shown in Figures 2(a) and 2(b), which show that the recorded data has a high signal-to-noise ratio, and the underwater explosion sound signal has good medium-low frequency spectrum characteristics;

[0070] According to the time-frequency spectrum distribution of the record data, and considering the characteristic quantity concerned in underwater acoustic signal processing, the lower limit f L = 20Hz and the upper limit f H = 120Hz of the filter frequency, the signal length estimation value T ch = 60s, the signal front interception reserved time T ca = 20s, the short-time window length T S = 2s and the long-time window length T L = 8s, and the long-short time characteristic function F(n) are selected:

[0071] F(n) = |x(n)-x(n-1)|;

[0072] According to the determined filter frequency upper and lower limits and long-short time parameters, signal length estimation value, etc., the short-time average energy S TA (n) of each data point n in the record file, the long-time average energy L TA (n) and the long-short time average energy ratio N SL (n) are preliminarily calculated, and the trigger threshold T h = 1.5 is determined, and the long-short time energy ratio distribution is shown in Figure 3, which shows that the long-short time parameters and the trigger threshold are effective.

[0073] Based on the determined upper and lower limits of the filtering frequency and the long and short time parameters, batch loop processing is performed on M=40 record files. The long and short time windows are shifted to determine the triggering situation, as shown in Figure 3. For each trigger value of 1, the signal is truncated and saved according to the estimated signal length and the reserved time value before signal truncation. The truncated results for the vector channel and scalar channel of the underwater acoustic device are shown below. Figure 4 As shown, the results demonstrate that this method can effectively achieve automatic interception of underwater explosion acoustic signals, significantly reducing the time and workload of manual interception. It also avoids data interception errors caused by time drift and asynchrony between different devices, improving signal interception quality. This method can effectively achieve rapid and accurate interception of underwater explosion acoustic signals without the need for manual estimation of the detonation time, greatly saving signal interception time. It provides effective support for subsequent research on underwater acoustic signal analysis and underwater acoustic detection.

[0074] 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 automatic intercepting of underwater explosion sound signals based on long-short time average energy ratio, the method comprising: Step 1) selecting an underwater explosion sound signal file from experimental line recording results of a hydroacoustic device, and performing time-frequency analysis on the recorded data to obtain a time-frequency spectrum; Step 2) setting filter upper and lower limits, determining signal intercepting duration and signal pre-reserved duration according to time-frequency spectrum characteristics of explosion sound signals; Step 3) calculating long-short time average energy ratio of filtered data according to the set filter upper and lower limits; Step 4) determining long-short time parameters and explosion sound signal triggering threshold in combination with the long-short time average energy ratio; Step 5) moving long-short time time windows according to the specified long-short time parameters and filter upper and lower limits, performing batch processing on the recording data of the entire line, calculating long-short time average energy ratio of each data point of each file in the line, and comparing the long-short time average energy ratio with the explosion sound signal triggering threshold to obtain triggering determination results of each file; Step 6) intercepting underwater explosion sound signals of each file according to the triggering determination results of each file and the determined signal intercepting duration and signal pre-reserved duration.

2. The method for automatically intercepting underwater explosion acoustic signals based on the long-short time average energy ratio according to claim 1, characterized in that, The step 1) specifically comprises: selecting an underwater explosion sound signal file from experimental line recording results of a hydroacoustic device, and the recorded data is x(n), n is the number of data points, n=1, 2,...N, N is the total number of data points; The short-time Fourier transform is used to obtain the transformed time-frequency spectrum P STFT (t,f) is: wherein s(τ) represents recorded data in time domain, w(τ-t) is a time window function shifted by t time length; t represents a time variable in the time-frequency spectrum, f represents a frequency variable in the time-frequency spectrum, τ represents a time point corresponding to the recorded data, and i represents an imaginary part.

3. The method for automatically intercepting underwater explosion acoustic signals based on the long-short time average energy ratio according to claim 2, characterized in that, The step 2) specifically comprises: According to the time-frequency spectrum characteristics of the underwater explosion sound signal, the lower limit of the filtering frequency is f L , the upper limit is f H , the interception time length is T S , and the signal front reserved time length is T ca .

4. The method for automatically intercepting underwater explosion acoustic signals based on the long-short time average energy ratio according to claim 3, characterized in that, The step 3) specifically comprises: According to the selected long and short time characteristic function F(n), the preliminarily set short time window length T S and long time window length T L , the short time average energy S TA (n) and long time average energy L TA (n) at each data point n after filtering are calculated. wherein L S is the short time window point number, L s = T s * f s , L L is the long time window point number, L L = T L * f s , f s is the sampling frequency, and j represents each data point within the time window range; The long-short time average energy ratio N is obtained according to the following formula SL (n) is:

5. The method according to claim 4, wherein the long-short time average energy ratio is defined as: ###0001### where E is the long-time average energy, E is the short-time average energy, and T is the time interval of the long-time average energy. 5 The long-short time characteristic function F(n) satisfies the following formula: F(n)=K1|x(n)|+K2|x(n)-x(n-1) wherein |x(n)| represents amplitude characteristics of the signal to be processed, |x(n)-x(n-1)| represents amplitude variation trend and frequency characteristics of the signal, and K1 and K2 are weight values set according to signal sampling frequency and inherent noise properties of the device.

6. The method for automatically intercepting underwater explosion acoustic signals based on the long-short time average energy ratio according to claim 5, characterized in that, The short time window length T S And the long time window length T L The selection is based on the time domain characteristics of underwater explosion sound signals and the characteristics of ocean ambient noise, and is preliminarily determined.

7. The method for automatically intercepting underwater explosion acoustic signals based on the long-short time average energy ratio according to claim 6, characterized in that, The step 4) specifically comprises: According to the long-short time average energy ratio N SL (n), the effectiveness of the characteristic function F(n) and the size of the explosion sound signal triggering threshold T h are determined, and the short time window length T S and the long time window length T L are further adjusted.

8. The method for automatically intercepting underwater explosion acoustic signals based on the long-short time average energy ratio according to claim 7, characterized in that, The step 5) specifically comprises: According to the specified long-short time parameters and the upper and lower limits of the filtering frequency, the long-short time window is moved, and each record file of the experimental measuring line is traversed, wherein the data of the mth record file is x m (n), n is the number of data points, n = 1, 2,...N, N is the total number of data points of each record file, m is the file serial number saved according to the device itself setting parameters in the experimental measuring line, m = 1, 2,...M, M is the total number of recorded files on the experimental measuring line; Compute x m (n) the long-short time average energy ratio N SL-m (n) of the data point n of the file, compare the long-short time average energy ratio N SL-m (n) of the data point n of the file with the set explosion sound signal trigger threshold T h (n) and obtain the trigger determination result J cen-m (n) of the mth record file 9. The method for automatically intercepting underwater explosion acoustic signals based on the long-short time average energy ratio according to claim 8, characterized in that, The step 6) specifically comprises: According to the signal length estimation value T ch Trigger determination J cen-m Further processing in the case of (n) = 1 for data length T ch *f s The first data point n1 and other trigger data points n that satisfy the trigger condition within the inner satisfy the following formula: Wherein, n1 is the automatic intercept trigger data point of the underwater explosion sound signal, and the automatic intercept trigger time t n = n1 / f s ; According to the intercepting duration T S and the signal pre-reserved duration T ca , the underwater explosion sound signal in the data x m (n) of the recording file is intercepted and saved.

Images