Motion sound source weak sound feature extraction method

Abstract

The invention discloses a motion sound source weak sound feature extraction method. The method comprises the following steps: step 1, processing a motion sound source time domain signal received by a receiver to obtain energy time history data of each frequency / frequency band; 2, selecting one frequency / frequency band, removing pulse interference of energy time history data of the frequency / frequency band by adopting a filtering algorithm, and carrying out moving average processing; 3, performing N-order Fourier series decomposition on the frequency / frequency band energy time history data; step 4, keeping the first M orders of the Fourier series amplitude, and performing N-order inverse Fourier transform after the rest order amplitudes are zeroed; and step 5, removing the background noise fixed component according to the criterion of removing the background noise fixed component from the received signal to obtain the maximum value of the signal energy of the frequency / frequency band at the positive transverse moment. According to the invention, the background noise fixed component and pulse interference can be removed from the received signal, the signal-to-noise ratio of the received signal is improved, and the weak sound feature extraction of the moving sound source is realized.

Description

Technical Field

[0002] The present invention relates to signal processing technology, and particularly to a method for extracting weak sound characteristics of a moving sound source. Background Art

[0004] When measuring the sound energy of a moving sound source in an environment with interference and background noise, the signal received by the receiver simultaneously includes the moving sound source signal, background noise, and impulse interference. When the background noise level is relatively high, the moving sound source signal is often submerged in the background noise, resulting in a low signal-to-noise ratio of the received signal, and it is impossible to accurately extract the weak sound characteristics of the moving sound source and accurately evaluate the magnitude. Summary of the Invention

[0006] Object of the Invention: In order to overcome the deficiencies in the prior art, the present invention provides a method for extracting weak sound characteristics of a moving sound source. Starting from the characteristic that the energy of the moving sound source signal at the receiver end changes with time due to the relative position change between the moving sound source and the receiver from far to near and then from near to far during the measurement process, through the Fourier series decomposition and inverse Fourier transform methods, and based on the criterion for removing the fixed component of the background noise from the received signal, a method for removing the fixed component of the background noise and random interference is established to improve the signal-to-noise ratio of the received signal and achieve the extraction of the weak sound characteristics of the moving sound source.

[0007] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0008] A method for extracting weak sound characteristics of a moving sound source, comprising the following steps:

[0009] Step 1: For the time-domain signal of the moving sound source received by the receiver, with the beam-on moment as the center, select a period of data with a certain time length, segment the time-domain signal of the moving sound source received by the receiver at equal time intervals, equal time lengths, and a fixed aliasing rate, calculate the power spectrum of each frequency / band in each segment of the time-domain signal within a certain frequency range, and sort the calculation results of the power spectrum of the same frequency / band in chronological order to obtain the energy time history data of each frequency / band;

[0010] Step 2: For the energy time history data of each frequency / band in Step 1, select one of the frequencies / band, use a filtering algorithm to remove the impulse interference in the energy time history data of this frequency / band and perform a moving average process to obtain the processed energy time history data of this frequency / band;

[0011] Step 3: Perform -order Fourier series decomposition on the processed energy time history data of this frequency / band;

[0012] Step 4. For the Fourier series amplitude obtained by the Nth-order Fourier series decomposition, retain the first order, set the amplitudes of the remaining orders to zero, and then perform the Nth-order inverse Fourier transform to obtain an array ;

[0013] Step 5. According to the criterion for removing the fixed component of the background noise from the received signal, remove the fixed component of the background noise from the array obtained in Step 4 to obtain the maximum value of the signal energy at the beam reach moment of this frequency / band. The criterion for removing the fixed component of the background noise from the received signal is as follows:

[0014] Set the threshold to , with the unit of , substitute the maximum value and the minimum value of the array into it, and determine whether Equation (1) holds:

[0015] (1)

[0016] If Equation (1) holds, assign to the maximum value of the signal energy at the beam reach moment of this frequency / band :

[0017] ;

[0018] Otherwise, subtract the fixed small value from the array , substitute the maximum value and the minimum value of the array into it, and determine whether Equation (2) holds:

[0019] (2)

[0020] If Equation (2) holds, assign to the maximum value of the signal energy at the beam reach moment of this frequency / band :

[0021] ;

[0022] Otherwise, continue to subtract the fixed small value from the array , and so on, until the value that makes Equation (3) hold is obtained:

[0023] (3)

[0024] Assign The maximum value of the signal energy at the beam aspect moment of the frequency / band :

[0025] ;

[0026] To prevent the situation of excessive calculation times or non-convergence, a reasonable fixed small value needs to be set according to the average energy of the received signal during the processing period and the maximum number of calculations , if the maximum number of calculations set is exceeded and equation (3) still cannot hold, then the maximum value of the array described in step 4 is assigned to the maximum value of the signal energy at the beam aspect moment of the frequency / band , that is ;

[0027] For each of the frequencies / bands, repeat steps 2 to 5 until the maximum values of the signal energies at the beam aspect moments of all frequencies / bands are obtained. The total energy of the sound energy of the moving sound source received within a certain frequency range can be further calculated by summing the energies.

[0028] Preferably, the time-domain signal of the moving sound source received in step 1 is composed of the superposition of the sound source signal, background noise, and impulse interference; among them, the sound source signal is a smooth, slowly changing function with a maximum value near the center of the data, tending to the edge of the data, and the signal decays by no less than a certain magnitude from the maximum value; the main component of the background noise is a Gaussian random process with a fixed component; the impulse interference can be a random impulse interference of a single data point or a random impulse interference composed of multiple consecutive data points; the sound source signal can be lower than the background noise by a certain magnitude; the beam aspect moment described in step 1 is the moment when the distance between the moving sound source and the receiver is the closest.

[0029] Preferably, the certain frequency range described in step 1 is determined according to the moving sound source and the receiver, the frequency is the line spectrum frequency, and the band is the 1 / 3 octave band or the 1 octave band.

[0030] Preferably, for the selection of one of the frequencies / bands in step 2, a filtering algorithm is used to remove the impulse interference from the energy time history data of the frequency / band and perform a moving average process. The specific steps are as follows:

[0031] Step S2-1: Remove the single-point random impulse interference from the energy time history data of the frequency / band

[0032] Step S2-2: Remove the random impulse interference composed of two consecutive data points from the energy time history data of the frequency / band

[0033] Step S2-3: Perform moving average processing on the energy time history data of this frequency / band after removing pulse interference.

[0034] Preferably, the previous order refers to the order containing the main energy of this frequency / band.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] Starting from the characteristic that the received signal energy from a moving sound source changes with time, the present invention adopts filtering algorithms, Fourier series decomposition and inverse Fourier transform methods, and based on the criterion of removing the fixed component of background noise from the received signal, removes the fixed component of background noise and pulse interference from the received signal with low signal-to-noise ratio, improves the signal-to-noise ratio of the received signal, and realizes the extraction of weak sound characteristics of the moving sound source. Description of the Drawings

[0038] Figure 1 is a flowchart of a method for extracting weak sound characteristics of a moving sound source according to the present invention;

[0039] Figure 2 is a signal comparison diagram before and after removing pulse interference by adopting a filtering algorithm and moving average according to an embodiment of the present invention;

[0040] Figure 3 is an energy time history curve of a moving sound source signal according to an embodiment of the present invention;

[0041] Figure 4 is a background noise curve according to an embodiment of the present invention;

[0042] Figure 5 is an energy time history curve of a received signal mixed with a moving sound source signal, background noise and pulse interference according to an embodiment of the present invention;

[0043] Figure 6 is an energy time history curve of a moving sound source signal obtained after removing background noise and pulse interference from the received signal according to an embodiment of the present invention. Detailed Embodiments

[0045] The following clearly and completely describes the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. It should be understood that these examples are only used to illustrate the present invention and are not used to limit the scope of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0046] AsFigure 1 As shown in Figure 1 , an embodiment of the present invention discloses a method for extracting weak sound characteristics of a moving sound source, including the following steps:

[0047] Step 1: Process the time-domain signal of the moving sound source received by the receiver to obtain the energy time history data of each frequency / band;

[0048] Step 2: Select one of the frequencies / bands, and use a filtering algorithm to remove the impulse interference of the energy time history data of this frequency / band and perform a moving average process;

[0049] Step 3: Decompose the energy time history data of this frequency / band into N-order Fourier series;

[0050] Step 4: Retain the first M orders of the Fourier series amplitudes, set the amplitudes of the remaining orders to zero, and then perform an N-order inverse Fourier transform;

[0051] Step 5: According to the criterion for removing the fixed component of the background noise from the received signal, remove the fixed component of the background noise to obtain the maximum value of the signal energy at the beam aspect moment of this frequency / band.

[0052] The following further describes the above steps in combination with specific embodiments.

[0053] In one embodiment, in Step 1, for the time-domain signal of the moving sound source received by the receiver, with the beam aspect moment as the center, select data with a length of 263 s, segment the selected data at a time interval of 1 s, a data segment time length of 8 s, and an aliasing rate of 87.5%, obtain 256 data segments, calculate the power spectrum of each 1 / 3 octave band of each segment signal within the concerned frequency range, and sort the calculation results of the power spectrum of the same 1 / 3 octave band in chronological order to obtain the energy time history data of 256 points for each 1 / 3 octave band.

[0054] In one embodiment, in Step 1, for the time-domain signal of the moving sound source received by the receiver, with the beam aspect moment as the center, select a data segment with a certain time length, segment the selected data at an equal time interval, equal time length, and a fixed aliasing rate, calculate the power spectrum of each line spectrum frequency of each segment signal within the concerned frequency range, and sort the calculation results of the power spectrum of the same line spectrum frequency in chronological order to obtain the energy time history data of each line spectrum frequency;

[0055] In one embodiment, in Step 2, select one of the frequencies / bands, and the energy time history data of this frequency / band is expressed as , and use a filtering algorithm to remove the impulse interference of the energy time history data of this frequency / band and perform a moving average process, as described in Steps S2-1, S2-2, and S2-3.

[0056] Step S2-1: Remove the single-point random pulse interference in the frequency / band energy time history data. The specific method is as follows:

[0057] Analyze every three consecutive data values. If is satisfied, no change is made to the data. Here, represents the set of data .

[0058] (4)

[0059] If , then replace in the data with .

[0060] If , then replace in the data with .

[0061] In the compared sequence , when , only use each number in the original data.

[0062] Step S2-2: Remove the random pulse interference composed of two consecutive data points in the frequency / band energy time history data. The specific method is as follows:

[0063] Analyze every six consecutive data values:

[0064] First, use inequalities (5) - (8) to test the random pulse interference composed of two consecutive data points. If all four inequalities (5) - (8) hold, then assign to , that is, ; if any one of the inequalities (5) - (8) does not hold, no replacement is made in the data. Here, represents the set of data , represents the set of data .

[0065] (5)

[0066] (6)

[0067] (7)

[0068] (8)

[0069] Then, use inequalities (9) to (12) to continue to check the random pulse interference composed of two consecutive data points. If all four inequalities (9) to (12) hold, then assign to be , that is ; if any one of the inequalities (9) to (12) does not hold, then no replacement is made in the data.

[0070] (9)

[0071] (10)

[0072] (11)

[0073] (12)

[0074] In the compared sequence , when , only use each number in the original data.

[0075] Step S2-3: Perform a moving average process on the energy time history data of this frequency / band after removing the pulse interference, specifically, perform a moving average process of three consecutive points.

[0076] In this embodiment, the energy time history data of the 1 / 3 octave band with a center frequency of 200 Hz of the received signal is generated through simulation calculation. The filtering algorithm is used to first remove the single-point pulse interference of the energy time history data of the received signal in this band, then remove the pulse interference composed of two consecutive points, and perform a three-point moving average process. The processing result is as Figure 2 shown. It can be seen from Figure 2 that using the filtering algorithm and the moving average process can obtain an energy time history curve of the 1 / 3 octave band of the received signal with less fluctuation.

[0077] In one embodiment, simulation calculation is used to give the result of weak sound feature extraction of a moving sound source. As Figure 3 shown, the energy time history data of the 1 / 3 octave band with a center frequency of 1 kHz of the received moving sound source signal is set to be a smooth, slowly changing function with a maximum value of 80 dB near the center of the data, tending to the edge of the data, and the signal decays by 6 dB from the maximum value. As Figure 4 shown, the background noise is set to be a Gaussian random process with a mean of 86 dB and a standard deviation of 0.45 dB, and there are random pulse interferences distributed. As Figure 5 shown, the energy time history of the 1 / 3 octave band of the received signal is composed of the superposition of the moving sound source signal, background noise and pulse interference, and is composed of Figure 5It can be seen that the moving sound source signal is submerged in the background noise, and the signal-to-noise ratio of the received signal is relatively low. In this embodiment, set . In step 3, the energy time history data in the 1 / 3 octave band is decomposed by a 256-order Fourier series. In step 4, the first 8 orders of the Fourier series amplitudes are retained, and after setting the amplitudes of the remaining orders to zero, a 256-order inverse Fourier transform is performed. In step 5, set the threshold , set the maximum number of calculations , set the fixed small value , where represents the average value of the array . As Figure 6 shows, after removing the fixed component of the background noise and the impulse interference, the maximum value of the signal energy at the beam aspect moment in this frequency band is 79.8 dB, which is close to the set maximum value of the signal energy at the beam aspect moment in this frequency band of 80 dB. From the simulation calculation results of Figure 6 , it can be seen that when the energy of the moving sound source signal is 6 dB lower than the background noise, the method of the present invention can effectively extract the moving sound source signal from the received signal.

[0078] It should be understood that the above embodiments of the present invention are only used for exemplary illustration or explanation of the principle of the present invention, and do not constitute a limitation to the present invention. Therefore, any modifications, equivalent replacements, improvements, etc. made without departing from the spirit and scope of the present invention shall be included within the protection scope of the present invention. In addition, the appended claims of the present invention are intended to cover all variations and modifications that fall within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. A method for extracting the weak sound characteristics of a moving sound source, characterized in that, It includes the following steps: Step 1: For the time-domain signal of the moving sound source received by the receiver, select a data segment of a certain time length centered on the beam-crossing moment. Segment the received time-domain signal of the moving sound source at equal time intervals, equal time lengths, and a fixed aliasing rate. Calculate the power spectrum of each frequency / band in each time-domain signal segment within a certain frequency range. Sort the calculation results of the power spectrum of the same frequency / band in chronological order to obtain the energy time history data of each frequency / band; Step 2: For the energy time history data of each frequency / band in Step 1, select one of the frequencies / bands, and use a filtering algorithm to remove the impulse interference in the energy time history data of this frequency / band and perform a moving average process to obtain the processed energy time history data of this frequency / band; Step 3. Perform order Fourier series decomposition on the processed frequency / frequency band energy time history data described in Step 2; Step 4. For the Fourier series amplitude obtained by the Nth-order Fourier series decomposition, retain the first order, set the amplitudes of the remaining orders to zero, and then perform the Nth-order inverse Fourier transform to obtain the array ; Step 5. According to the criterion for removing the fixed component of background noise from the received signal, for the array described in Step 4 Remove the fixed component of background noise to obtain the maximum value of the signal energy at the beam - reach moment of this frequency / band , the method for the criterion for removing the fixed component of background noise from the received signal is as follows: Set the threshold value to , with the unit of , substitute into the maximum value and the minimum value of the array , and determine whether equation (1) holds: （1） If equation (1) holds, assign to the maximum value of the signal energy at the beam aspect moment of this frequency / band : ； Otherwise, subtract the fixed small value from the array , substitute into the maximum value and the minimum value of the array , and determine whether equation (2) holds: （2） If equation (2) holds, assign to the maximum value of the signal energy at the beam aspect moment of this frequency / band : ； Otherwise, continue to subtract the fixed value small from the array , and so on until the value that makes equation (3) hold is obtained: （3） Assign to the maximum value of the signal energy at the beam aspect moment of this frequency / band : ； To prevent the situation of excessive calculation times or non-convergence, it is necessary to set a reasonable fixed small value according to the average energy of the received signal during the processing period and the maximum number of calculations . If the maximum number of calculations is exceeded and the condition of formula (3) still cannot be satisfied, then assign the maximum value of the array described in step 4 to the maximum value of the signal energy at the beam aspect moment of this frequency / band , that is ; For each of the frequencies / bands, repeat Step 2 to Step 5 until the maximum value of the signal energy at the beam-crossing moment of all frequencies / bands is obtained. The total energy of the radiated sound of the moving sound source received within the certain frequency range can be further calculated by summing the energies.

2. The method for extracting weak sound characteristics of a moving sound source according to claim 1, wherein The time-domain signal of the moving sound source received by the receiver in Step 1 is composed of the source signal, background noise, and impulse interference superimposed. Among them, the source signal is a smooth, slowly changing function with a maximum value near the center of the data, tending to the edge of the data, and the signal decays by no less than a certain magnitude from the maximum value; the main component of the background noise is a Gaussian random process containing a fixed component; the impulse interference can be a random impulse interference of a single data point or a random impulse interference composed of multiple consecutive data points; the source signal can be lower than the background noise by a certain magnitude; the beam-crossing moment in Step 1 is the moment when the distance between the moving sound source and the receiver is the closest.

3. The method for extracting the weak sound characteristics of a moving sound source according to claim 1, wherein, The certain frequency range in Step 1 is determined according to the moving sound source and the receiver. The frequency is the line spectrum frequency, and the band is the 1 / 3 octave band or the 1 octave band.

4. The method for extracting the weak sound characteristics of a moving sound source according to claim 1, wherein In Step 2, when selecting one of the frequencies / bands and using a filtering algorithm to remove the impulse interference in the energy time history data of this frequency / band and perform a moving average process, the specific steps are as follows: Step S2-1: Remove the single-point random impulse interference in the energy time history data of this frequency / band; Step S2-2: Remove the random impulse interference composed of two consecutive data points in the energy time history data of this frequency / band; Step S2-3: Perform a moving average process on the energy time history data of this frequency / band after removing the impulse interference.

5. The method for extracting weak sound characteristics of a moving sound source according to claim 1, characterized in that, The previous stage described in step 4 refers to the order containing the main energy of the frequency / band.

Images