A method and device for reducing noise of track acoustic signal

Abstract

The present application discloses a method and device for reducing noise of track acoustic signals, which relates to the field of rail transit technology. The short-time Fourier transform is used to process the track acoustic signal, and the full-band energy value of the track acoustic signal corresponding to each moment is obtained, and the energy of the characteristic frequency band of the corrugation at each moment and the proportion of the full-band energy sum of the track acoustic signal at the same moment are calculated. The present application uses the energy of the common frequency band of the corrugation signal and the proportion of the full-band energy sum of the track acoustic signal at the same moment as the weight of the acoustic signal excited by the driving track to reduce irrelevant noise, which can effectively avoid the defects of the loss of soundprint features caused by methods such as digital filters and wavelet threshold noise reduction, and can also effectively avoid the defects of "music noise" generated when the traditional spectral subtraction method uses reference noise for noise reduction, and can achieve the purpose of noise reduction to highlight the characteristics of track corrugation without losing the soundprint features of the track acoustic signal.

Description

Technical Field

[0001] The present application relates to the field of rail transit technology, and in particular to a method and device for reducing noise of rail acoustic signals. Background Art

[0002] Because rail-based trains run at high speeds for a long time, coupled with complex and changeable environmental conditions, periodic uneven deformation wear will occur on the track. This wear is called rail corrugation.

[0003] Rail corrugation can cause a lot of additional wheel-rail dynamics, resulting in additional traction energy consumption. This additional dynamic effect may cause severe vibration of the train and even lead to derailment accidents. Therefore, it is necessary to promptly identify whether there is rail corrugation on the track where the train is running.

[0004] In the prior art, acoustic sensors are used to collect track acoustic signals when vehicles are running, and the track acoustic signals are input into the recognition model to determine whether the track has corrugation damage. However, the track acoustic signals collected by the acoustic sensors contain a lot of noise interference, which can easily cause the recognition model to misjudge track corrugation.

[0005] In order to solve the problem of noise interference in the track acoustic signal collected by the acoustic sensor, noise reduction technology is used in the prior art to process the acoustic signal. However, these existing noise reduction technologies will filter out the useful soundprint features in the track acoustic signal, achieving the purpose of noise reduction while sacrificing the integrity of the track acoustic signal. Summary of the invention

[0006] In view of the above problems, the present application provides a method and device for reducing noise of track acoustic signals, so as to achieve the purpose of reducing noise without destroying the integrity of the acoustic signal. The specific scheme is as follows:

[0007] The first aspect of the present application provides a method for reducing noise of a track acoustic signal, comprising:

[0008] Acquiring track acoustic signals;

[0009] The track acoustic signal is processed by short-time Fourier transform to obtain a frequency energy group; the frequency energy group includes the full-band energy value of the track acoustic signal corresponding to each moment at different moments;

[0010] According to the frequency energy group, the energy of the corrugation characteristic frequency band in the track acoustic signal at each moment and the proportion of the total frequency band energy of the track acoustic signal at the same moment are calculated to obtain an initial track acoustic signal weight set; the initial track acoustic signal weight set includes the initial track acoustic signal weight corresponding to each moment;

[0011] The track acoustic signal weight corresponding to each moment of the track acoustic signal is obtained by upsampling the initial track acoustic signal weight set, and the track acoustic signal is processed based on the track acoustic signal weight corresponding to each moment to obtain the track acoustic signal after noise reduction.

[0012] In a possible implementation, short-time Fourier transform is used to process the track acoustic signal to obtain a frequency energy group, including:

[0013] Perform short-time Fourier transform processing on the track acoustic signal to obtain the processing result of the short-time Fourier transform;

[0014] Take the modulus in the processing result to obtain the full-band energy distribution matrix in the time dimension;

[0015] The full-band energy distribution matrix in the time dimension is regarded as the frequency energy group.

[0016] In a possible implementation, based on the frequency energy group, the ratio of the energy of the corrugation characteristic frequency band in the track acoustic signal at each moment to the total energy of the full frequency band of the track acoustic signal at the same moment is calculated to obtain an initial track acoustic signal weight set, including:

[0017] Determine the corrugation characteristic frequency band in the full frequency band of the track acoustic signal at each moment according to the preset corrugation characteristic frequency band range;

[0018] Calculate the energy sum of the corrugation characteristic frequency band in the track acoustic signal at each moment, and the energy sum of the full frequency band of the track acoustic signal at each moment;

[0019] Calculate the energy of the corrugation characteristic frequency band at the same moment and its proportion in the energy of the full frequency band of the track acoustic signal to obtain the initial track acoustic signal weight corresponding to each moment;

[0020] Based on the initial track acoustic signal weight corresponding to each moment, an initial track acoustic signal weight set is obtained.

[0021] In a possible implementation, the initial track acoustic signal weight set is upsampled to obtain the track acoustic signal weight corresponding to each moment of the track acoustic signal, and the track acoustic signal is processed based on the track acoustic signal weight corresponding to each moment to obtain the track acoustic signal after noise reduction, including:

[0022] The length of the initial track acoustic signal weight set is expanded by upsampling to obtain an upsampled track acoustic signal weight set that is consistent with the length of the original track acoustic signal;

[0023] The track acoustic signal weight corresponding to each moment in the upsampled track acoustic signal weight set is used for weighting, and the track acoustic signal corresponding to the moment is processed to obtain the track acoustic signal after noise reduction.

[0024] In a possible implementation, before acquiring the track acoustic signal, the method further includes:

[0025] Obtaining raw track acoustic signals from acoustic sensors;

[0026] The original track acoustic signal is normalized to obtain a preprocessed track acoustic signal.

[0027] A second aspect of the present application provides a track acoustic signal noise reduction device, comprising:

[0028] An acquisition unit, a first processing unit, a calculation unit, and a second processing unit; wherein:

[0029] An acquisition unit, used for acquiring a track acoustic signal;

[0030] A first processing unit is used to process the track acoustic signal by using short-time Fourier transform to obtain a frequency energy group; the frequency energy group includes a full-band energy value of the track acoustic signal corresponding to each moment at different moments;

[0031] A calculation unit is used to calculate the proportion of the energy of the corrugation characteristic frequency band in the track acoustic signal at each moment and the sum of the full frequency band energy of the track acoustic signal at the same moment according to the frequency energy group, so as to obtain an initial track acoustic signal weight set; the initial track acoustic signal weight set includes the initial track acoustic signal weight corresponding to each moment;

[0032] The second processing unit is used to obtain the track acoustic signal weight corresponding to each moment of the track acoustic signal by upsampling the initial track acoustic signal weight set, and process the track acoustic signal based on the track acoustic signal weight corresponding to each moment to obtain the track acoustic signal after noise reduction.

[0033] In a possible implementation, the first processing unit is configured to:

[0034] Perform short-time Fourier transform processing on the track acoustic signal to obtain the processing result of the short-time Fourier transform;

[0035] Take the modulus of the processing result to obtain the energy distribution matrix of the full frequency band in the time dimension;

[0036] The full-band energy distribution matrix in the time dimension is regarded as the frequency energy group.

[0037] A third aspect of the present application provides a computer program product, comprising computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements the track acoustic signal noise reduction method as described in any one of the above.

[0038] In a fourth aspect, the present application provides a track acoustic signal noise reduction device, comprising at least one processor and a memory connected to the processor, wherein:

[0039] The memory is used to store computer programs;

[0040] The processor is used to execute the computer program so that the track acoustic signal noise reduction device can implement the track acoustic signal noise reduction method as described in any one of the above.

[0041] A fifth aspect of the present application provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement a track acoustic signal noise reduction method as described in any one of the above.

[0042] By means of the above technical scheme, the method and device for reducing noise of track acoustic signals provided by the present application use short-time Fourier transform to process the track acoustic signals, obtain the full-band energy value of the track acoustic signals corresponding to each moment in different moments, and calculate the energy of the characteristic frequency band of the corrugation at each moment and the proportion of the total energy of the full-band of the track acoustic signal at the same moment. The present application uses the energy of the common frequency band of the corrugation signal and the total energy of the full-band of the track acoustic signal at the same moment as the weight of the acoustic signal excited by the vehicle-track to reduce irrelevant noise, which can effectively avoid the defects of the loss of soundprint features caused by methods such as digital filters and wavelet threshold noise reduction, and can also effectively avoid the defects of "music noise" generated by the traditional spectral subtraction method when using reference noise for noise reduction, and can achieve the purpose of noise reduction to highlight the characteristics of track corrugation without losing the soundprint features of the track acoustic signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0044] Figure 1 Example diagram of rail corrugation damage scenario provided for this application;

[0045] Figure 2 A schematic diagram of a process of a method for reducing noise of track acoustic signals provided in this application;

[0046] Figure 3 An example flow chart of a method for reducing noise of track acoustic signals provided in this application;

[0047] Figure 4 An example diagram of the noise reduction effect of the noise reduction method provided in this application on rail corrugation acoustic signals;

[0048] Figure 5 This is an example diagram of the noise reduction effect of the noise reduction method provided in this application on normal track acoustic signals;

[0049] Figure 6 A structural diagram of the track acoustic signal noise reduction device provided in this application;

[0050] Figure 7 This is a schematic diagram of the structure of the track acoustic signal noise reduction device provided in this application. DETAILED DESCRIPTION

[0051] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0052] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0053] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0054] Rail corrugation, also known as rail waveform wear or track corrugation, mainly refers to the wear phenomenon of regular longitudinal fluctuations on the top surface of the rail.

[0055] For example, see Figure 1 , an example diagram of a rail corrugation damage scenario provided in this application.

[0056] The presence of rail corrugation not only causes additional power consumption during train operation, but also accelerates damage to the track structure. For example, it causes rail surface wear and permanent deformation of the track bed, which not only increases the frequency and cost of maintenance, but also may affect the safety of train operation.

[0057] In view of the many negative effects of rail corrugation mentioned above, various methods for identifying rail corrugation have been developed in the prior art:

[0058] Method 1: Use traditional vibration sensors. Specifically, install the vibration sensors near the wheels of the train, and can be installed on the axle box of the train. The vibration sensors collect the vibration signals of the train and the track, and then process and analyze the vibration signals. Based on the analysis results, it is determined whether the train track has corrugation damage and the severity of the damage.

[0059] Method 2: Use the emerging acoustic sensor. The acoustic sensor is flexible to install and can be installed near the door in the car. In addition, the acoustic sensor has the advantages of high sensitivity and wide signal acquisition bandwidth. Therefore, the acoustic sensor quickly became the main signal acquisition device in the rail corrugation identification method.

[0060] Specifically, the acoustic sensor collects the track acoustic signal when the vehicle is running, and inputs the track acoustic signal into the recognition model to determine whether the track has corrugation damage. However, the track acoustic signal collected by the acoustic sensor contains a lot of noise interference, making it difficult for the recognition model to accurately determine whether the track has corrugation damage.

[0061] In order to solve the problem of noise interference in the track acoustic signal collected by the acoustic sensor, the existing technology generally uses various filters to filter out the noise in the track acoustic signal, such as digital filter, wavelet threshold filter, EMD (Empirical Mode Decomposition) and VMD (Variational Mode Decomposition).

[0062] However, the above-mentioned methods for filtering noise in track acoustic signals in the prior art will inevitably filter out useful voiceprint features in track acoustic signals.

[0063] For digital filters and wavelet threshold filters, these two filters usually distinguish track acoustic signals from noise based on frequency. Inevitably, the useful track acoustic signal will overlap with the noise frequency range, which will result in the useful track acoustic signal being mistakenly removed while filtering out the noise.

[0064] For EMD and VMD, both noise reduction methods distinguish useful track acoustic signals from noise by decomposing the signal. Inevitably, useful voiceprint features in useful track acoustic signals, such as specific frequency components, may overlap with the frequency range of noise, which will result in the removal of useful voiceprint features while filtering out noise.

[0065] It can be seen that the method for filtering noise in the track acoustic signal in the known technology can achieve the purpose of noise reduction only on the basis of sacrificing the track acoustic signal.

[0066] In order to solve the above problems, the present application proposes and designs a method and device for reducing noise of track acoustic signals.

[0067] See also Figure 2 , a schematic flow chart of a method for reducing noise of track acoustic signals provided in the present application.

[0068] like Figure 2 As shown, the method for reducing noise of track acoustic signals includes the following steps:

[0069] Step 101: Acquire track acoustic signals.

[0070] It should be noted that the track acoustic signal is the track acoustic signal including noise before noise reduction, and the track acoustic signal here is specifically the time domain acoustic signal of the train-track. Moreover, what is obtained here is the track acoustic signal after preprocessing, and the preprocessing process can be as follows:

[0071] An original track acoustic signal is acquired from an acoustic sensor, and the original track acoustic signal is normalized to obtain a preprocessed track acoustic signal.

[0072] Specifically, after obtaining the original train-track time domain acoustic signal from the acoustic sensor, the original train-track time domain acoustic signal is normalized. Suppose the normalized train-track time domain acoustic signal is , the normalization formula is as follows:

[0073]

[0074] in, is the normalized train-track time domain acoustic signal, is the original train-track time domain acoustic signal; is the minimum value of the original train-track time domain acoustic signal collected by the acoustic sensor. Correspondingly, It is the maximum value of the original train-track time domain acoustic signal collected by the acoustic sensor.

[0075] In the above normalized formula, the numerator The denominator is the original train-track time domain acoustic signal minus the minimum value of this signal. is the maximum value of the original train-track time domain acoustic signal Subtract the minimum value in this signal The purpose of the normalization process in this application is to eliminate the dimensional error of the original acquired train-track time-domain acoustic signal.

[0076] The present application is not limited to the above-mentioned method of eliminating dimensional errors, and other methods may also be used.

[0077] It should be noted that the acoustic sensor used to obtain the original track acoustic signal in this application can also be called a voiceprint acquisition terminal. The voiceprint acquisition terminal includes any sensor combination that can collect the original track acoustic signal, such as a capacitive microphone array, a MEMS (Micro-Electro-Mechanical Systems array) array, a patch voiceprint sensor, and a distributed acoustic sensor system.

[0078] Among them, capacitive microphones have high sensitivity and a wide frequency response range, and can capture a wide range of sound signals from low frequency to high frequency; MEMS arrays allow the manufacture of very small microphone arrays, which can be flexibly installed in different parts of rail vehicles to achieve all-round sound collection; patch-type voiceprint sensors can be installed on the surface of the equipment and can obtain track acoustic signals using non-invasive detection methods; distributed acoustic sensor systems are composed of multiple independent acoustic sensor nodes, which are distributed in different locations in the train compartments, and can form a widely covered monitoring network to comprehensively capture sound signals.

[0079] It is understandable that when different acoustic sensor combinations are used to collect track acoustic signals in the present application, it is necessary to eliminate the dimensional errors from the different acoustic sensors.

[0080] The voiceprint collection terminal described above can be installed in the car near the door.

[0081] Optionally, a normalized train-track time domain acoustic signal is obtained.

[0082] Specifically, the time domain acoustic signal of the train-track is a signal waveform that changes with time, describing the value of the signal at different time points. In the time domain, the independent variable is time, the horizontal axis represents time, and the vertical axis represents the amplitude or intensity of the track acoustic signal.

[0083] Step 102: Process the track acoustic signal using short-time Fourier transform to obtain a frequency energy group; the frequency energy group includes the full-band energy value of the track acoustic signal corresponding to each moment at different moments.

[0084] It should be noted that the short-time Fourier transform can divide the track acoustic signal into multiple short, overlapping time windows, and perform Fourier transform on the time domain acoustic signal in each window, so as to obtain the instantaneous frequency information of the track acoustic signal in the frequency domain. It can be understood that each divided time point in each time window divided in the short-time Fourier transform process corresponds to each moment in the different moments in the frequency energy group.

[0085] The results obtained after short-time Fourier transform of the track acoustic signal mainly include the real part and the imaginary part, where the real part is the energy distribution of different frequencies of the track acoustic signal, and the imaginary part reflects the phase information. The frequency energy group is specifically the module after the track acoustic signal is processed by short-time Fourier transform. The specific form of the module is the energy distribution matrix containing phase information, that is, the energy distribution matrix of the full frequency band in the time dimension.

[0086] Optionally, the track acoustic signal is subjected to short-time Fourier transform processing and modulo-processed to obtain a time-frequency domain matrix.

[0087] Exemplarily, the formula for performing short-time Fourier transform processing on the track acoustic signal may be as follows:

[0088] x ( t i , f j ) = abs | STFT ( x ( t ))| = [ x t 1 , f n x t 2 , f n ... x t n , f n ... ... ... ... x t 1 , f 2 x t 2 , f 2 ... x t n , f 2 x t 1 , f 1 x t 2 , f 1 ... x t n , f 1 ]

[0089] Among them, x(t i , f i ) represents the result of taking the modulus after short-time Fourier transform of the track acoustic signal. represents the short-time Fourier transform of the track acoustic signal. Express The result after transformation is modulo;

[0090] [ x t 1 , f n x t 2 , f n ... x t n , f n ... ... ... ... x t 1 , f 2 x t 2 , f 2 ... x t n , f 2 x t 1 , f 1 x t 2 , f 1 ... x t n , f 1 ] is the frequency energy group;

[0091] in, represents the energy value at time t1 and frequency f1, represents the energy value at time t1 and frequency f2, Indicates that at time t1, the frequency is f n The energy value at represents the energy value at time t2 and frequency f1, represents the energy value at time t2 and frequency f2, Indicates that at time t2, the frequency is f n The energy value at Indicates that at t n The energy value at time and frequency f1, Indicates that at t n The energy value at time and frequency f2, Indicates that at t n Time, frequency is f n The energy value at time.

[0092] It should be noted that the short-time Fourier transform divides the track acoustic signal into multiple short, overlapping time windows during the process of processing the track acoustic signal, and performs Fourier transform on the track acoustic signal in units of the intercepted windows to capture the local characteristics of the track acoustic signal in time and frequency. Therefore, short-time Fourier processing of the track acoustic signal will shorten the time resolution.

[0093] Exemplarily, the acquisition time of the track acoustic signal is assumed to be 10s, and the track acoustic signal is acquired every 0.1s; the track acoustic signal is subjected to short-time Fourier transform, with a window length of 0.5s, that is, after the short-time Fourier transform is processed, a time-frequency feature corresponds to every 0.5s.

[0094] Step 103: Calculate the ratio of the energy of the corrugation characteristic frequency band in the track acoustic signal at each moment to the total energy of the full frequency band of the track acoustic signal at the same moment based on the frequency energy group, and obtain an initial track acoustic signal weight set; the initial track acoustic signal weight set includes the initial track acoustic signal weight corresponding to each moment.

[0095] It should be noted that the background noise in the track acoustic signal collected by the acoustic sensor is generally processed as Gaussian white noise. Compared with the frequency domain characteristics of the corrugation signal, Gaussian white noise is evenly distributed in the frequency domain and has lower energy.

[0096] The track acoustic signal at the same moment when track corrugation occurs includes corrugation characteristic frequencies and noise characteristic frequencies, which are generally different; correspondingly, different frequency band energy values ​​of the corrugation signal and frequency band energy values ​​of the noise signal are included at the same moment. The present application proposes to use the proportion of energy within the range of the corrugation characteristic frequency at the same moment to represent the proportion of the corrugation characteristics in the track acoustic signal at that moment. The higher the proportion, the higher the content of the corrugation characteristics in the track acoustic signal, and the more important the track acoustic signal at that moment.

[0097] Optionally, the corrugation characteristic frequency band in the full frequency band of the track acoustic signal at the same moment is determined according to a preset corrugation characteristic frequency band range, the energy sum of the corrugation characteristic frequency band and the energy sum of the full frequency band of the track acoustic signal are calculated, and the proportion of the energy sum of the corrugation characteristic frequency band in the energy sum of the full frequency band of the track acoustic signal is calculated.

[0098] The preset wave grinding characteristic frequency band range is the preset wave grinding characteristic frequency range. Exemplarily, it can be known from subway statistical data that the frequency range of the wave grinding signal is generally about 40Hz to 600Hz, so the frequency signal within this frequency range at the same time can be approximately regarded as the wave grinding signal, and the wave grinding characteristic frequency band located in the above preset wave grinding characteristic frequency band range and the energy value corresponding to the wave grinding characteristic frequency band are determined from the time-frequency domain matrix (i.e., the full-band energy distribution matrix), and the energy sum of each wave grinding characteristic frequency band is calculated. Then, the energy sum of the full frequency band at the same time is calculated. It should be noted that the full frequency band of the track acoustic signal for wave grinding identification is generally within 1000Hz. Finally, the ratio between the energy sum of the wave grinding characteristic frequency band and the energy sum of the full frequency band is calculated, that is, the proportion of the energy sum of the wave grinding characteristic frequency band in the energy sum of the full frequency band. This proportion can be used as the signal weight of the wave grinding signal in the track acoustic signal, and the signal weight can also be used as the signal weight of the useful track acoustic signal at each moment in the time domain.

[0099] Exemplarily, taking time t1 as an example, the energy sum of all corrugation characteristic frequency bands in the range of n1 to n2 at time t1 in the time-frequency domain matrix is ​​calculated. The calculation formula may be as follows:

[0100] ;

[0101] in, It represents the energy sum of all corrugation characteristic frequency bands within the range of corrugation characteristic frequency bands n1 to n2 at time t1.

[0102] Then, calculate the full frequency band f1 to f at time t1 in the time-frequency domain matrix n The energy and within the range. The calculation formula can be as follows:

[0103] ;

[0104] Finally, the ratio of the energy of the corrugation characteristic frequency band at time t1 to the total energy of the full frequency band at this time, w(t1), is calculated using the following formula:

[0105]

[0106] Among them, w(t1) is the time domain signal weight at time t1, is the energy sum of all corrugation characteristic frequency bands within the range of corrugation characteristic frequency bands n1 to n2 at time t1, is the full frequency band f1 to f at time t1 n Energy within the range and.

[0107] Afterwards, the time domain signal weight corresponding to each signal acquisition time point is obtained according to the length of the track acoustic signal after short-time Fourier transform processing, and the initial track acoustic signal weight set is composed of the time domain signal weights corresponding to different moments. As mentioned above, short-time Fourier processing of the track acoustic signal will shorten the time resolution, so the time resolution of each initial track acoustic signal weight corresponding to the obtained initial track acoustic signal weight set is low. That is, the length of the initial track acoustic signal weight set composed of the time domain signal weights corresponding to different moments is inconsistent with the length of the original track acoustic signal. Therefore, the initial track acoustic signal weight set cannot be directly used to process the original track acoustic signal. The initial track acoustic signal weight set needs to be processed to the same length as the original track acoustic signal. The specific processing process is detailed in step 104.

[0108] Step 104: The initial track acoustic signal weight set is upsampled to obtain the track acoustic signal weight corresponding to each moment of the track acoustic signal, and the track acoustic signal is processed based on the track acoustic signal weight corresponding to each moment to obtain a denoised track acoustic signal.

[0109] It should be noted that the initial track acoustic signal weight set is calculated according to the length after short-time Fourier transform processing, but the length of the initial track acoustic signal weight set is inconsistent with the length of the original track acoustic signal, and the time domain weight signal in the initial track acoustic signal weight set cannot be weighted with the initial track acoustic signal for noise reduction.

[0110] Optionally, the present application uses upsampling to expand the length of the initial track acoustic signal weight set, and expands the length of the initial track acoustic signal weight set to be consistent with the length of the original track acoustic signal. It should be noted that the means of upsampling can specifically be linear interpolation, etc. The expansion principle of linear interpolation is: based on known data points, more unknown values ​​between these data are estimated to achieve the purpose of upsampling. The core idea is to assume that the changes between known data points are linear, and expand the data length through multiple linear predictions.

[0111] Assume that the initial track acoustic signal weight set is w(t i ), the initial track acoustic signal is x(t), the length of the time domain signal weight in the initial track acoustic signal weight set is inconsistent with the length of the original track acoustic signal, and weighting cannot be performed directly. Specifically, the formula for linear interpolation expansion can be:

[0112] ;

[0113] Among them, w(t i ) is t i The time domain signal weight at time, w(t i+1 ) is t i+1The time domain signal weight at time t can be obtained by upsampling through linear interpolation. j The expected signal weight at each moment .

[0114] Based on this, the track acoustic signal weight set is processed by linear interpolation to obtain an upsampled track acoustic signal weight set that is consistent with the length of the original track acoustic signal.

[0115] Then, the track acoustic signal weight corresponding to each moment in the upsampled track acoustic signal weight set is weighted with the track acoustic signal at the same moment to obtain the denoised track acoustic signal.

[0116] Specifically, the track acoustic signal after noise reduction is assumed to be , the calculation formula of the track acoustic signal after noise reduction is:

[0117]

[0118] in, is the track acoustic signal after noise reduction, To transpose the original track acoustic signal x(t), It is the signal weight after the initial time domain signal weight is upsampled to the same length as the original track acoustic signal.

[0119] For example, see Figure 3 , a flow chart showing an example of a method for reducing noise of track acoustic signals provided in the present application.

[0120] Firstly, the measured track acoustic signal is obtained, and the track acoustic signal is short-time Fourier transformed to obtain a frequency energy group. Then, the corrugation characteristic frequency band is selected based on the preset corrugation characteristic frequency band range, that is, the corrugation characteristic frequency band in the track acoustic signal is determined. Then, the time-frequency energy ratio of the corrugation characteristic frequency band is calculated, and the time-frequency energy ratio is used as the initial signal weight set. The initial signal weight set is upsampled to obtain a signal weight set consistent with the length of the track acoustic signal. Finally, the signal weight set obtained after upsampling is weighted with the track acoustic signal according to the signal acquisition time point to obtain the denoised track acoustic signal.

[0121] Optional, see Figure 4 , Figure 4This is an example diagram of the noise reduction effect of the noise reduction method provided in the present application on the rail corrugation signal. Among them, the noise reduction effect diagram on the left is an example diagram of the effect after the original corrugation signal is denoised by various filters in the prior art, and the noise reduction effect diagram on the right is an example diagram of the effect after the original corrugation signal is denoised by the noise reduction method provided in the present application. Obviously, the waveform of the noise-reduced corrugation signal obtained by the noise reduction method provided in the present application is highly similar to the waveform of the original corrugation signal. Therefore, the noise reduction method provided in the present application is superior to the noise reduction method for the original corrugation signal in the prior art in retaining the useful soundprint features in the original corrugation signal.

[0122] See also Figure 5 , Figure 5 This is an example diagram of the noise reduction effect of the noise reduction method provided in this application on the normal track acoustic signal. Among them, the noise reduction effect diagram on the left is an example diagram of the effect of using various filters in the prior art to reduce the noise of the original normal signal, and the noise reduction effect diagram on the right is an example diagram of the effect of using the noise reduction method provided in this application to reduce the noise of the original normal signal. Obviously, the waveform of the normal signal after noise reduction obtained by using the noise reduction method provided in this application is highly similar to the waveform of the original normal signal. Therefore, the noise reduction method provided in this application is superior to the noise reduction method for the original normal signal in the prior art in retaining the useful voiceprint features in the original normal signal.

[0123] Depend on Figure 4 and Figure 5 It can be seen that the noise reduction method of track acoustic signals provided in this application is effective.

[0124] In summary, the track acoustic signal denoising method provided by the present application uses short-time Fourier transform to process the track acoustic signal, obtains the full-band energy value of the track acoustic signal corresponding to each moment at different moments, and calculates the energy of the corrugation characteristic frequency band at each moment and the proportion of the full-band energy sum of the track acoustic signal at the same moment. The present application uses the energy of the common frequency band of the corrugation signal and the proportion of the full-band energy sum of the track acoustic signal at the same moment as the acoustic signal weight excited by the vehicle-track to reduce irrelevant noise, which can effectively avoid the defects of the loss of soundprint features caused by methods such as digital filters and wavelet threshold denoising, and can also effectively avoid the defects of 'music noise' generated by the traditional spectral subtraction method using reference noise denoising, and can achieve the purpose of noise reduction to highlight the track corrugation features without losing the soundprint features of the track acoustic signal.

[0125] For the above-mentioned track acoustic signal noise reduction method, the present application also provides a track acoustic signal noise reduction device. Figure 6 , a composition structure diagram of the track acoustic signal noise reduction device provided in this application.

[0126] like Figure 6As shown, the track acoustic signal noise reduction device includes an acquisition unit 10, a first processing unit 20, a calculation unit 30 and a second processing unit 40; wherein:

[0127] An acquisition unit 10, used for acquiring a track acoustic signal;

[0128] The first processing unit 20 is used to process the track acoustic signal by using short-time Fourier transform to obtain a frequency energy group; the frequency energy group includes a full-frequency energy value of the track acoustic signal corresponding to each moment at different moments;

[0129] The calculation unit 30 is used to calculate the energy of the corrugation characteristic frequency band in the track acoustic signal at each moment and the proportion of the total frequency band energy of the track acoustic signal at the same moment according to the frequency energy group, so as to obtain an initial track acoustic signal weight set; the initial track acoustic signal weight set includes the initial track acoustic signal weight corresponding to each moment;

[0130] The second processing unit 40 is used to execute step 4: obtain the track acoustic signal weight corresponding to each moment of the track acoustic signal by upsampling the initial track acoustic signal weight set, and process the track acoustic signal by each track acoustic signal weight to obtain the track acoustic signal after noise reduction.

[0131] In one embodiment, the first processing unit 20 is specifically configured to:

[0132] Perform short-time Fourier transform processing on the track acoustic signal to obtain the processing result of the short-time Fourier transform;

[0133] Take the modulus in the processing result to obtain the full-band energy distribution matrix in the time dimension;

[0134] The full-band energy distribution matrix in the time dimension is regarded as the frequency energy group.

[0135] In one embodiment, the calculation unit 30 is specifically configured to:

[0136] Determine the corrugation characteristic frequency band in the full frequency band of the track acoustic signal at each moment according to the preset corrugation characteristic frequency band range;

[0137] Calculate the energy sum of the corrugation characteristic frequency band in the track acoustic signal at each moment, and the energy sum of the full frequency band of the track acoustic signal at each moment;

[0138] Calculate the energy of the corrugation characteristic frequency band at the same moment and its proportion in the energy of the full frequency band of the track acoustic signal to obtain the initial track acoustic signal weight corresponding to each moment;

[0139] Based on the initial track acoustic signal weight corresponding to each moment, an initial track acoustic signal weight set is obtained.

[0140] In one embodiment, the second processing unit 40 is specifically configured to:

[0141] The length of the initial track acoustic signal weight set is expanded by upsampling to obtain an upsampled track acoustic signal weight set that is consistent with the length of the original track acoustic signal;

[0142] The track acoustic signal weight corresponding to each moment in the upsampled track acoustic signal weight set is used for weighting, and the track acoustic signal corresponding to the moment is processed to obtain the track acoustic signal after noise reduction.

[0143] In one embodiment, the track acoustic signal noise reduction device further includes a preprocessing unit, which is specifically used to:

[0144] Obtaining raw track acoustic signals from acoustic sensors;

[0145] The original track acoustic signal is normalized to obtain a preprocessed track acoustic signal.

[0146] An embodiment of the present application also provides a computer program product, including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any one of the track acoustic signal noise reduction methods provided in the embodiments of the present application.

[0147] The present application also provides a track acoustic signal noise reduction device. Figure 7 As shown, it shows a schematic diagram of the structure of the track acoustic signal noise reduction device suitable for implementing the embodiment of the present application. The track acoustic signal noise reduction device in the embodiment of the present application may include but is not limited to fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 7 The track acoustic signal noise reduction device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0148] like Figure 7 As shown, the track acoustic signal noise reduction device may include a processing device (such as a central processing unit, a graphics processor, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 to a random access memory (RAM) 603. When the track acoustic signal noise reduction device is powered on, various programs and data required for the operation of the electronic device are also stored in the RAM 603. The processing device 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0149] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a memory card, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 7 The track acoustic signal noise reduction device with various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0150] A computer storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any one of the track acoustic signal noise reduction methods provided in the embodiment of the present application.

[0151] It should also be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0152] Through the description of the above implementation mode, the technicians in the field can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present application, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0153] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0154] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, a computer, a training device or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, training device or data center. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integration. Available media can be magnetic media, (e.g., floppy disk, hard disk, tape), optical media (e.g., DVD), or semiconductor media (e.g., solid-state hard disk (SSD)), etc.

Claims

1. A method for reducing noise of track acoustic signals, characterized in that: include: Acquiring track acoustic signals; Processing the track acoustic signal by short-time Fourier transform to obtain a frequency energy group; The frequency energy group includes the full-band energy value of the track acoustic signal corresponding to each moment at different moments; According to the frequency energy group, the energy of the corrugation characteristic frequency band in the track acoustic signal at each moment and the proportion of the total frequency band energy of the track acoustic signal at the same moment are calculated to obtain an initial track acoustic signal weight set; the initial track acoustic signal weight set includes the initial track acoustic signal weight corresponding to each moment; The initial track acoustic signal weight set is upsampled to obtain a track acoustic signal weight corresponding to each moment of the track acoustic signal, and the track acoustic signal is processed based on the track acoustic signal weight corresponding to each moment to obtain a track acoustic signal after noise reduction; The method of processing the track acoustic signal by short-time Fourier transform to obtain a frequency energy group includes: Performing short-time Fourier transform processing on the track acoustic signal to obtain a short-time Fourier transform processing result; Taking the modulus of the processing result, obtaining a full-band energy distribution matrix in the time dimension; Using the full-band energy distribution matrix in the time dimension as the frequency energy group; The step of obtaining the track acoustic signal weight corresponding to each moment of the track acoustic signal by upsampling the initial track acoustic signal weight set, and processing the track acoustic signal based on the track acoustic signal weight corresponding to each moment to obtain the track acoustic signal after noise reduction includes: Expanding the length of the initial track acoustic signal weight set by upsampling to obtain an upsampled track acoustic signal weight set that is consistent with the length of the original track acoustic signal; The track acoustic signal weight corresponding to each moment in the up-sampled track acoustic signal weight set is used for weighting, and the track acoustic signal corresponding to the moment is processed to obtain the denoised track acoustic signal.

2. The method for reducing noise of track acoustic signals according to claim 1, characterized in that: The method of calculating the energy of the corrugation characteristic frequency band in the track acoustic signal at each moment and the proportion of the energy of the full frequency band of the track acoustic signal at the same moment based on the frequency energy group to obtain an initial track acoustic signal weight set includes: Determine the corrugation characteristic frequency band in the full frequency band of the track acoustic signal at each moment according to the preset corrugation characteristic frequency band range; Calculating the energy sum of the corrugation characteristic frequency band in the track acoustic signal at each moment, and the energy sum of the full frequency band of the track acoustic signal at each moment; Calculate the ratio of the energy of the corrugation characteristic frequency band to the energy of the full frequency band of the track acoustic signal at the same moment, and obtain the initial track acoustic signal weight corresponding to each moment; Based on the initial track acoustic signal weight corresponding to each moment, the initial track acoustic signal weight set is obtained.

3. The method for reducing noise of track acoustic signals according to claim 1, characterized in that: Before acquiring the track acoustic signal, it also includes: Obtaining raw track acoustic signals from acoustic sensors; The original track acoustic signal is normalized to obtain the preprocessed track acoustic signal.

4. A track acoustic signal noise reduction device, characterized in that: include: An acquisition unit, a first processing unit, a calculation unit, and a second processing unit; wherein: The acquisition unit is used to acquire the track acoustic signal; The first processing unit is used to process the track acoustic signal by short-time Fourier transform to obtain a frequency energy group; the frequency energy group includes a full-band energy value of the track acoustic signal corresponding to each moment at different moments; The calculation unit is used to calculate the proportion of the corrugation characteristic frequency band energy in the track acoustic signal at each moment and the total frequency band energy of the track acoustic signal at the same moment according to the frequency energy group, so as to obtain an initial track acoustic signal weight set; the initial track acoustic signal weight set includes the initial track acoustic signal weight corresponding to each moment; The second processing unit is configured to obtain a track acoustic signal weight corresponding to each moment of the track acoustic signal by upsampling the initial track acoustic signal weight set, and process the track acoustic signal based on the track acoustic signal weight corresponding to each moment to obtain a track acoustic signal after noise reduction; Wherein, the first processing unit is used to: Performing short-time Fourier transform processing on the track acoustic signal to obtain a short-time Fourier transform processing result; Taking the modulus of the processing result, obtaining a full-band energy distribution matrix in the time dimension; Using the full-band energy distribution matrix in the time dimension as the frequency energy group; Wherein, the second processing unit is used for: The step of obtaining a track acoustic signal weight corresponding to each moment of the track acoustic signal by upsampling the initial track acoustic signal weight set, and processing the track acoustic signal based on the track acoustic signal weight corresponding to each moment to obtain a track acoustic signal after noise reduction includes: Expanding the length of the initial track acoustic signal weight set by upsampling to obtain an upsampled track acoustic signal weight set that is consistent with the length of the original track acoustic signal; The track acoustic signal weight corresponding to each moment in the up-sampled track acoustic signal weight set is used for weighting, and the track acoustic signal corresponding to the moment is processed to obtain the denoised track acoustic signal.

5. A computer program product, characterized in that The invention comprises computer-readable instructions, and when the computer-readable instructions are executed on an electronic device, the electronic device implements the method for reducing noise of a track acoustic signal according to any one of claims 1 to 3.

6. A track acoustic signal noise reduction device, characterized in that: The method comprises at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the track acoustic signal noise reduction device can implement the track acoustic signal noise reduction method according to any one of claims 1 to 3.

7. A computer storage medium, characterized in that: The storage medium carries one or more computer programs, and when the one or more computer programs are executed by an electronic device, the electronic device can implement the track acoustic signal noise reduction method as described in any one of claims 1 to 3.

Images