CN115376540A - Biological radar voice enhancement method and system based on variational modal decomposition - Google Patents
Biological radar voice enhancement method and system based on variational modal decomposition Download PDFInfo
- Publication number
- CN115376540A CN115376540A CN202210993119.8A CN202210993119A CN115376540A CN 115376540 A CN115376540 A CN 115376540A CN 202210993119 A CN202210993119 A CN 202210993119A CN 115376540 A CN115376540 A CN 115376540A
- Authority
- CN
- China
- Prior art keywords
- decomposition
- mode
- radar
- radar voice
- voice
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 149
- 238000000034 method Methods 0.000 title claims abstract description 66
- 230000006870 function Effects 0.000 claims description 42
- 238000005070 sampling Methods 0.000 claims description 8
- 230000010355 oscillation Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 3
- 230000003595 spectral effect Effects 0.000 claims description 2
- 238000004590 computer program Methods 0.000 description 10
- 238000001514 detection method Methods 0.000 description 10
- 230000006872 improvement Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000005534 acoustic noise Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000009545 invasion Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
- G10L21/0216—Noise filtering characterised by the method used for estimating noise
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Speech or voice signal processing techniques to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
- G10L21/0216—Noise filtering characterised by the method used for estimating noise
- G10L21/0232—Processing in the frequency domain
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Quality & Reliability (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a biological radar voice enhancement method and system based on variational modal decomposition. Acquiring biological radar voice to obtain an original noisy radar voice signal; carrying out intrinsic mode function characteristics of empirical mode decomposition on the original noisy radar voice signal, and determining the number of layers of variable mode decomposition according to the intrinsic mode function characteristics; carrying out variational modal decomposition VMD on the original noisy radar voice according to the decomposition layer number to obtain each order of intrinsic modal decomposition; solving Pearson coefficients for each order of eigenmode decomposition and original radar voice signals to determine a useful decomposition reconstruction mode; and denoising the useful modes by using the improved threshold value, and reconstructing to obtain the enhanced radar voice. The invention has stronger adaptability and can obviously improve the quality and intelligibility of the voice on the basis of eliminating the noise in the radar voice. Therefore, the method has stronger use value and application prospect in the aspect of radar voice enhancement.
Description
Technical Field
The invention relates to the technical field of voice enhancement, in particular to a biological radar voice enhancement method and system based on variational modal decomposition.
Background
The biological radar voice detection technology overcomes the defects that the traditional microphone voice detection device is easily interfered by environmental acoustic noise, the throat microphone and the bone conduction microphone need to contact the detection device with the skin of a human body, and the optical voice detection sensor is easily influenced by environmental factors such as temperature, climate and the like, and has the advantages of non-contact, non-invasion, safety, good directivity, high sensitivity, long detection distance, strong anti-interference capability and the like. The biological radar technology provides a new way for acquiring the voice signal. At present, a 94GHz biological radar can provide a good compromise between detection distance and detection sensitivity, but still has the problem of short detection distance, and the radar voice using electromagnetic waves as media is often superimposed with circuit noise, electromagnetic interference noise and harmonic noise. These noises reduce the quality of radar voice to a large extent. Therefore, the research on the radar voice enhancement method has important significance for the development of the new voice detection technology.
Disclosure of Invention
The invention aims to provide a method for enhancing a 94GHz asymmetric antenna biological radar voice, which determines the layer number of variable mode decomposition according to empirical mode decomposition, then performs variable mode decomposition on an original radar voice signal according to the layer number, and finally effectively removes the noise content in the radar voice by utilizing an improved threshold strategy. Technical support in the aspect of voice denoising is provided for the development of the biological radar voice detection technology.
In order to achieve the purpose, the invention adopts the following technical scheme.
A biological radar voice enhancement method based on variational modal decomposition comprises the following steps:
acquiring a biological radar voice to obtain an original radar voice signal with noise;
carrying out intrinsic mode function characteristics of empirical mode decomposition on the original noisy radar voice signal, and determining the number of variable mode decomposition layers according to the intrinsic mode function characteristics;
performing variable modal decomposition VMD on the original radar voice with noise according to the decomposition layer number to obtain each order of intrinsic modal decomposition;
solving Pearson coefficients for each order of eigenmode decomposition and original radar voice signals to determine a useful decomposition reconstruction mode;
and denoising the useful modes by using the improved threshold value, and reconstructing to obtain the enhanced radar voice.
As a further improvement of the invention, the acquisition of the biological radar voice is to utilize an 8-channel Powerlab physiological recorder to sample the biological radar voice, and the sampling rate is 10kHz.
As a further improvement of the present invention, the determining the number of layers of the variation modal decomposition according to the eigenmode function characteristics of the empirical modal decomposition includes:
given a voice signal as x (t), the empirical mode decomposition method adaptively decomposes the original signal into local oscillation mode functions through a screening process, and each local oscillation mode function IMF is a sub-frequency component of the original signal; analyzing the amplitude characteristics of each IMF from high frequency to low frequency, wherein when the amplitude value of the IMF is smaller than a set value, the mode is voice detail information, the rest modes are main information, and the number of the main information modes is the number of layers of variational mode decomposition;
from top to bottom, the IMF order of the IMF after EMD decomposition, of which the first appearance amplitude value is smaller than a set value, is a main information mode number and is obtained by the following formula:
in the formula, IMF i For each order of the eigenmode function after empirical mode decomposition, i represents the order of the eigenmode function.
As a further improvement of the present invention, the performing variational modal decomposition VMD on the original noisy radar voice according to the number of decomposition layers to obtain each order of eigenmodal decomposition includes:
given an original radar voice signal as x (t), each mode u is obtained through Hilbert transformation k And constructing a variation constraint problem:
decomposing x (t) into K modes, where u k (K =1,2, …, K) is the K modes of the VMD method, w k (K =1,2, …, K) is the center frequency of each mode; introduce lagrange multiplier:
wherein, α is a penalty factor for controlling bandwidth, λ (t) is Lagrange multiplier, which represents convolution, δ (t) unit pulse function;
adopting a multiplicative operator alternating direction method and Parseval Fourier equidistant transformation, converting each mode from a time domain to a frequency domain by a formula (2) as follows:
wherein, each modal center frequency is:
wherein w k n+1 Is the power spectrum center of the kth mode.
As a further improvement of the present invention, the solving for the pearson coefficient for each order of eigenmode decomposition and the original radar voice signal includes:
wherein x is t And y t Representing two random variables.
As a further improvement of the present invention, the improved threshold is represented by the following formula:
N j the length of the jth scale signal, σ, is the estimated noise of each modal signal; wherein the improved estimated noise of each mode can be calculated by the following formula:
where L is the estimated length of the initial silence period of the radar speech signal.
As a further improvement of the invention, the useful mode is denoised by using an improved threshold value, and the method comprises the following steps:
wherein m is a compensation factor.
As a further improvement of the invention, the value of the compensation factor m is 0.001.
As a further improvement of the invention, the reconstruction to obtain the enhanced radar voice is to reconstruct the enhanced voice of the 94GHz radar voice denoised by the improved threshold strategy, and the specific reconstruction is calculated according to the following formula:
wherein, IMF is each mode of K of VMD decomposition, and a is useful mode number determined by calculating Pearson coefficient.
A biological radar voice enhancement system based on variational modal decomposition comprises the following steps:
the acquisition module is used for acquiring the biological radar voice to obtain an original noisy radar voice signal;
the decomposition module is used for carrying out intrinsic mode function characteristics of empirical mode decomposition on the original noisy radar voice signal and determining the number of variable-fractional mode decomposition layers according to the intrinsic mode function characteristics; carrying out variational modal decomposition VMD on the original noisy radar voice according to the decomposition layer number to obtain each order of intrinsic modal decomposition; solving Pearson coefficients for each order of eigenmode decomposition and original radar voice signals to determine a useful decomposition reconstruction mode;
and the reconstruction module is used for denoising the useful mode by utilizing the improved threshold value and reconstructing to obtain the enhanced radar voice.
Due to the adoption of the technology, compared with the existing radar voice enhancement technology, the invention has the following beneficial effects:
according to the method, after empirical mode decomposition is carried out on the acquired 94GHz asymmetric antenna radar voice signal, the number of decomposition layers is determined, variable mode decomposition is carried out according to the number of decomposition layers, then the noise variance is estimated according to the voice silence period information, the quality and intelligibility of voice are effectively improved by using an improved soft threshold method on the premise of not causing voice signal distortion, and the method has strong adaptability and effectiveness. The example of the method shows that the variable state decomposition method can effectively improve the quality and intelligibility of radar voice, and compared with the traditional voice enhancement method, the method can improve the scoring of the original radar voice in general. Therefore, the invention can effectively enhance the original radar voice signal. The invention can provide effective technical support for detecting human voice by using the biological radar. Therefore, the method has stronger use value and application prospect in the aspect of eliminating radar voice noise.
Drawings
FIG. 1 is a flow chart of a biological radar voice enhancement method based on variational modal decomposition and improved threshold strategy;
FIG. 2 is a section of an empirical mode decomposed radar speech material, where (a) is the original radar speech signal containing noise; (b) is the eigenmode function of each order after decomposition;
FIG. 3 is a section of a radar speech material after a metamorphic mode decomposition, where (a) is the eigenmode function of each order after the decomposition; (b) is the frequency spectrum of the eigenmode functions of each order;
FIG. 4 is a Pearson coefficient for the original Lee speech signal and the mode after the variational mode decomposition;
FIG. 5 is a raw radar voice signal collected by a radar system, where (a) is a time domain waveform of the radar voice signal; (b) is a spectrogram of a radar speech signal;
FIG. 6 is a radar speech signal enhanced by the wavelet soft threshold algorithm, wherein (a) is a time domain waveform of the enhanced radar speech signal; (b) is a spectrogram of an enhanced radar speech signal;
FIG. 7 is a radar speech signal enhanced by an empirical mode decomposition algorithm, wherein (a) is a time domain waveform of the enhanced radar speech signal; (b) is a spectrogram of an enhanced radar speech signal;
FIG. 8 is a radar speech signal enhanced based on a variational modal decomposition and refinement threshold strategy as described in the present patent, wherein (a) is a time domain waveform of the enhanced radar speech signal; and (b) enhancing the spectrogram of the radar voice signal.
Detailed Description
In order to make those skilled in the art better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The invention provides a biological radar voice enhancement method based on variational modal decomposition, which comprises the following steps:
1) Sampling the biological radar voice to obtain an original radar voice signal with noise;
in the step 1), an 8-channel Powerlab physiological recorder is used for sampling the biological radar voice, and the sampling rate is 10kHz.
2) Firstly, carrying out empirical mode decomposition on the collected original noisy radar voice signals, and determining the number of layers of variable mode decomposition according to the intrinsic mode function characteristics after the empirical mode decomposition;
determining the decomposition layer number by using empirical mode decomposition in the step 2) as follows:
assuming that a given voice signal is x (t), the empirical mode decomposition method adaptively decomposes the original signal into local oscillation mode functions through a screening process, and each local oscillation mode function IMF is a sub-frequency component of the original signal. And analyzing the amplitude characteristics of each IMF from high frequency to low frequency, and determining that the mode is voice detail information when the amplitude value of the IMF is less than 0.1, the rest modes are main information, and the number of the main information modes is the number of layers of the variational mode decomposition. Then, from top to bottom, the IMF order of the EMD after the EMD is the main information mode number, and can be obtained by the following formula:
3) Carrying out variation modal decomposition on the original radar voice with noise according to the decomposition layer number;
step 3) the method of metamorphic modal decomposition (VMD) is as follows:
given an original radar voice signal as x (t), each mode u is obtained by Hilbert transformation k And constructing a variation constraint problem:
decomposing x (t) into K modes, where u k (K =1,2, …, K) is the K modes of the VMD method, w k (K =1,2, …, K) is the center frequency of each mode. For solving each mode, the constraint problem of the formula (1) needs to be converted into a non-constraint variational problem. Introduce lagrange multiplier:
where α is a penalty factor for controlling bandwidth, λ (t) is a lagrange multiplier, which represents convolution, and δ (t) is a unit pulse function.
Adopting a multiplication operator alternating direction method and Parseval Fourier equidistant transformation, converting each mode from a time domain to a frequency domain by a formula (2) as follows:
wherein, the center frequency of each mode is:
wherein w k n+1 Is the power spectral center of the k-th mode.
4) Solving a Pearson coefficient for each order of eigenmode decomposition of the variation mode decomposition and the original radar voice signal; to determine useful decomposition reconstruction modalities;
5) Solving a Pearson coefficient for each order of eigenmode decomposition of the variational mode decomposition and an original radar voice signal; the method for determining the Pearson coefficient of each mode of the original radar voice signal and the VMD decomposition in the step 3) is as follows:
x t and y t Representing two random variables. For radar speech enhancement, however, experiments have shown that a correlation coefficient greater than 0.1 is a threshold for classifying it as a useful modality.
The improved threshold strategy method is as follows, and each modal adaptive soft threshold is represented by the following formula:
N j the length of the jth scale signal, σ, is the estimated noise of the individual modal signals. Wherein the improved estimated noise of each mode can be calculated by the following formula:
where L is the estimated length of the initial silence period of the radar speech signal.
The improved threshold strategy denoises the useful modes according to a soft threshold function improved by the following formula:
where m is a compensation factor to prevent distortion of the speech signal caused by the past noise.
The value of the compensation factor m is 0.001.
6) And denoising the useful modes by using the improved threshold value, and then reconstructing to obtain the enhanced radar voice.
The 94GHz radar voice denoised by the improved threshold strategy is reconstructed to obtain enhanced voice, and the specific reconstruction is calculated according to the following formula:
wherein, IMF is each mode of K of VMD decomposition, and a is useful mode number determined by calculating Pearson coefficient.
The following detailed description of the present invention is provided with reference to the accompanying drawings, but the present invention is not limited to the embodiments. In the following description of the preferred embodiments of the present invention, specific details are set forth in order to provide a thorough understanding of the present invention.
Referring to fig. 1, the basic principle of the method for decomposing and improving the threshold strategy based on the variation mode of the invention is as follows: firstly, decomposing acquired radar voice signals by empirical mode decomposition to determine the number of decomposition layers; then decomposing the acquired radar voice signals by empirical mode decomposition according to the number of decomposition layers; calculating the original radar voice signal and the Pearson coefficient of each order of modal function, and determining a useful mode; carrying out improved threshold strategy denoising treatment on a useful mode; and reconstructing the processed useful mode to obtain the enhanced voice.
For comparison, please refer to fig. 6, fig. 7, and fig. 8, wherein (a) is a time domain waveform of the radar voice signal, and (b) is a spectrogram of the radar voice signal.
As can be seen from fig. 5, noise exists in the original radar voice signal in the whole frequency band, and noise interference exists in a single frequency band, which results in low quality of the collected radar voice. FIG. 6 shows a radar speech signal enhanced by a wavelet soft threshold algorithm. As can be seen from the figure, the wavelet soft threshold effectively removes noise in radar speech. But observing the spectrogram, the horizontal stripes of the voice have certain deformation, new noise is also introduced, and in addition, residual noise still exists in a low frequency band.
Fig. 7 is a radar voice signal after being enhanced by empirical mode decomposition, and it can be seen from the figure that noise in the original radar voice is suppressed to a great extent, and it can be seen that voice quality is greatly improved in both time domain waveforms and spectrogram, but the visibility between horizontal stripes is not high when observing the spectrogram, and therefore, it is still insufficient in radar voice enhancement.
FIG. 8 shows the radar voice effect enhanced by the variation modal decomposition and threshold improvement strategy method. As can be seen from the figure, the noise of the radar voice enhanced by the method in the whole frequency band is weakened, and the objective evaluation method experiment carried out synchronously shows that the method can effectively improve the quality and intelligibility of the radar voice on the premise of not causing the distortion of the original radar voice signal.
Although the voice enhancement method based on the variational modal decomposition and the improved threshold strategy has pertinence and is particularly suitable for 94GHz asymmetric antenna biological radar voice signals, the application range of the method is not limited to the 94GHz asymmetric antenna radar voice signals, and the method also has important guiding significance and reference value for other millimeter wave and centimeter wave radar voice signals and acoustic voice signals collected under the same environment. The above embodiments of the present invention are described in detail with reference to the accompanying drawings, without any limitation to the present invention, and any simple modifications, changes and equivalent structural changes made on the above embodiments according to the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.
In order to ensure the consistency of sound sources, a recording material '1-2-3-4-5-6' of a male experimenter is selected and played by a sound box at a distance of 10 meters from a 94GHz asymmetric antenna biological radar system.
The radar voice is enhanced by the following steps:
1) Sampling the biological radar voice by using an 8-channel Powerlab physiological recorder, wherein the sampling rate is 10kHz; sampling to obtain an original noisy radar voice signal;
2) Decomposing the noisy radar voice signal into an intrinsic mode function through empirical mode decomposition, and determining the number of main information layers;
3) Carrying out variation modal decomposition on the noisy radar voice signal according to the determined decomposition layer number;
4) Solving a Pearson correlation coefficient of each decomposed mode and the original radar voice signal, and determining a useful mode;
5) Denoising useful modes by improving a threshold strategy;
6) Reconstructing the denoised useful mode to obtain the enhanced radar voice.
And (3) comparing treatment results:
as can be seen from fig. 5, noise exists in the original radar voice signal in the whole frequency band, and noise interference exists in a single frequency band, which results in low quality of the collected radar voice. FIG. 6 shows a radar speech signal enhanced by a wavelet soft threshold algorithm. As can be seen from the figure, the wavelet soft threshold effectively removes noise in radar speech. But observing the spectrogram, the horizontal stripes of the voice have certain deformation, new noise is also introduced, and in addition, residual noise still exists in a low frequency band. Fig. 7 is a radar voice signal after being enhanced by empirical mode decomposition, and it can be seen from the figure that noise in the original radar voice is suppressed to a great extent, and it can be seen that voice quality is greatly improved in both time domain waveforms and spectrogram, but the visibility between horizontal stripes is not high when observing the spectrogram, and therefore, it is still insufficient in radar voice enhancement. FIG. 8 shows the radar voice effect enhanced by the variation modal decomposition and improved threshold strategy method. As can be seen from the figure, the noise of the radar voice enhanced by the method in the whole frequency band is weakened, and the objective evaluation method experiment carried out synchronously shows that the method can effectively improve the quality and intelligibility of the radar voice on the premise of not causing the distortion of the original radar voice signal. The method is mainly characterized in that the original radar voice signal can be effectively decomposed into modal components with different frequencies, so that the occurrence of over-decomposition and under-decomposition can be effectively avoided, and modal aliasing can be inhibited.
In addition, the improved threshold strategy of the invention calculates the estimation value of the modal noise after the variational modal decomposition according to the average energy of the initial silent section, thereby having higher accuracy. In addition, the original soft threshold function is corrected, and the soft threshold function is corrected by using the compensation factor, so that not only can the excessive damage of the voice signal be effectively avoided, but also the residual noise in the noise radar voice can be effectively inhibited.
The invention also provides a biological radar voice enhancement system based on variational modal decomposition, which comprises the following steps:
the acquisition module is used for acquiring the biological radar voice to obtain an original noisy radar voice signal;
the decomposition module is used for carrying out intrinsic mode function characteristics of empirical mode decomposition on the original noisy radar voice signal and determining the number of variable-fractional mode decomposition layers according to the intrinsic mode function characteristics; carrying out variational modal decomposition VMD on the original noisy radar voice according to the decomposition layer number to obtain each order of intrinsic modal decomposition; solving Pearson coefficients for each order of eigenmode decomposition and original radar voice signals to determine a useful decomposition reconstruction mode;
and the reconstruction module is used for denoising the useful mode by utilizing the improved threshold value and reconstructing to obtain the enhanced radar voice.
The invention provides an electronic device, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the steps of the biological radar voice enhancement method based on the variational modal decomposition when executing the computer program.
The biological radar voice enhancement method based on variational modal decomposition comprises the following steps:
acquiring a biological radar voice to obtain an original radar voice signal with noise;
carrying out intrinsic mode function characteristics of empirical mode decomposition on the original noisy radar voice signal, and determining the number of layers of variable mode decomposition according to the intrinsic mode function characteristics;
carrying out variational modal decomposition VMD on the original noisy radar voice according to the decomposition layer number to obtain each order of intrinsic modal decomposition;
solving Pearson coefficients for each order of eigenmode decomposition and original radar voice signals to determine a useful decomposition reconstruction mode;
and denoising the useful modes by using the improved threshold value, and reconstructing to obtain the enhanced radar voice.
The present invention also provides a computer-readable storage medium, which stores a computer program that, when being executed by a processor, implements the steps of the voice enhancement method for biological radar based on variational modal decomposition.
The biological radar voice enhancement method based on the variational modal decomposition comprises the following steps of:
acquiring biological radar voice to obtain an original noisy radar voice signal;
carrying out intrinsic mode function characteristics of empirical mode decomposition on the original noisy radar voice signal, and determining the number of layers of variable mode decomposition according to the intrinsic mode function characteristics;
performing variable modal decomposition VMD on the original radar voice with noise according to the decomposition layer number to obtain each order of intrinsic modal decomposition;
solving Pearson coefficient for each order of eigenmode decomposition and original radar voice signal to determine useful decomposition reconstruction mode;
and denoising the useful modes by using the improved threshold value, and reconstructing to obtain the enhanced radar voice.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (10)
1. A biological radar voice enhancement method based on variational modal decomposition is characterized by comprising the following steps:
acquiring a biological radar voice to obtain an original radar voice signal with noise;
performing empirical mode decomposition on the original noisy radar voice signal to obtain intrinsic mode function characteristics, and determining the number of layers of variable mode decomposition according to the intrinsic mode function characteristics;
carrying out variational modal decomposition VMD on the original noisy radar voice according to the decomposition layer number to obtain each order of intrinsic modal decomposition;
solving Pearson coefficients for each order of eigenmode decomposition and original radar voice signals to determine a useful decomposition reconstruction mode;
and denoising the useful modes by using the improved threshold value, and reconstructing to obtain the enhanced radar voice.
2. The biological radar voice enhancement method based on variational modal decomposition according to claim 1, wherein: the step of acquiring the biological radar voice is to sample the biological radar voice by using an 8-channel Powerlab physiological recorder, and the sampling rate is 10kHz.
3. The biological radar voice enhancement method based on variational modal decomposition according to claim 1, wherein: the determining the number of layers of the variation modal decomposition according to the characteristic of the intrinsic modal function comprises the following steps:
given a voice signal as x (t), the empirical mode decomposition method adaptively decomposes an original signal into local oscillation mode functions through a screening process, and each local oscillation mode function IMF is a sub-frequency component of the original signal; analyzing the amplitude characteristics of each IMF from high frequency to low frequency, wherein when the amplitude value of the IMF is smaller than a set value, the mode is voice detail information, the rest modes are main information, and the number of the main information modes is the number of layers of variational mode decomposition;
from top to bottom, the IMF decomposed by the empirical mode decomposition method takes the IMF order with the first occurrence amplitude value smaller than the set value as the main information mode number, and is obtained by the following formula:
in the formula, IMF i For each order of the eigenmode function after empirical mode decomposition, i represents the order of the eigenmode function.
4. The biological radar voice enhancement method based on variational modal decomposition according to claim 1, wherein: performing variational modal decomposition VMD on the original noisy radar voice according to the decomposition layer number to obtain each order of eigenmode decomposition, including:
given an original radar voice signal as x (t), each mode u is obtained through Hilbert transformation k And constructing a variation constraint problem:
decomposing x (t) into K modes, where u k (K =1,2, …, K) is the K modes of the VMD method, w k (K =1,2, …, K) is the center frequency of each mode; introduce lagrange multiplier:
wherein, α is a penalty factor for controlling bandwidth, λ (t) is Lagrange multiplier, which represents convolution, δ (t) unit pulse function;
adopting a multiplication operator alternating direction method and Parseval Fourier equidistant transformation, converting each mode from a time domain to a frequency domain by a formula (2) as follows:
wherein, the center frequency of each mode is:
wherein w k n+1 Is the power spectral center of the k-th mode.
5. The biological radar voice enhancement method based on variational modal decomposition according to claim 3, wherein: solving the Pearson coefficient for each order of eigenmode decomposition and the original radar voice signal comprises the following steps:
wherein x is t And y t Representing two random variables.
6. The biological radar voice enhancement method based on variational modal decomposition according to claim 1, wherein: the improved threshold is represented by:
N j the length of the jth scale signal, σ, is the estimated noise of each modal signal; wherein the improved estimated noise of each mode can be calculated by the following formula:
where L is the estimated length of the initial silence period of the radar speech signal.
7. The biological radar voice enhancement method based on variational modal decomposition according to claim 1, wherein: denoising useful modalities with improved thresholds, comprising:
wherein m is a compensation factor.
8. The biological radar voice enhancement method based on variational modal decomposition according to claim 7, wherein: the value of the compensation factor m is 0.001.
9. The biological radar voice enhancement method based on variational modal decomposition according to claim 8, wherein: reconstructing to obtain enhanced radar voice, wherein the enhanced radar voice is obtained by reconstructing 94GHz radar voice subjected to denoising by an improved threshold strategy, and the specific reconstruction is calculated according to the following formula:
wherein, IMF is each mode of K of VMD decomposition, and a is useful mode number determined by calculating Pearson coefficient.
10. A biological radar voice enhancement system based on variational modal decomposition is characterized by comprising the following steps:
the acquisition module is used for acquiring the biological radar voice to obtain an original noisy radar voice signal;
the decomposition module is used for carrying out empirical mode decomposition on the original noisy radar voice signal to obtain intrinsic mode function characteristics, and determining the number of variable mode decomposition layers according to the intrinsic mode function characteristics; performing variable modal decomposition VMD on the original radar voice with noise according to the decomposition layer number to obtain each order of intrinsic modal decomposition; solving Pearson coefficients for each order of eigenmode decomposition and original radar voice signals to determine a useful decomposition reconstruction mode;
and the reconstruction module is used for denoising the useful mode by utilizing the improved threshold value and reconstructing to obtain the enhanced radar voice.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210993119.8A CN115376540A (en) | 2022-08-18 | 2022-08-18 | Biological radar voice enhancement method and system based on variational modal decomposition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210993119.8A CN115376540A (en) | 2022-08-18 | 2022-08-18 | Biological radar voice enhancement method and system based on variational modal decomposition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115376540A true CN115376540A (en) | 2022-11-22 |
Family
ID=84066015
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210993119.8A Pending CN115376540A (en) | 2022-08-18 | 2022-08-18 | Biological radar voice enhancement method and system based on variational modal decomposition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115376540A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116013240A (en) * | 2023-01-07 | 2023-04-25 | 广西大学 | Steel pipe concrete signal noise reduction method based on variational modal decomposition and digital filtering |
-
2022
- 2022-08-18 CN CN202210993119.8A patent/CN115376540A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116013240A (en) * | 2023-01-07 | 2023-04-25 | 广西大学 | Steel pipe concrete signal noise reduction method based on variational modal decomposition and digital filtering |
| CN116013240B (en) * | 2023-01-07 | 2023-10-31 | 广西大学 | Steel pipe concrete signal noise reduction method based on variational modal decomposition and digital filtering |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102697495B (en) | Second-generation wavelet electromyographic signal noise eliminating method based on ensemble empirical mode decomposition | |
| Alfaouri et al. | ECG signal denoising by wavelet transform thresholding | |
| CN109871733B (en) | Self-adaptive sea clutter signal denoising method | |
| CN101625869B (en) | Non-air conduction speech enhancement method based on wavelet-packet energy | |
| CN103984866A (en) | Signal denoising method based on local mean value decomposition | |
| CN110151175A (en) | Surface electromyogram signal noise-eliminating method based on CEEMD and improvement wavelet threshold | |
| CN108154081A (en) | Based on instantaneous frequency stability SWT logistics equipment vibration signal noise-reduction methods | |
| El Bouny et al. | ECG signal filtering based on CEEMDAN with hybrid interval thresholding and higher order statistics to select relevant modes | |
| CN106601265B (en) | A method of eliminating millimeter wave bioradar noise in voice | |
| Wang et al. | Ensemble EMD‐based signal denoising using modified interval thresholding | |
| CN113397496B (en) | Pulse wave acquisition method, system and storage medium based on signal-to-noise ratio improvement technology | |
| CN115700544A (en) | Chromatographic signal denoising method combining empirical mode decomposition and wavelet soft threshold | |
| CN111795931A (en) | Reconstruction extraction method for laser ultrasonic defect detection diffraction echo signal | |
| CN115840879A (en) | Additive manufacturing component ultrasonic detection echo signal processing method and system | |
| CN115376540A (en) | Biological radar voice enhancement method and system based on variational modal decomposition | |
| CN104280776B (en) | Self-adaption wavelet threshold solving method | |
| Ashraf et al. | Underwater ambient-noise removing GAN based on magnitude and phase spectra | |
| Li | Speech denoising based on improved discrete wavelet packet decomposition | |
| Wang et al. | Research on denoising algorithm for ECG signals | |
| El Bouny et al. | ECG signal denoising based on ensemble emd thresholding and higher order statistics | |
| Ou et al. | Frame-based time-scale filters for underwater acoustic noise reduction | |
| CN106782588A (en) | A kind of millimeter wave bioradar sound enhancement method based on sense of hearing wavelet packet | |
| Hussein et al. | Spectrogram enhancement by edge detection approach applied to bioacoustics calls classification | |
| Malleswari et al. | An improved denoising of electrocardiogram signals based on wavelet thresholding | |
| Malhotra et al. | Electrocardiogram signals denoising using improved variational mode decomposition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |