CN113869107A

CN113869107A - Signal denoising method, signal denoising device, electronic device and storage medium

Info

Publication number: CN113869107A
Application number: CN202110959698.XA
Authority: CN
Inventors: 陈子豪; 易昊翔
Original assignee: Hangzhou Enter Electronic Technology Co ltd
Current assignee: Hangzhou Enter Electronic Technology Co ltd
Priority date: 2021-08-20
Filing date: 2021-08-20
Publication date: 2021-12-31

Abstract

The application relates to a signal denoising method, a signal denoising device, an electronic device and a storage medium, wherein the signal denoising method is used for obtaining a target signal segment to be denoised, extracting characteristic parameters of the target signal segment, obtaining a noise category of the target signal segment according to the characteristic parameters, and denoising the target signal segment by adopting a preset denoising mode according to the noise category of the target signal segment, wherein the preset denoising mode comprises different denoising thresholds and strategies which are set aiming at different types of noise categories. The method and the device realize self-adaptive denoising based on the self characteristics of the target signal, reduce the noise content of the EEG signal, thereby improving the user experience and improving the accuracy of the EEG signal analysis result.

Description

Signal denoising method, signal denoising device, electronic device and storage medium

Technical Field

The present application relates to the field of signal processing, and in particular, to a signal denoising method, apparatus, electronic apparatus, and storage medium.

Background

EEG (Electroencephalogram) signals contain rich information related to brain states, can be used for analyzing attention, relaxation and other states of the brain, and the signal quality of the EEG signals has an important influence on the accuracy of analysis results in the analysis process of the EEG signals. The EEG signal can be acquired from the head point of the subject, so the eye movement and facial muscle movement of the subject will affect the signal quality of the EEG signal, and in order to improve the accuracy of the analysis of the EEG signal, the noise of the original EEG signal needs to be removed.

At present, the original EEG signal is denoised by performing single band pass filtering on the EEG signal, and then directly using the energy of the signal to determine whether the single channel EEG signal contains other noise, if not, the signal is directly used for calculation, and if there is interference, the signal is removed without participating in calculation. The method of directly abandoning the EEG signal with large noise seriously damages the integrity of signal data, causes the situation of no data for a long time or sometimes data in actual use, influences the user experience, and influences the accuracy of the signal analysis result.

Aiming at the problems that the user experience is poor and the accuracy of an EEG signal analysis result is influenced due to the fact that a single denoising mode for the EEG signal exists in the related technology, an effective solution is not provided at present.

Disclosure of Invention

The embodiment provides a signal denoising method, a signal denoising device, an electronic device and a storage medium, which are used for solving the problems that the user experience is poor and the accuracy of an EEG signal analysis result is influenced due to the fact that the denoising mode of an EEG signal is single in the related art.

In a first aspect, in this embodiment, a signal denoising method is provided, including:

acquiring a target signal segment to be denoised;

extracting characteristic parameters of the target signal segment;

obtaining the noise category of the target signal segment according to the characteristic parameters;

denoising the target signal segment by adopting a preset denoising mode according to the noise category of the target signal segment; the preset denoising mode comprises different denoising thresholds and strategies set for different types of noise categories.

In some embodiments, the extracting the feature parameters of the target signal segment includes:

filtering the target signal segment by using a band-pass filter to obtain a filtered signal segment of the target signal segment;

and calculating a first characteristic parameter indicating the noise content in the target signal segment, and calculating a second characteristic parameter indicating the noise content in the filtering signal segment.

In some embodiments, the extracting the feature parameters of the target signal segment further includes:

filtering the target signal segment by using a band-pass filter to obtain a filtering signal of the target signal segment;

and calculating the spectral characteristics of the filtering signal segment, and taking the filtering signal segment and the spectral characteristics as the characteristic parameters of the target signal segment.

In some embodiments, the obtaining the noise category of the target signal segment according to the characteristic parameter includes:

inputting the characteristic parameters into a completely trained decision tree model to obtain the signal quality grade corresponding to the target signal segment;

and determining the noise category of the target signal segment according to the signal quality grade.

In some embodiments, denoising the target signal segment by adopting a preset denoising manner according to the noise category of the target signal segment includes:

and under the condition that the signal quality grade reaches a denoising condition, selecting a preset denoising mode to denoise the target signal segment according to the noise category corresponding to the signal quality grade.

In some embodiments, the obtaining the noise category of the target signal segment according to the characteristic parameter further includes:

and inputting the characteristic parameters into a neural network with complete training to obtain the noise category corresponding to the target signal segment.

In some embodiments, the denoising the target signal segment by adopting a preset denoising manner according to the noise category of the target signal segment further includes:

performing wavelet decomposition on a filtering signal obtained after filtering the target signal segment to obtain a wavelet coefficient after the wavelet decomposition of the target signal segment;

and after the wavelet coefficient is processed by adopting a preset denoising threshold value and a preset denoising strategy, reconstructing the target signal segment according to the processed wavelet coefficient.

In some embodiments, the acquiring a target signal segment to be denoised includes:

acquiring an EEG original signal acquired by an electrode site, performing segment interception on the EEG original signal with a preset length, and taking a signal segment of the intercepted EEG original signal as a target signal segment.

In a second aspect, in this embodiment, there is provided a signal denoising apparatus, including: the device comprises an acquisition module, a feature extraction module, a category determination module and a denoising module, wherein:

the acquisition module is used for acquiring a target signal segment to be denoised;

the characteristic extraction module is used for extracting characteristic parameters of the target signal segment;

the category determining module is used for obtaining the noise category of the target signal segment according to the characteristic parameters;

the denoising module is used for denoising the target signal segment by adopting a preset denoising mode according to the noise category of the target signal segment; the preset denoising mode comprises different denoising thresholds and strategies set for different types of noise categories.

In a third aspect, an electronic device is provided in this embodiment, and includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the signal denoising method of the first aspect when executing the computer program.

In a fourth aspect, in the present embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the steps of the signal denoising method according to the first aspect.

The signal denoising method, the signal denoising device, the electronic device and the storage medium acquire a target signal segment to be denoised, extract characteristic parameters of the target signal segment, obtain a noise category of the target signal segment according to the characteristic parameters, and denoise the target signal segment by adopting a preset denoising mode according to the noise category of the target signal segment, wherein the preset denoising mode comprises different denoising thresholds and strategies set aiming at different types of noise categories. The method and the device realize self-adaptive denoising based on the self characteristics of the target signal, reduce the noise content of the EEG signal, thereby improving the user experience and improving the accuracy of the EEG signal analysis result.

The details of one or more embodiments of the application are set forth in the accompanying drawings and the description below to provide a more thorough understanding of the application.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

fig. 1 is a block diagram of a hardware structure of a terminal of a signal denoising method of the related art;

FIG. 2 is a flowchart of a signal denoising method according to the present embodiment;

fig. 3 is a block diagram of the signal denoising apparatus according to the present embodiment.

Detailed Description

For a clearer understanding of the objects, aspects and advantages of the present application, reference is made to the following description and accompanying drawings.

Unless defined otherwise, technical or scientific terms used herein shall have the same general meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The use of the terms "a" and "an" and "the" and similar referents in the context of this application do not denote a limitation of quantity, either in the singular or the plural. The terms "comprises," "comprising," "has," "having," and any variations thereof, as referred to in this application, are intended to cover non-exclusive inclusions; for example, a process, method, and system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or modules, but may include other steps or modules (elements) not listed or inherent to such process, method, article, or apparatus. Reference throughout this application to "connected," "coupled," and the like is not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. Reference to "a plurality" in this application means two or more. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. In general, the character "/" indicates a relationship in which the objects associated before and after are an "or". The terms "first," "second," "third," and the like in this application are used for distinguishing between similar items and not necessarily for describing a particular sequential or chronological order.

The method embodiments provided in the present embodiment may be executed in a terminal, a computer, or a similar computing device. For example, the method is executed on a terminal, and fig. 1 is a block diagram of a hardware structure of the terminal of the signal denoising method in this embodiment. As shown in fig. 1, the terminal may include one or more processors 102 (only one shown in fig. 1) and a memory 104 for storing data, wherein the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA. The terminal may also include a transmission device 106 for communication functions and an input-output device 108. It will be understood by those of ordinary skill in the art that the structure shown in fig. 1 is merely an illustration and is not intended to limit the structure of the terminal described above. For example, the terminal may also include more or fewer components than shown in FIG. 1, or have a different configuration than shown in FIG. 1.

The memory 104 may be used for storing a computer program, for example, a software program of an application software and a module, such as a computer program corresponding to the signal denoising method in the present embodiment, and the processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, so as to implement the method. The memory 104 may include high speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memory located remotely from the processor 102, which may be connected to the terminal over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission device 106 is used to receive or transmit data via a network. The network described above includes a wireless network provided by a communication provider of the terminal. In one example, the transmission device 106 includes a Network adapter (NIC) that can be connected to other Network devices through a base station to communicate with the internet. In one example, the transmission device 106 may be a Radio Frequency (RF) module, which is used to communicate with the internet in a wireless manner.

In this embodiment, a signal denoising method is provided, and fig. 2 is a flowchart of the signal denoising method of this embodiment, as shown in fig. 2, the flowchart includes the following steps:

step S210, obtaining a target signal segment to be denoised.

Specifically, the target signal segment may be an original signal segment within a preset time length. The raw signals may be EEG signals containing information about the state of the brain, but also other types of electrical signals, such as EMG (electromyography) signals or EOG (electro-oculography) signals. When the target signal segment is an EEG signal segment, the target signal segment may be obtained by inputting an EEG raw signal collected by the electrode site into a sliding window with a preset time length for signal interception to obtain a plurality of EEG signal segments of the EEG raw signal. The target signal segment is one of the EEG signal segments. In particular, the electrode sites may be electrode sites distributed across the forehead. The signal segment may be truncated for a length of 4 seconds.

Step S220, extracting characteristic parameters of the target signal segment.

Specifically, the characteristic parameter of the target signal segment may be a parameter indicating the noise content in the target signal, such as the signal-to-noise ratio and the drift slope of the target signal segment, or a typical frequency band energy ratio, a signal effective value, a signal artifact maximum amplitude value, and the like obtained according to the filtered signal after filtering the target signal segment. Or, the characteristic parameter may be a frequency spectrum of a target signal segment obtained by filtering the target signal and calculating the filtered signal according to the filtered signal corresponding to the target signal, and then the frequency spectrum of the target signal and the filtered signal are used as the characteristic parameter of the target signal segment. In addition, the characteristic parameters of the target signal segment can also select other parameters for representing the noise content according to the category of the target signal segment. Additionally, before extracting the characteristic parameters of the target signal segment, main information of the target signal segment is retained for filtering drift and power frequency noise, and the target signal segment can be subjected to band-pass filtering. The passband frequency of the selected bandpass filter may be 2Hz to 45 Hz. It is understood that the passband frequency of the bandpass filter may also be selected from other ranges according to the signal category of the processed target signal segment and the actual application scenario, and is not limited herein.

And step S230, obtaining the noise category of the target signal segment according to the characteristic parameters.

The noise type of the target signal segment may be non-periodic noise such as polarization noise or muscle electricity, or noise such as blinking, eye movement, frown, and tooth biting. The characteristic parameters can be input into a preset classification model for processing, and the noise category corresponding to the target signal segment is obtained. Specifically, the preset classification model may be a completely pre-trained decision tree model, or may be another learning model, such as a completely pre-trained neural network.

Before the feature parameters are processed by using a preset classification model, the classification model can be trained by using a training set. The training set may be a data set that includes noise class labels. For example, in the process of training the decision tree model, the decision tree model may be trained by using a data set that includes the feature parameters of the training signal and is labeled with the noise category, so as to obtain a completely trained decision tree model. The characteristic parameters used for training and subsequent classification of the decision tree model may specifically, but not limited to, signal-to-noise ratio and drift slope extracted from a target signal segment, and typical frequency band energy ratio, signal effective value, and signal artifact maximum amplitude extracted from a filtered signal obtained by band-pass filtering the target signal segment. Additionally, in the process of training the decision tree model, labels of signal quality grades can be labeled for the characteristic parameters, and the signal quality grades can be divided based on the content and the category of noise according to actual application scenarios. Preferably, the signal quality levels can be divided into no-load, very bad, medium, normal and very good 5 levels. The device for acquiring the target signal segment is indicated by no load, the device is not worn by the detected object, the range indicates that an effective target signal segment is not acquired, the range indicates that the content of non-periodic noise such as polarization noise and muscle electricity in the target signal segment is higher than that of other noise, the target signal segment is normally indicated to have artifact noise such as blink, and the target signal segment is well indicated to have no obvious noise. And processing the characteristic parameters of the target signal segment by using a completely trained decision tree model to obtain the signal quality grade corresponding to the target signal segment. And determining the noise class corresponding to the target signal segment according to the signal quality grade.

Additionally, frequency spectrum calculation may be performed on a filtered signal segment obtained by band-pass filtering a target signal segment, and the frequency spectrum of the target signal segment and the filtered signal segment are input as characteristic parameters into a neural network trained completely in advance by using a training set labeled with a noise category, so as to obtain a noise category corresponding to the target signal segment.

Step S240, denoising the target signal segment by adopting a preset denoising mode according to the noise category of the target signal segment; the preset denoising mode comprises different denoising thresholds and strategies set for different types of noise categories.

The denoising threshold and strategy may be a threshold and strategy for processing a wavelet coefficient obtained after filtering and wavelet decomposing a target signal segment. In addition, the applicable denoising methods are different for different noise types. Therefore, the correspondence between the noise category and the denoising method can be predetermined, and after the noise category of the target signal segment is obtained based on the step S230, the denoising method corresponding to the noise category can be adopted for processing based on the correspondence. For example, for muscle electrical noise and polarization noise with higher content relative to other noise classes, a relatively low threshold may be set for the cd3 wavelet coefficients in the wavelet coefficients to remove large non-periodic noise in the target signal segment. And for signal segments containing artifacts such as winks, the signal segments can be removed by using a big-Massart adaptive strategy. In addition, for the case where a valid target signal segment is not detected or a target signal segment is not acquired, the 0 level may be directly output. For target signal segments without significant noise, the wavelet coefficients may not be processed. Additionally, other wavelet thresholds and strategies may be selected for the noise category, such as hard or soft threshold functions, etc.

Compared with the conventional single denoising method directly based on the energy of the filtered signal, the method has the advantages that the accuracy of the analysis result of the target signal segment can be improved and the damage to the integrity of the denoised signal is avoided by extracting the characteristic parameters, calculating the noise type corresponding to the target signal segment based on the characteristic parameters and denoising the noise type by adopting the corresponding denoising method aiming at the noise type.

In the above steps S210 to S240, a target signal segment to be denoised is obtained, a characteristic parameter of the target signal segment is extracted, a noise category of the target signal segment is obtained according to the characteristic parameter, a preset denoising manner is adopted to denoise the target signal segment according to the noise category of the target signal segment, and the preset denoising manner includes different denoising thresholds and strategies set for different types of noise categories. The method and the device realize self-adaptive denoising based on the self characteristics of the target signal, reduce the noise content of the EEG signal, thereby improving the user experience and improving the accuracy of the EEG signal analysis result.

Further, in an embodiment, based on the step S220, the feature parameters include a first feature parameter and a second feature parameter, and the extracting the feature parameters of the target signal segment specifically includes the following steps:

step S221, filtering the target signal segment by using a band-pass filter to obtain a filtered signal segment of the target signal segment.

Specifically, the pass band frequency of the band pass filter may be determined according to an actual application scenario. Preferably, in case of processing the EEG raw signal segments, the pass band frequency of the band pass filter may be set to 2Hz to 45 Hz. After the target signal segment passes through the band-pass filter, drift and power frequency noise in the target signal can be taken out, and main information of the target signal is reserved.

Step S222, calculating a first characteristic parameter indicating the noise content in the target signal segment, and calculating a second characteristic parameter indicating the noise content in the filtered signal segment.

Specifically, the first characteristic parameter may be a signal-to-noise ratio and a drift slope directly calculated from the target signal segment. The second characteristic parameter may be a typical frequency band energy ratio, a signal effective value and a signal artifact maximum amplitude value calculated according to the filtered signal segment. The content of different types of noise in the target signal can be determined through the parameters, and then the signal quality of the target signal segment can be graded.

Additionally, in an embodiment, based on the step S220, extracting the feature parameters of the target signal segment, further includes the following steps:

step S223, filtering the target signal segment by using the band-pass filter to obtain a filtered signal of the target signal segment.

Step S224, calculating the spectral characteristics of the filtered signal segment, and using the filtered signal segment and the spectral characteristics as the characteristic parameters of the target signal segment.

Additionally, in an embodiment, based on the step S230, the obtaining the noise category of the target signal segment according to the characteristic parameter specifically includes the following steps:

step S231, inputting the characteristic parameters into the completely trained decision tree model to obtain the signal quality level corresponding to the target signal segment.

Specifically, the first characteristic parameter and the second characteristic parameter extracted in the above steps S221 and S222 may be input into the decision tree model together to obtain the signal quality level corresponding to the target signal segment.

Step S232, determining the noise category of the target signal segment according to the signal quality grade.

Different signal quality levels can represent the content of different types of noise in the target signal, so that the correspondence between the signal quality level and the noise type can be preset, and after the signal quality level corresponding to the target signal segment is obtained according to the step S231, the noise type in the target signal segment can be determined based on the signal quality level.

Further, in an embodiment, based on the step S240, a preset denoising method is adopted to denoise the target signal segment according to the noise category of the target signal segment, and the method specifically includes the following steps:

and step S241, under the condition that the signal quality level reaches the denoising condition, selecting a preset denoising mode to denoise the target signal segment according to the noise category corresponding to the signal quality level.

The denoising condition may be a predetermined signal quality level at which denoising processing may be performed. For example, when the signal quality level including polarization noise and muscle electrical noise is medium, or when the signal quality level including artifacts such as blinking is normal, a preset denoising method is selected for denoising according to the noise category corresponding to the signal quality level. The denoising mode can be specifically a denoising threshold and a strategy for processing a wavelet coefficient after wavelet decomposition is performed on a target signal.

Additionally, in an embodiment, based on the step S230, obtaining the noise category of the target signal segment according to the characteristic parameter, further includes the following steps:

step S233, the characteristic parameters are input into the neural network with complete training, and the noise category corresponding to the target signal segment is obtained.

Additionally, in an embodiment, a preset denoising method is adopted to denoise the target signal segment according to the noise category of the target signal segment, and the method specifically includes the following steps:

step S242, performing wavelet decomposition on the filtered signal obtained by filtering the target signal segment to obtain a wavelet coefficient after wavelet decomposition of the target signal segment.

Specifically, the wavelet decomposition may be performed on the filtered signal obtained by band-pass filtering the target signal segment with the db4 wavelet basis to obtain the wavelet coefficient of the target signal segment.

And step S243, processing the wavelet coefficient by adopting a preset denoising threshold value and a preset denoising strategy, and reconstructing the target signal segment according to the processed wavelet coefficient.

Additionally, in an embodiment, based on the step S210, acquiring a target signal segment to be denoised specifically includes the following steps:

step S211, obtaining an EEG original signal collected by an electrode site, performing segment interception on the EEG original signal with a preset length, and taking a signal segment of the intercepted EEG original signal as a target signal segment.

In the above steps S210 to S240, the band-pass filter is used to filter the target signal segment to obtain a filtered signal segment of the target signal segment, and a first characteristic parameter indicating the noise content in the target signal segment and a second characteristic parameter indicating the noise content in the filtered signal segment are calculated, so as to improve the accuracy of determining the noise category of the target signal segment; calculating the frequency spectrum characteristics of the filtering signal segment, and taking the filtering signal segment and the frequency spectrum characteristics as the characteristic parameters of the target signal segment, thereby improving the accuracy of judging the noise category of the target signal segment; inputting the characteristic parameters into a decision tree model which is trained completely to obtain a signal quality grade corresponding to a target signal segment, determining the noise category of the target signal segment according to the signal quality grade, and selecting a preset denoising mode to denoise the target signal according to the noise category corresponding to the signal quality grade under the condition that the signal quality grade reaches a denoising condition, so that the signal denoising accuracy can be improved; the wavelet decomposition is carried out on the filtering signal obtained after the target signal segment is filtered to obtain the wavelet coefficient of the target signal after the wavelet decomposition, the preset denoising threshold value and the strategy are adopted to process the wavelet coefficient, and the target signal segment is reconstructed according to the processed wavelet coefficient, so that the self-adaptive denoising based on the self characteristic of the target signal is realized, the noise content of the EEG signal is reduced, and the accuracy of the EEG signal analysis result is improved.

In the present embodiment, a signal denoising apparatus 30 is also provided. As shown in fig. 3, the signal denoising apparatus 30 includes: an acquisition module 32, a feature extraction module 34, a category determination module 36, and a denoising module 38, wherein:

an obtaining module 32, configured to obtain a target signal segment to be denoised;

a feature extraction module 34, configured to extract feature parameters of the target signal segment;

a category determining module 36, configured to obtain a noise category of the target signal segment according to the characteristic parameter;

the denoising module 38 is configured to denoise the target signal segment in a preset denoising manner according to the noise category of the target signal segment; the preset denoising mode comprises different denoising thresholds and strategies set for different types of noise categories.

The signal denoising device 30 obtains a target signal segment to be denoised, extracts characteristic parameters of the target signal segment, obtains a noise category of the target signal segment according to the characteristic parameters, and denoises the target signal segment by adopting a preset denoising manner according to the noise category of the target signal segment, wherein the preset denoising manner includes different denoising thresholds and strategies set for different types of noise categories. The method and the device realize self-adaptive denoising based on the self characteristics of the target signal segment, reduce the noise content of the EEG signal, thereby improving the user experience and improving the accuracy of the EEG signal analysis result.

The above modules may be functional modules or program modules, and may be implemented by software or hardware. For a module implemented by hardware, the modules may be located in the same processor; or the modules can be respectively positioned in different processors in any combination.

There is also provided in this embodiment an electronic device comprising a memory having a computer program stored therein and a processor arranged to run the computer program to perform the steps of any of the above method embodiments.

Optionally, the electronic apparatus may further include a transmission device and an input/output device, wherein the transmission device is connected to the processor, and the input/output device is connected to the processor.

Optionally, in this embodiment, the processor may be configured to execute the following steps by a computer program:

acquiring a target signal segment to be denoised;

extracting characteristic parameters of the target signal segment;

In one embodiment, the processor executes the computer program to further implement the following steps:

a first characteristic parameter indicative of the noise content in the target signal segment is calculated, and a second characteristic parameter indicative of the noise content in the filtered signal segment is calculated.

and under the condition that the signal quality grade reaches the denoising condition, selecting a preset denoising mode to denoise the target signal segment according to the noise category corresponding to the signal quality grade.

and after the wavelet coefficient is processed by adopting a preset denoising threshold value and a preset denoising strategy, reconstructing a target signal segment according to the processed wavelet coefficient.

obtaining an EEG original signal collected by an electrode site, carrying out fragment interception on the EEG original signal with a preset length, and taking a signal fragment of the intercepted EEG original signal as a target signal fragment.

In addition, in combination with the signal denoising method provided in the foregoing embodiment, a storage medium may also be provided in this embodiment to implement the method. The storage medium having stored thereon a computer program; the computer program, when executed by a processor, implements any of the signal denoising methods in the above embodiments.

It should be noted that, for specific examples in this embodiment, reference may be made to the examples described in the foregoing embodiments and optional implementations, and details are not described again in this embodiment.

It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to be limiting. All other embodiments, which can be derived by a person skilled in the art from the examples provided herein without any inventive step, shall fall within the scope of protection of the present application.

It is obvious that the drawings are only examples or embodiments of the present application, and it is obvious to those skilled in the art that the present application can be applied to other similar cases according to the drawings without creative efforts. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.

The term "embodiment" is used herein to mean that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the present application. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is to be expressly or implicitly understood by one of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the patent protection. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for denoising a signal, comprising:

acquiring a target signal segment to be denoised;

extracting characteristic parameters of the target signal segment;

2. The signal denoising method according to claim 1, wherein the feature parameters include a first feature parameter and a second feature parameter, and the extracting the feature parameters of the target signal segment includes:

3. The signal denoising method according to claim 1, wherein the extracting feature parameters of the target signal segment further comprises:

4. The signal denoising method of claim 1, wherein the obtaining a noise class of the target signal segment according to the characteristic parameter comprises:

5. The signal denoising method according to claim 4, wherein denoising the target signal segment by adopting a preset denoising manner according to the noise category of the target signal segment comprises:

6. The signal denoising method of claim 1, wherein the obtaining a noise class of the target signal segment according to the characteristic parameter further comprises:

7. The signal denoising method of claim 1, wherein the denoising the target signal segment by a preset denoising method according to the noise category of the target signal segment, further comprises:

8. The signal denoising method according to any one of claims 1 to 7, wherein the obtaining a target signal segment to be denoised comprises:

9. A signal denoising apparatus, comprising: the device comprises an acquisition module, a feature extraction module, a category determination module and a denoising module, wherein:

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the computer program to perform the signal denoising method according to any one of claims 1 to 8.

11. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the signal denoising method according to any one of claims 1 to 8.