A method and system for processing audio signals of cardiopulmonary recording
By performing multi-layer processing of the audio signals of cardiopulmonary recordings, including noise reduction, digital-to-analog conversion, signal expansion and periodic analysis, the problems of low signal-to-noise ratio and serious distortion in the prior art are solved, and the cardiopulmonary characteristic information is significantly improved. accuracy.
Patent Information
- Application Number
- CN202410169761.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-02-06
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-02-06
AI Technical Summary
The audio signal processing of existing cardiopulmonary recordings has low signal-to-noise ratio and severe distortion, resulting in poor accuracy of acquired cardiopulmonary characteristic information.
The respiratory sound recorder is used to collect the respiratory source signal of the target user's predetermined time zone, perform global noise reduction and local noise reduction processing, digital-to-analog conversion processing and signal sampling, perform signal expansion to compensate for spectral aliasing distortion, and standard digital signals Periodic analysis was performed to extract the target digital signal and identify the cardiopulmonary characteristics alone.
Effectively filter out environmental noise, improve signal resolution and integrity, enhance the accuracy of cardiopulmonary characteristic information, and realize high-quality cardiopulmonary audio signal processing.
Smart Images

Figure CN118280380B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of signal processing, and in particular to a method and system for processing audio signals of cardiopulmonary recordings. Background Art
[0002] With the development of society and the aging of the population, in order to prevent cardiopulmonary diseases, it is necessary to record and monitor the heart and lungs. However, the audio signal processing of existing cardiopulmonary recordings is sensitive to environmental noise, and the cardiopulmonary audio signal is easily contaminated by various noises, and the signal-to-noise ratio is generally low. In addition, the audio signal processing process of cardiopulmonary recordings will also introduce varying degrees of distortion. These problems greatly reduce the quality of the cardiopulmonary audio signal, resulting in poor accuracy of the acquired cardiopulmonary feature information. Summary of the invention
[0003] The present application provides an audio signal processing method and system for cardiopulmonary recordings, aiming to solve the technical problems in the prior art of low signal-to-noise ratio and severe distortion of cardiopulmonary audio signal processing, resulting in poor accuracy of acquired cardiopulmonary characteristic information.
[0004] In view of the above problems, the present application provides an audio signal processing method and system for cardiopulmonary recording.
[0005] The first aspect disclosed in the present application provides an audio signal processing method for cardiopulmonary recording, the method comprising: based on a respiratory sound recorder, collecting a respiratory source signal of a target user in a predetermined time zone, wherein the respiratory source signal is a continuous time signal; transmitting the respiratory source signal to a signal preprocessing unit, performing global noise reduction processing and local noise reduction processing, and determining a preprocessed signal, the signal preprocessing unit comprising a first noise reduction unit and a second noise reduction unit, and the second noise reduction unit comprising a plurality of noise reduction branches; setting a predetermined phase interval, performing digital-to-analog conversion processing and signal sampling on the preprocessed signal, and determining a sampled digital signal, wherein the sampled digital signal is a discrete time signal; integrating the sampled digital signal, performing signal expansion, and determining a standard digital signal, wherein the signal expansion factor is consistent with the predetermined phase interval, and there is distortion compensation based on spectrum aliasing; performing periodic analysis on the standard digital signal, integrating periodic signal segments as the target digital signal; identifying the target digital signal, and determining a single column of the cardiopulmonary characteristics of the target user.
[0006] Another aspect disclosed in the present application provides an audio signal processing system for cardiopulmonary recording, the system comprising: a respiratory source signal module, used to collect the respiratory source signal of the target user in a predetermined time zone based on a respiratory sound recorder, wherein the respiratory source signal is a continuous time signal; a preprocessing signal module, used to transmit the respiratory source signal to a signal preprocessing unit, perform global noise reduction processing and local noise reduction processing, and determine the preprocessing signal, the signal preprocessing unit comprises a first noise reduction unit and a second noise reduction unit, and the second noise reduction unit comprises a plurality of noise reduction branches; a sampling digital signal module, used to set a predetermined phase interval, perform digital-to-analog conversion processing and signal sampling on the preprocessing signal, and determine the sampled digital signal, the sampled digital signal is a discrete time signal; a standard digital signal module, used to integrate the sampled digital signal, perform signal expansion, and determine the standard digital signal, wherein the signal expansion ratio is consistent with the predetermined phase interval, and there is distortion compensation based on spectrum aliasing; a target digital signal module, used to perform periodic analysis on the standard digital signal, integrate the periodic signal segments as the target digital signal; a cardiopulmonary feature single column module, used to identify the target digital signal and determine the cardiopulmonary feature single column of the target user.
[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0008] The respiratory source signal of the target user's predetermined time zone is used as the original signal for cardiopulmonary audio analysis, laying the foundation for subsequent processing; two-level noise reduction processing is performed, including global noise reduction and local noise reduction, to effectively filter out environmental noise and random noise, improve the distinguishability of the signal, and lay the foundation for retaining effective feature information; set a predetermined phase interval, perform digital-to-analog conversion processing and signal sampling, digitize the signal and obtain a standardized discrete signal, determine the sampled digital signal, and facilitate subsequent digital signal processing and analysis; expand the time domain length of the sampled digital signal, perform distortion compensation based on spectrum aliasing, and make up for the distortion introduced during the acquisition process , obtain a standard digital signal and enhance signal integrity; periodically analyze the standard digital signal, extract the target signal segment, remove the non-periodic components, obtain the target digital signal, and enhance the parameter representation of effective features; identify the target digital signal, generate cardiopulmonary state characteristic parameters, determine the cardiopulmonary characteristic single column of the target user, and realize the technical solution of high-quality cardiopulmonary audio signal processing, which solves the technical problems of low signal-to-noise ratio and serious distortion in cardiopulmonary audio signal processing in the prior art, resulting in poor accuracy of the acquired cardiopulmonary characteristic information, and achieves the technical effect of retaining effective signal information and improving the accuracy of cardiopulmonary characteristic information through multi-layer signal processing.
[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Figure 1 A flowchart of an audio signal processing method for cardiopulmonary recording is provided for an embodiment of the present application;
[0011] Figure 2 A flowchart of setting a predetermined phase interval in a method for processing an audio signal of a cardiopulmonary recording is provided for an embodiment of the present application;
[0012] Figure 3 A structural schematic diagram of an audio signal processing system for cardiopulmonary recording is provided for an embodiment of the present application.
[0013] Explanation of the reference numerals: respiratory source signal module 11 , preprocessing signal module 12 , sampling digital signal module 13 , standard digital signal module 14 , target digital signal module 15 , cardiopulmonary feature single column module 16 . DETAILED DESCRIPTION
[0014] The overall idea of the technical solution provided by this application is as follows:
[0015] The embodiment of the present application provides an audio signal processing method and system for cardiopulmonary recording, which retains effective signal information and accurately obtains the cardiopulmonary status information of the user through multi-layer signal processing. First, the respiratory source signal of the target user in the predetermined time zone is collected through a respiratory sound recorder to lay the source signal foundation for subsequent signal processing. Then, the respiratory source signal is globally and locally denoised by the preprocessing unit to effectively filter out the noise and obtain a preprocessed signal. Subsequently, a predetermined phase interval is set, and the preprocessed signal is subjected to digital-to-analog conversion processing and signal sampling to obtain a sampled digital signal, and the continuous signal is converted into a discrete signal suitable for digital signal analysis. Next, the spectrum aliasing technology is applied to expand the sampled digital signal to compensate for the distortion and obtain a standard digital signal. Thereafter, the standard digital signal is periodically analyzed to extract the periodic signal and construct a stable target digital signal. Finally, according to the standard target digital signal, the cardiopulmonary audio feature parameters are accurately obtained, and the cardiopulmonary feature single column of the target user is determined.
[0016] By performing multi-layer signal processing on the audio signals of cardiopulmonary recordings, the cardiopulmonary audio signal information is effectively retained and the accuracy of the cardiopulmonary characteristic information is significantly obtained.
[0017] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.
[0018] Embodiment 1
[0019] like Figure 1 As shown, the embodiment of the present application provides an audio signal processing method for cardiopulmonary recording, the method comprising:
[0020] Based on the respiratory sound recorder, a respiratory source signal of a target user in a predetermined time zone is collected, wherein the respiratory source signal is a continuous time signal;
[0021] In the embodiment of the present application, the device for collecting the respiratory source signal is a respiratory sound recorder, which can record the sound produced during human breathing. The operator pre-sets the working time of the respiratory sound recorder according to the physical condition of the target user, that is, the predetermined time zone.
[0022] The target user wears a respiratory sound recorder and turns on the device. When entering the predetermined time zone, the respiratory sound recorder automatically starts working, captures the target user's breathing sound in real time, and stores it in the form of a continuous analog signal in the time domain. After the predetermined time zone ends, the respiratory sound recorder stops working, completes the collection of the respiratory source signal, and obtains the respiratory source signal.
[0023] The obtained breathing source signal contains all breathing information of the target user in a predetermined time zone, with time as the horizontal axis and signal amplitude as the vertical axis, forming a continuous waveform, that is, a continuous time signal.
[0024] By setting a predetermined time zone, continuous respiratory source signals are collected to lay the signal source foundation for subsequent signal processing.
[0025] The respiratory source signal is transmitted to a signal preprocessing unit to perform global noise reduction processing and local noise reduction processing to determine a preprocessed signal, wherein the signal preprocessing unit includes a first noise reduction unit and a second noise reduction unit, and the second noise reduction unit includes a plurality of noise reduction branches;
[0026] Furthermore, this step also includes:
[0027] The configuration algorithm of the first noise reduction unit is universal, and the configuration algorithm of the second noise reduction unit is targeted;
[0028] Based on the first noise reduction unit, perform global noise reduction processing on the respiratory source signal to determine a primary processing signal;
[0029] Transferring the first processed signal flow to the second noise reduction unit, intercepting the noise signal segment and performing noise reduction branch matching to determine the target noise reduction branch;
[0030] Based on the target noise reduction branches, the noise signal segments are processed in parallel and globally replaced to generate the preprocessed signal, wherein the number of the target noise reduction branches is greater than or equal to zero.
[0031] In a feasible implementation, the signal preprocessing unit includes a first noise reduction unit and a second noise reduction unit to perform two-stage noise reduction processing to improve the noise suppression effect. Among them, the first noise reduction unit adopts a universal algorithm to analyze the time-frequency characteristics of the signal, identify the mainstream conventional noise in the signal, and process these conventional and high-frequency noises for preliminary signal noise reduction; the second noise reduction unit includes multiple noise reduction branches optimized for specific noise types, and the algorithm of each branch is more complex and sophisticated, such as specific noise reduction branches for ECG interference, bandwidth noise, pulse noise, etc.; multiple noise reduction branches work in parallel, identify and screen signal fragments, and then process them in a targeted manner. On the one hand, the source signal information can be retained to the maximum extent, and on the other hand, the residual noise can be eliminated in a targeted manner, achieving the dual goals of maximizing information extraction and minimizing noise suppression.
[0032] After acquiring the respiratory source signal, the respiratory source signal is transmitted to the signal preprocessing unit. The preprocessing unit transmits the collected respiratory source signal to the first noise reduction unit. The first noise reduction unit automatically identifies the noise components in the respiratory source signal by analyzing the frequency domain and time domain characteristics of the respiratory source signal, and filters the noise components. Under the processing of the first noise reduction unit, the conventional noise in the respiratory source signal is effectively suppressed, and a respiratory source signal that has undergone preliminary noise reduction is generated, which is defined as a primary processing signal. Then, the primary processing signal output by the first noise reduction unit is transmitted to the second noise reduction unit. The second noise reduction unit intercepts the block segments in the primary processing signal in units of time slices to determine whether there is residual noise. Then, the second noise reduction unit matches the identified noise signal segments one by one with the models contained in the multiple noise reduction branches in the unit, calculates the matching degree, and determines the best matching target noise reduction branch. Among them, the number of target noise reduction branches is greater than or equal to zero. When the number of target noise reduction branches is zero, it means that there is no obvious residual noise. At this time, the preprocessing signal is the primary processing signal. Then, the algorithm inside the target noise reduction branch is started, and only the corresponding noise signal segment is subjected to accurate noise reduction. After that, the noise signal segment processed by the target noise reduction branch replaces the original noise signal segment in the primary processing signal to obtain the preprocessed signal.
[0033] Furthermore, before performing the global noise reduction process and the local noise reduction process, the process further includes:
[0034] Identify the respiratory source signal, perform signal separation, and determine the respiratory signal and the noise signal, wherein the noise signal is caused by abnormal breathing;
[0035] The breathing signal and the noise signal are mapped.
[0036] In a preferred embodiment, before the first noise reduction unit and the second noise reduction unit perform global noise reduction processing and local noise reduction processing on the respiratory source signal, a new process of identifying and mapping the respiratory signal and the noise signal is added. This is because the respiratory source signal may be mixed with noise components caused by abnormal breathing, such as respiratory noise caused by secretions in the respiratory tract due to colds, pharyngitis, bronchitis, pneumonia, bronchial asthma, etc. Therefore, the original respiratory source signal obtained can be separated into a respiratory signal and a noise signal, and then separately enter the subsequent global noise reduction processing and local noise reduction processing, thereby improving the scope of application and refinement of the audio signal processing of the cardiopulmonary recording.
[0037] First, set up a respiratory signal model and a noise signal model, and compare the respiratory source signal with the respiratory signal model and the noise signal model at the same time. Use time-frequency domain analysis, signal recognition and other technologies to calculate the fit between the respiratory source signal segment and the two types of models. Then, based on the fit, divide the respiratory source signal into two paths, one for normal breathing sound signal, called the respiratory signal; the other for abnormal breathing noise signal, called the noise signal; the two signals are aligned and synchronized on the time axis. Next, establish a point-to-point mapping relationship between the respiratory signal and the noise signal on the time axis. For example, assuming that the total length of the respiratory source signal is T, divide it into n short blocks of the same length, and each block corresponds to a time slice. For each time slice, if there is a breathing signal, it is marked as O in the mapping structure; if there is a noise signal, it is marked as 1, forming a 01 sequence of length n to achieve the mapping of breathing signals and noise signals.
[0038] By identifying and mapping the respiratory signal and the noise signal, the distribution correspondence of the two signals on the time axis can be clearly understood, which helps to perform separate and synchronous processing operations on the two signals.
[0039] Setting a predetermined phase interval, performing digital-to-analog conversion processing and signal sampling on the preprocessed signal, and determining a sampled digital signal, wherein the sampled digital signal is a discrete time signal;
[0040] Further, such as Figure 2 As shown, setting the predetermined phase interval includes:
[0041] Identifying the preprocessed signal and determining the spectrum amplitude characteristics;
[0042] Based on the spectrum amplitude characteristics, fuzzy identification of effective signal points is performed, and phase distribution based on the identification result is determined;
[0043] The phase distribution is identified, and the minimum phase interval is taken as the predetermined phase interval.
[0044] In a feasible implementation, first, the preprocessed signal is Fourier transformed using algorithms such as fast Fourier transform to obtain its frequency domain information, including the amplitude and initial phase corresponding to each frequency point. Then, the amplitude of each frequency point is counted, the amplitude distribution in the entire frequency band is sorted, and the main energy concentration area, that is, the frequency interval with a large amplitude, is determined to obtain the spectrum amplitude characteristics. Subsequently, in the spectrum amplitude characteristics, since a large amplitude means that the signal energy contained in the frequency point is large, it is necessary for the reconstruction of the fidelity signal, so those frequency points with large amplitude values are selected as valid signal points. Then, in the phase information of these valid signal points, the phase size is extracted to obtain a phase value set containing only valid signal points, and the phase values in the phase value set are sorted by digital size, that is, the phase distribution of these valid signal points is obtained, which reflects the phase characteristics of the signal on these main frequency points. Thereafter, the numerical difference between adjacent phase values in the phase distribution is analyzed to obtain a series of phase differences, and the smallest phase difference is found after sorting these phase differences as the predetermined phase interval.
[0045] After that, the continuous preprocessed signal is input into the analog-to-digital converter for sampling. The sampling frequency of the analog-to-digital converter satisfies the Nyquist sampling theorem to ensure that no signal detailed information is lost, and the phase of the preprocessed signal is uniformly quantized with a predetermined phase interval as the quantization step, thereby obtaining a digital signal sampled in the time domain and the phase domain at the same time. However, phase domain quantization will introduce a sampling effect, and the sampled digital signal needs to be expanded by multiplying the predetermined phase interval, but there may be spectrum aliasing distortion, that is, the spectrum overlaps before and after. Based on this, the expansion factor is adjusted for the aliased part, or it is performed by interpolation, zero interpolation is performed, zero is inserted in the aliased area, and the spectrum trend is adjusted to determine the sampled digital signal.
[0046] Integrate the sampled digital signal, perform signal expansion, and determine a standard digital signal, wherein the signal expansion ratio is consistent with the predetermined phase interval, and there is distortion compensation based on spectrum aliasing;
[0047] Furthermore, this step includes:
[0048] Performing time sequence integration on the sampled digital signal to determine a sampled signal segment;
[0049] Based on the predetermined phase interval, determining the signal expansion factor, performing signal expansion on the sampled signal segment, and determining an expanded signal segment;
[0050] Identify the extended signal segment and determine whether there is signal aliasing distortion;
[0051] If not, the extended signal segment is used as the standard digital signal.
[0052] Furthermore, this step also includes:
[0053] If so, intercept the aliasing distortion signal segment, perform aliasing regularity and frequency identification, and determine whether the threshold standard is met;
[0054] If the conditions are met, a global compensation adjustment is performed on the extended signal segment; if the conditions are not met, a local compensation adjustment is performed on the aliasing distortion signal segment, wherein the compensation adjustment method includes an expansion ratio adjustment and a zero interpolation adjustment.
[0055] In a preferred embodiment, first, the sampled digital signal is analyzed to check whether there are intermittent gap segments. For each gap segment, the time domain characteristics of the adjacent signal blocks before and after it are recorded, such as sampling point timestamps, signal waveform correlation, etc. Then, these features are matched to determine the correct timing position of the gap segment under the complete time axis. For each gap segment with a determined position, the previous and next blocks are connected to form a seamless and uninterrupted complete sampled signal segment in the time domain. Subsequently, the determined predetermined phase interval is extracted, the duration of the sampled signal segment is quantitatively analyzed, the ratio of the duration of the sampled signal segment to the predetermined phase interval is calculated, and the ratio is rounded to an integer as the signal expansion factor of the sampled signal segment. Thereafter, the current signal segment is signal expanded at the determined signal expansion factor to obtain an extended signal segment. Next, the spectrum analysis is performed on the acquired extended signal segment to obtain its frequency component components, and the spectrum is detected to see whether there are two or more unconnected spectrum components in a frequency band. If so, it is determined that the spectrum is aliased or overlapped in the frequency band, that is, the overlap is caused by the reverse expansion of adjacent signal components. Then, the extended signal segment determined to be free of aliasing distortion is output as a standard digital signal.
[0056] If there is signal aliasing distortion in the extended signal segment, at this time, analyze the aliasing type and distortion regularity in this extended signal segment, and obtain the center frequency of the distortion and the affected frequency bandwidth as the aliasing feature. Then, compare the obtained aliasing feature with the threshold standard preset according to experience to determine whether the aliasing feature is within the range of the threshold standard. If the aliasing feature is within the range of the threshold standard, a global compensation adjustment is performed on the extended signal segment, that is, the signal expansion ratio of the sampled signal segment corresponding to the extended signal segment is fine-tuned, and the frequency domain analysis is performed again until the aliasing is eliminated. If the aliasing feature is not within the range of the threshold standard, a local correction is performed on the aliasing segment, that is, the aliasing distortion signal segment, that is, only the waveform points of the aliasing distortion signal segment are filled with zero values, and the spectrum overlap of the front and rear signals is separated and compensated. By adjusting the expansion ratio and zero interpolation adjustment for the extended signal segment according to the situation, the extended aliasing distortion is effectively eliminated to obtain a standard digital signal.
[0057] Performing periodic analysis on the standard digital signal and integrating periodic signal segments as the target digital signal;
[0058] Furthermore, the embodiment of the present application also includes:
[0059] Identifying regular signal segments and irregular signal segments of the standard digital signal;
[0060] Periodically integrate and identify regular signal segments, determine periodic signal segments, and identify the overall trend;
[0061] For irregular signal segments, the difference is measured in combination with the overall trend to identify irregular characteristic values;
[0062] The periodic signal segment and the irregular signal segment are integrated to determine the target digital signal.
[0063] In a feasible implementation, first, the standard digital signal is segmented in the time domain to obtain multiple signal segments corresponding to the time domain, and each signal segment is individually analyzed for periodicity to detect the existence of a periodic cycle pattern. If there is a stable periodic pattern, the signal segment is determined as a regular signal segment; if it does not have periodic characteristics, it is determined as an irregular segment. Regularity detection and classification are completed for all time domain segments to obtain regular signal segments and irregular signal segments. Subsequently, it is determined whether the periodic patterns between different regular signal segments are highly consistent. When the periodic patterns of multiple regular signal segments are similar, it is considered to reflect the same law. At this time, only one representative regular signal segment is retained as a periodic signal segment, and the same category identifier is assigned to the remaining similar segments, indicating that the overall trend is the same as that reflected by the periodic signal segment, thereby reducing display redundancy and highlighting common laws.
[0064] At the same time, for each irregular signal segment, analyze the difference between its overall trend and that of the periodic signal segment, measure the degree of difference, determine whether there is a specific sudden situation, and whether the mutation is random or deterministic. Then, for each irregular signal segment, determine the irregular characteristic value that describes its characteristics based on the above difference analysis. Thereafter, the periodic signal segment and the irregular signal segment with irregular characteristic values are reintegrated and arranged in the time axis order of the standard digital signal, and the irregular signal segment is sequentially embedded between the periodic signal segments to obtain a target digital signal that contains both common regular information and retains unique details.
[0065] The target digital signal is identified, and the cardiopulmonary characteristic list of the target user is determined.
[0066] In the embodiment of the present application, a matching analysis model is established for the target digital signal, and the analysis model includes a periodic feature extraction module and a state evaluation decision module. Among them, the periodic feature extraction module identifies the basic periodic components of the target digital signal, such as heart rate and breathing rate, through methods such as FFT, and outputs periodic parameters expressed in digital form; the state evaluation decision module compares the target digital signal with the normal mode benchmark to determine whether the difference part meets the preset abnormal mode, and if so, it is determined as abnormal breathing.
[0067] After acquiring the target digital signal, the target digital signal is input into the analysis model. The analysis model calls the cycle feature extraction module and the state evaluation decision module to identify and process the target digital signal, and obtains two types of output results. One is a series of digital cardiopulmonary cycle parameters, such as heart rate, respiratory rate, respiratory cycle, etc., reflecting the cardiopulmonary physiological cycle of the target user; the other is a normal or abnormal binary classification state result, reflecting the respiratory patency of the target user. Subsequently, the above model outputs are integrated to form a single column of the target user's cardiopulmonary features.
[0068] In summary, the audio signal processing method for cardiopulmonary recording provided by the embodiment of the present application has the following technical effects:
[0069] Based on the respiratory sound recorder, the respiratory source signal of the target user's predetermined time zone is collected to obtain the original cardiopulmonary audio data. The respiratory source signal is transmitted to the signal preprocessing unit for global and local noise reduction processing, and the preprocessed signal is determined to effectively reduce the noise component in the signal and improve the signal quality. The predetermined phase interval is set, and the preprocessed signal is subjected to digital-to-analog conversion processing and signal sampling to determine the sampled digital signal, and the continuous signal is converted into a digital signal to facilitate subsequent digital signal processing. The sampled digital signal is integrated, the signal is expanded, and the standard digital signal is determined to provide a basis for obtaining the target digital signal. The standard digital signal is periodically analyzed, and the periodic signal segment is integrated as the target digital signal to obtain a stable cardiopulmonary audio waveform. The target digital signal is identified, the cardiopulmonary feature single column of the target user is determined, the characteristic information of the cardiopulmonary audio is accurately obtained, and the processing quality of the cardiopulmonary audio signal is improved.
[0070] Embodiment 2
[0071] Based on the same inventive concept as the audio signal processing method for cardiopulmonary recording in the aforementioned embodiment, Figure 3 As shown, an embodiment of the present application provides an audio signal processing system for cardiopulmonary recording, the system comprising:
[0072] A respiratory source signal module 11 is used to collect a respiratory source signal of a target user in a predetermined time zone based on a respiratory sound recorder, wherein the respiratory source signal is a continuous time signal;
[0073] A preprocessing signal module 12 is used to transmit the respiratory source signal to a signal preprocessing unit, perform global noise reduction processing and local noise reduction processing, and determine a preprocessing signal, wherein the signal preprocessing unit includes a first noise reduction unit and a second noise reduction unit, and the second noise reduction unit includes a plurality of noise reduction branches;
[0074] The sampling digital signal module 13 is used to set a predetermined phase interval, perform digital-to-analog conversion and signal sampling on the pre-processed signal, and determine a sampled digital signal, wherein the sampled digital signal is a discrete time signal;
[0075] A standard digital signal module 14, used to integrate the sampled digital signal, perform signal expansion, and determine a standard digital signal, wherein the signal expansion ratio is consistent with the predetermined phase interval, and there is a distortion compensation based on spectrum aliasing;
[0076] The target digital signal module 15 is used to perform periodic analysis on the standard digital signal and integrate periodic signal segments as the target digital signal;
[0077] The cardiopulmonary characteristic single column module 16 is used to identify the target digital signal and determine the cardiopulmonary characteristic single column of the target user.
[0078] Furthermore, the preprocessing signal module 12 further includes the following execution steps:
[0079] Identify the respiratory source signal, perform signal separation, and determine the respiratory signal and the noise signal, wherein the noise signal is caused by abnormal breathing;
[0080] The breathing signal and the noise signal are mapped.
[0081] Furthermore, the preprocessing signal module 12 further includes the following execution steps:
[0082] The configuration algorithm of the first noise reduction unit is universal, and the configuration algorithm of the second noise reduction unit is targeted;
[0083] Based on the first noise reduction unit, perform global noise reduction processing on the respiratory source signal to determine a primary processing signal;
[0084] Transferring the first processed signal flow to the second noise reduction unit, intercepting the noise signal segment and performing noise reduction branch matching to determine the target noise reduction branch;
[0085] Based on the target noise reduction branches, the noise signal segments are processed in parallel and globally replaced to generate the preprocessed signal, wherein the number of the target noise reduction branches is greater than or equal to zero.
[0086] Furthermore, the sampling digital signal module 13 includes the following execution steps:
[0087] Identifying the preprocessed signal and determining the spectrum amplitude characteristics;
[0088] Based on the spectrum amplitude characteristics, fuzzy identification of effective signal points is performed, and phase distribution based on the identification result is determined;
[0089] The phase distribution is identified, and the minimum phase interval is taken as the predetermined phase interval.
[0090] Furthermore, the standard digital signal module 14 includes the following execution steps:
[0091] Performing time sequence integration on the sampled digital signal to determine a sampled signal segment;
[0092] Based on the predetermined phase interval, determining the signal expansion factor, performing signal expansion on the sampled signal segment, and determining an expanded signal segment;
[0093] Identify the extended signal segment and determine whether there is signal aliasing distortion;
[0094] If not, the extended signal segment is used as the standard digital signal.
[0095] Furthermore, the standard digital signal module 14 further includes the following execution steps:
[0096] If so, intercept the aliasing distortion signal segment, perform aliasing regularity and frequency identification, and determine whether the threshold standard is met;
[0097] If the conditions are met, a global compensation adjustment is performed on the extended signal segment; if the conditions are not met, a local compensation adjustment is performed on the aliasing distortion signal segment, wherein the compensation adjustment method includes an expansion ratio adjustment and a zero interpolation adjustment.
[0098] Furthermore, the target digital signal module 15 includes the following execution steps:
[0099] Identifying regular signal segments and irregular signal segments of the standard digital signal;
[0100] Periodically integrate and identify regular signal segments, determine periodic signal segments, and identify the overall trend;
[0101] For irregular signal segments, the difference is measured in combination with the overall trend to identify irregular characteristic values;
[0102] The periodic signal segment and the irregular signal segment are integrated to determine the target digital signal.
[0103] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.
[0104] Furthermore, the first or second mentioned above may not only represent an order relationship, but may also represent a specific concept, and / or refer to multiple elements that can be selected individually or in full. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.
Claims
1. A method for processing audio signals of cardiopulmonary recordings, characterized in that: The method comprises: Based on the respiratory sound recorder, a respiratory source signal of a target user in a predetermined time zone is collected, wherein the respiratory source signal is a continuous time signal; The respiratory source signal is transmitted to a signal preprocessing unit to perform global noise reduction processing and local noise reduction processing to determine a preprocessed signal, wherein the signal preprocessing unit includes a first noise reduction unit and a second noise reduction unit, and the second noise reduction unit includes a plurality of noise reduction branches; Setting a predetermined phase interval, performing digital-to-analog conversion processing and signal sampling on the preprocessed signal, and determining a sampled digital signal, wherein the sampled digital signal is a discrete time signal; Integrate the sampled digital signal, perform signal expansion, and determine a standard digital signal, wherein the signal expansion ratio is consistent with the predetermined phase interval, and there is distortion compensation based on spectrum aliasing; Performing periodic analysis on the standard digital signal and integrating periodic signal segments as the target digital signal; The target digital signal is identified, and the cardiopulmonary characteristic list of the target user is determined.
2. The method according to claim 1, characterized in that Before performing the global noise reduction processing and the local noise reduction processing, the method further includes: Identify the respiratory source signal, perform signal separation, and determine the respiratory signal and the noise signal, wherein the noise signal is caused by abnormal breathing; The breathing signal and the noise signal are mapped.
3. The method according to claim 2, characterized in that The configuration algorithm of the first noise reduction unit is universal, and the configuration algorithm of the second noise reduction unit is targeted. The method further includes: Based on the first noise reduction unit, perform global noise reduction processing on the respiratory source signal to determine a primary processing signal; Transferring the first processed signal flow to the second noise reduction unit, intercepting the noise signal segment and performing noise reduction branch matching to determine the target noise reduction branch; Based on the target noise reduction branches, the noise signal segments are processed in parallel and globally replaced to generate the preprocessed signal, wherein the number of the target noise reduction branches is greater than or equal to zero.
4. The method according to claim 1, characterized in that The method further comprises: Identifying the preprocessed signal and determining the spectrum amplitude characteristics; Based on the spectrum amplitude characteristics, fuzzy identification of effective signal points is performed, and phase distribution based on the identification result is determined; The phase distribution is identified, and the minimum phase interval is taken as the predetermined phase interval.
5. The method according to claim 1, characterized in that The method of integrating the sampled digital signal, performing signal expansion, and determining a standard digital signal further comprises: Performing time sequence integration on the sampled digital signal to determine a sampled signal segment; Based on the predetermined phase interval, determining the signal expansion factor, performing signal expansion on the sampled signal segment, and determining an expanded signal segment; Identify the extended signal segment and determine whether there is signal aliasing distortion; If not, the extended signal segment is used as the standard digital signal.
6. The method according to claim 5, characterized in that The method further comprises: If so, intercept the aliasing distortion signal segment, perform aliasing regularity and frequency identification, and determine whether the threshold standard is met; If the conditions are met, a global compensation adjustment is performed on the extended signal segment; if the conditions are not met, a local compensation adjustment is performed on the aliasing distortion signal segment, wherein the compensation adjustment method includes an expansion ratio adjustment and a zero interpolation adjustment.
7. The method according to claim 1, characterized in that Performing periodic analysis on the standard digital signal and integrating periodic signal segments as the target digital signal, the method further includes: Identifying regular signal segments and irregular signal segments of the standard digital signal; Periodically integrate and identify regular signal segments, determine periodic signal segments, and identify the overall trend; For irregular signal segments, the difference is measured in combination with the overall trend to identify irregular characteristic values; The periodic signal segment and the irregular signal segment are integrated to determine the target digital signal.
8. An audio signal processing system for cardiopulmonary recording, characterized in that: A method for processing an audio signal of a cardiopulmonary recording according to any one of claims 1 to 7, the system comprising: A respiratory source signal module, the respiratory source signal module is used to collect a respiratory source signal of a target user in a predetermined time zone based on a respiratory sound recorder, wherein the respiratory source signal is a continuous time signal; A preprocessing signal module, the preprocessing signal module is used to transmit the respiratory source signal to a signal preprocessing unit, perform global noise reduction processing and local noise reduction processing, and determine a preprocessing signal, the signal preprocessing unit includes a first noise reduction unit and a second noise reduction unit, and the second noise reduction unit includes a plurality of noise reduction branches; A sampling digital signal module, the sampling digital signal module is used to set a predetermined phase interval, perform digital-to-analog conversion processing and signal sampling on the pre-processed signal, and determine a sampled digital signal, the sampled digital signal being a discrete time signal; A standard digital signal module, the standard digital signal module is used to integrate the sampled digital signal, perform signal expansion, and determine a standard digital signal, wherein the signal expansion ratio is consistent with the predetermined phase interval, and there is a distortion compensation based on spectrum aliasing; A target digital signal module, the target digital signal module is used to perform periodic analysis on the standard digital signal and integrate periodic signal segments as a target digital signal; A cardiopulmonary feature single column module, wherein the cardiopulmonary feature single column module is used to identify the target digital signal and determine the cardiopulmonary feature single column of the target user.
