CN110772279A

CN110772279A - Lung sound signal acquisition device and analysis method

Info

Publication number: CN110772279A
Application number: CN201911190292.9A
Authority: CN
Inventors: 韦海成; 冯海青; 王生营; 何艳茹; 肖明霞
Original assignee: North Minzu University
Current assignee: North Minzu University
Priority date: 2019-11-28
Filing date: 2019-11-28
Publication date: 2020-02-11

Abstract

The invention relates to a lung sound signal acquisition device and an analysis method, which comprise a lung sound sensor, an amplifier, an A/D acquisition card and an acquisition device carried by an upper computer, wherein the acquisition device is electrically connected in sequence, and lung sound signals acquired by the device are uploaded to a cloud end to be analyzed: receiving an original lung sound signal; decomposing the original lung sound signals, and extracting a plurality of intrinsic mode functions IMF from the original lung sound signals; after obtaining the intrinsic mode function IMF, selecting the intrinsic mode function IMF closest to the original lung sound signal, and performing Hilbert-Huang transform based on the intrinsic mode function to obtain a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of the original lung sound signal; and analyzing the instantaneous frequency extreme value and the instantaneous frequency amplitude in the Hilbert marginal spectrogram to judge the lung condition of the tested person.

Description

Lung sound signal acquisition device and analysis method

Technical Field

The invention relates to the technical field of biomedical signal analysis, in particular to a lung sound signal acquisition device and an analysis method.

Background

The lung is the respiratory organ of the human body and also an important hematopoietic organ of the human body. The lung sound is a sound signal generated when the lung exchanges gas with the outside, and reflects the health degree of the lung of a human body by reflecting the condition of the respiratory tract of the human body. The lung sound signals are subjected to feature extraction and analysis to judge the types of the diseases, and the method is a hot research field at present.

Common respiratory diseases can be roughly classified into: asthma, chronic obstructive pulmonary disease, chronic bronchitis, emphysema, lung cancer, cystic fibrosis, pneumonia, pleural effusion and the like. In a hospital, a doctor usually adopts a traditional stethoscope to listen to lung sounds as a basis for judging lung diseases, but the doctor is limited by factors such as the hearing condition of the doctor and medical experience, and the judgment result has objectivity. For this reason, it is desirable to introduce a digital sound analysis algorithm to perform the lung sound identification. Due to the non-linear and non-stationary characteristics of the lung sounds, the results are not ideal by using the traditional time domain analysis or frequency domain analysis method.

Disclosure of Invention

The invention aims to overcome the defects in the prior art and provides a lung sound signal acquisition device and an analysis method.

In order to achieve the above object, the embodiments of the present invention provide the following technical solutions:

a lung sound signal acquisition apparatus comprising:

the lung sound sensor is used for collecting lung sound signals;

the amplifier is used for amplifying the lung sound signals collected by the lung sound sensor;

the A/D acquisition card is used for converting the amplified lung sound signals into lung sound signals which can be identified by an upper computer;

and the upper computer is used for receiving and analyzing the lung sound signals converted by the A/D acquisition card.

The lung sound sensor is used for collecting lung sound signals of a tested person, after the lung sound signals are filtered and amplified by the amplifier, the USB is connected with the A/D collecting card to convert the lung sound signals into lung sound signals which can be identified by the upper computer, and the upper computer temporarily stores the lung sound signals and uploads the lung sound signals to the cloud for storage and analysis.

Further, in order to better implement the present invention, the lung sound sensor is of the type HKY-06F.

The lung sound signal adopts a novel polymer material micro sound sensing element to collect heart pulsation signals and other body surface artery pulsation signals, and is processed by a highly integrated signal processing circuit to output low-impedance audio signals and increase the function of resisting environmental noise.

A lung sound signal analysis method specifically comprises the following steps:

step S1: receiving an original lung sound signal;

step S2: decomposing the original lung sound signals, and extracting a plurality of intrinsic mode functions IMF from the original lung sound signals;

step S3: after obtaining the intrinsic mode function IMF, selecting the intrinsic mode function IMF closest to the original lung sound signal, and performing Hilbert-Huang transform based on the intrinsic mode function to obtain a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of the original lung sound signal;

among a plurality of research methods, the Hilbert-Huang transform method has the characteristic of time-frequency joint analysis, can simultaneously analyze time domain and frequency domain information of signals, can analyze the signals in multiple layers and process specific component information, and therefore has better effect and development prospect in the field of lung sound signal identification.

Step S4: and analyzing the instantaneous frequency extreme value and the instantaneous frequency amplitude in the Hilbert marginal spectrogram to judge the lung condition of the tested person.

Performing Hilbert-Huang transformation on the lung sound signal to eliminate various clutter signals with different frequencies, which are superposed on the lung sound signal under the influence of internal organs; and the lung sound intrinsic signal is provided, the hilbert time frequency spectrum and the marginal spectrum are utilized to analyze the intrinsic lung sound signal, and the change of the frequency and the amplitude of the lung sound is used as a judgment basis for detecting respiratory diseases.

Furthermore, in order to better implement the present invention, the step of performing Hilbert yellow transform based on the eigenmode function to obtain a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of the original lung sound signal includes:

step S3-1: real signal p for the eigenmode function IMF closest to the original lung sound signal _i(t) performing Hilbert-Huang transform:

step S3-2: for real signal p transformed by Hilbert-Huang _i(t) constructing its complex function:

z _i(t)＝p _i(t)+jH[p _i(t)]

will complex function z _i(t) is expressed in exponential form:

wherein the amplitude a _i(t) is:

phase phi _i(t) is:

to phase phi _i(t) differentiating to obtain p _i(t) instantaneous frequency f _i(t)：

Step S3-3: for real signal p _i(t) performing Hilbert time-frequency spectrum calculation:

is recorded as:

step S3-4: integration over the time axis of H (w, t) yields the Hilbert marginal spectrum:

step S3-5: and drawing a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of the original lung sound signal according to the Hilbert time-frequency spectrum and the Hilbert marginal spectrum.

Compared with the prior art, the invention has the beneficial effects that:

(1) the lung sound acquisition device is constructed by the lung sound sensor, the amplifier, the A/D acquisition card and the upper computer, is simple to operate and small in size, is convenient to acquire lung sound signals, and has no side effect on a testee; in addition, the full electronic data acquisition mode eliminates the inaccurate influence of artificial lung sound monitoring, and the acquired data is stored in the cloud, so that the data management and extraction are facilitated;

(2) the invention carries out Hilbert-Huang transform on the collected lung sound signals, eliminates the influence of clutter signals with different frequencies, which are superposed on the lung sound signals under the influence of internal organs and external environment, on the lung sound, and improves the accuracy of respiratory disease detection for the auxiliary diagnosis of patients with respiratory diseases.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.

FIG. 1 is a block diagram of an acquisition device according to the present invention;

FIG. 2 is a flow chart of an analysis method of the present invention;

FIG. 3(a) is a waveform diagram of lung sound signal of normal healthy person;

FIG. 3(b) is a waveform diagram of lung sound signals of a patient with respiratory diseases;

FIG. 4(a) is a waveform diagram of an intrinsic mode function IMF of a lung sound signal of a normal healthy person;

FIG. 4(b) is an IMF waveform of the intrinsic mode function of lung sound signals of a patient with respiratory tract diseases;

FIG. 5(a) is a Hilbert time-frequency spectrogram of a lung sound signal of a normal healthy person;

FIG. 5(b) is a Hilbert time-frequency spectrogram of a lung sound signal of a respiratory disease patient;

FIG. 6(a) is a Hilbert marginal spectrum of a lung sound signal of a normal healthy person;

fig. 6(b) is a Hilbert marginal spectrum of a lung sound signal of a respiratory disease patient;

FIG. 7 is a graphical representation of Hilbert marginal spectral mean waveforms for normal healthy persons.

Detailed Description

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. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.

Example 1:

the invention is realized by the following technical scheme, as shown in fig. 1, the lung sound signal acquisition device comprises a lung sound sensor, an amplifier, an A/D acquisition card and an upper computer which are sequentially and electrically connected, wherein the lung sound sensor is used for acquiring lung sound signals of a tested person, the lung sound signals are hereinafter referred to as lung sound signals, and the lung sound signals are filtered and amplified by the amplifier.

The lung sound sensor comprises a highly integrated signal processing circuit, has a certain environmental noise resistance function, and has the advantages of high reliability, high sensitivity, small size, convenience in use and the like. The A/D acquisition card is 16 bits, converts the original lung sound signal into a digital signal which can be identified by an upper computer, and then uses audio recording software in the upper computer to realize the waveform and data storage of the lung sound signal.

The model of the lung sound sensor is HKY-06F, and the heart pulse and other body surface artery pulse signals are acquired by adopting a novel high polymer material micro-sound sensing element and are processed by a highly integrated signal processing circuit to output low-impedance audio signals. The sensor is added with the function of resisting environmental noise on the basis of HKY-06B.

This device directly obtains effectual human lung sound signal to real signal waveform reflects the characteristic of wave form in the host computer, thereby judges the person's of being surveyed lung condition, has got rid of the artificial inaccurate influence of monitoring lung sound.

Based on the lung sound signal acquisition device, the lung sound signal acquired is analyzed, and a lung sound signal analysis method is provided, as shown in fig. 2, the method specifically includes the following steps:

step S100: the original lung sound signal is received.

The lung sound sensor collects the lung sound signal of the tested person, and the lung sound signal is filtered and amplified by the amplifier and then converted by the external A/D acquisition card.

Step S200: and decomposing the original lung sound signals, and extracting a plurality of intrinsic mode functions IMF from the original lung sound signals.

The acquired original lung sound signals are decomposed into a plurality of intrinsic mode functions IMF as shown in fig. 4, wherein fig. 4(a) is a waveform diagram of normal healthy people, and fig. 4(b) is a waveform diagram of respiratory disease patients.

Step S300: and after obtaining the intrinsic mode functions IMF, selecting one intrinsic mode function IMF closest to the original lung sound signal, and performing Hilbert-Huang transform based on the intrinsic mode function to obtain a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of the original lung sound signal.

After the original lung sound signal is decomposed into a plurality of intrinsic mode functions IMFs in step S200, it can be determined according to the frequency that, in the intrinsic mode functions shown in fig. 4(a) or fig. 4(b), the frequencies of IMFs 1 to IMF5 are too high, which are background noise or introduced noise, the frequencies of IFMs 7 to IMF10 are too low, which belong to noise of other organs of the body, and the waveform characteristics of IMF6 conform to the waveform characteristics of the lung sound signal, so that IMF6 is selected as the closest one of the intrinsic mode functions to the original lung sound signal.

The method comprises the following steps of performing Hilbert-yellow transform based on the intrinsic mode function to obtain a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of original lung sound signals, wherein the Hilbert time-frequency spectrogram and the Hilbert marginal spectrogram of the original lung sound signals comprise the following steps:

step S301: real signal p to the eigenmode function IMF6 closest to the original lung sound signal _i(t) performing a Hilbert-Huang transform:

defining t as an interval (- ∞ < t < + ∞);

step S302: for real signal p transformed by Hilbert-Huang _i(t) constructing its complex function z _i(t) and calculating the phase derivative thereof, wherein the constructed complex function is as follows:

z _i(t)＝p _i(t)+jH[p _i(t)]

will complex function z _i(t) is expressed in exponential form:

wherein the amplitude a _i(t) is:

phase phi _i(t) is:

further, for the phase phi _i(t) differentiating to obtain p _i(t) instantaneous frequency f _i(t)：

It can be seen that f _i(t) is a single valued function of time t, i.e. a time corresponds to a frequency and the sequence for Hilbert yellow transform must be a single component signal, thus again proving IMF6 useful.

Step S303: for real signal p _i(t) performing Hilbert time-frequency spectrum calculation:

is recorded as:

h (w, t) accurately describes the change law of the amplitude of the signal along with time and frequency.

Step S304: integration over the time axis of H (w, t) yields the Hilbert marginal spectrum:

step S305: and drawing a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of the original lung sound signal according to the Hilbert time-frequency spectrum and the Hilbert marginal spectrum.

Passing the instantaneous frequency f in the Hilbert-time spectrum and the Hilbert-marginal spectrum _i(t) and amplitude a _i(t) and the like, a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of the original lung sound signal can be drawn. The original lung sound signal is self-adaptive when being decomposed, so that the obtained Hilbert marginal spectrogram can accurately and intuitively reflect the actual frequency components of the signal.

And analyzing instantaneous frequency extreme values and amplitudes in the obtained Hilbert marginal spectrogram, analyzing the spectrum difference of a normal testee and a typical lung disease patient, and briefly analyzing the result.

According to the method, hardware such as a lung sound sensor, an amplifier, an A/D acquisition card and an upper computer are adopted to build a lung sound signal acquisition device, original lung sound signals acquired by the device are uploaded to a cloud for storage and analysis, when the original lung sound signals are analyzed, Hilbert-Huang transformation is firstly carried out on the original lung sound signals, and clutter signals with various different frequencies, which are influenced by internal organs and are superposed, are eliminated; and providing a lung sound intrinsic mode function, analyzing an intrinsic lung sound signal by using a Hilbert time-frequency spectrogram and a marginal spectrogram, and taking the frequency amplitude change of the lung sound as a judgment basis for detecting respiratory diseases.

Example 2:

acquiring lung sound signals and analyzing Hilbert-Huang transform on a plurality of experimenters with known health conditions, wherein an original lung sound signal oscillogram of a typical normal health person is shown in fig. 3(a), an original lung sound signal oscillogram of a typical respiratory disease patient is shown in fig. 3(b), Hilbert-Huang transform is respectively performed on the graphs in fig. 3(a) and 3(b), an intrinsic mode function IMF is extracted, obtained results are respectively shown in fig. 4(a) and 4(b), and Hilbert time-frequency spectrograms shown in fig. 5(a) and 5(b) are obtained, and Hilbert marginal spectrograms shown in fig. 6(a) and 6 (b).

By comparing the time-frequency spectrograms of fig. 5(a) and 5(b), it can be seen that the energy is intensively distributed at the low-frequency part along with the time change of the time-frequency spectrogram of the normal healthy person; in contrast, the time-frequency spectrum of a respiratory disease patient changes with time, and energy is dispersed in high frequency parts.

Further, the mean value of the lung sound signal of the normal healthy person is selected for analysis, fig. 7 is a waveform diagram of the mean value of the marginal spectrum of the normal healthy person, and the parameters in fig. 7 are the frequency extreme value, the amplitude value, and the area of the marginal spectrum. The results of the calculation of the marginal spectral mean of normal healthy persons are shown in table 1:

marginal spectrum analysis	Frequency of	Amplitude value	Area of
				Mean value	11.945Hz	0.366	5.757

TABLE 1

The average of the marginal spectral frequency and amplitude of normal healthy persons was (11.945Hz, 0.366), the area of the graph region was 5.757; in addition, the mean value of the marginal spectral frequency and amplitude of the respiratory disease patient is (358.313Hz, 0.031), and the area of the graph area is 0.530. The results of the calculation of the marginal spectral mean of normal healthy persons and patients with respiratory diseases are shown in table 2:

marginal spectrum analysis	Frequency of	Amplitude value	Area of
				Normal sound	11.945Hz	0.366	5.757
Abnormal sound	358.313Hz	0.031	0.530

TABLE 2

Through the data comparison and analysis in table 2, the frequency corresponding to the maximum peak in the marginal spectrum of the normal lung sound signal is concentrated in the low frequency part, and the frequency corresponding to the maximum peak in the marginal spectrum of the abnormal lung sound signal is higher than that of the normal lung sound signal. The result shows that when diseases occur in the human body to form abnormal sounds in the lung, such as the sounds of the gouty, the peak value of the lung sound signal gradually shifts to a high-frequency part, the amplitude of the sounds is reduced, and the frequency domain area is reduced.

The invention provides a lung sound signal analysis method, which records the breathing sound of a testee through a lung sound signal acquisition device, performs characteristic analysis by using the Siebert transform and obtains the following conclusion:

the hubert time spectrum and the boundary spectrum of the lung sound signal of a respiratory disease patient may have a different condition from those of a healthy person. The energy of the lung sound signal time-frequency spectrogram of a healthy person is intensively distributed at the low-frequency part of dozens of hertz, and the energy of the lung sound signal time-frequency spectrogram of a lung disease patient is relatively dispersed and is distributed at the high-frequency part of hundreds of hertz. Therefore, the normal healthy population and the population suffering from the respiratory disease can be classified by comparing and analyzing the time-frequency spectrogram and the frequency, amplitude and area corresponding to the maximum peak value of the marginal spectrogram, so as to judge whether the testee suffers from the respiratory disease.

Other parts of the embodiment are the same as those of the above embodiment, and thus are not described again.

The present invention is not limited to the above embodiments, and any person skilled in the art can easily conceive of changes or substitutions within the technical scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. The utility model provides a lung sound signal pickup assembly which characterized in that: the method comprises the following steps:

the lung sound sensor is used for collecting lung sound signals;

2. The lung sound signal collecting device according to claim 1, wherein: the model of the lung sound sensor is HKY-06F.

3. A lung sound signal analysis method is characterized in that: the method specifically comprises the following steps:

step S1: receiving an original lung sound signal;

4. The method of claim 3, wherein the method further comprises: the method comprises the following steps of performing Hilbert-yellow transform based on the intrinsic mode function to obtain a Hilbert time-frequency spectrogram and a Hilbert marginal spectrogram of original lung sound signals, wherein the Hilbert time-frequency spectrogram and the Hilbert marginal spectrogram of the original lung sound signals comprise the following steps:

z _i(t)＝p _i(t)+jH[p _i(t)]

will complex function z _i(t) is expressed in exponential form:

wherein the amplitude a _i(t) is:

phase phi _i(t) is:

is recorded as: