Respiration rate automatic calculation method for respiration sound processing of electronic stethoscope

Abstract

The invention belongs to the technical field of biomedical signal processing, and particularly relates to an automatic respiratory rate calculation method for respiratory sound processing of an electronic stethoscope, which comprises the following steps of: S1, acquiring a period of respiratory sound signal with the duration of Ts (Ts is greater than or equal to 15 seconds) and the sampling rate of fs by using the electronic stethoscope; s2, performing short-time Fourier transform on the obtained breath sound signal vector x; s3, calculating the amplitude S of the time-frequency spectrum matrix S; s4, carrying out homomorphic envelope st on the time domain vector # imgabs0 # of the intermediate frequency component; and S5, searching all peak values greater than a preset threshold Eth on the time domain homomorphic envelope st of the intermediate frequency component. By calculating the time domain homomorphic envelope of the intermediate frequency component, main interference components including heart sound can be automatically eliminated, the detection specificity of the breath sound signal component is greatly improved, and then the calculation precision of the breath rate is improved; respiratory cycle detection and respiratory rate calculation can be realized only by searching a peak value meeting a threshold condition on the time domain homomorphic envelope of the intermediate frequency component without any pattern recognition method with relatively high complexity.

Description

Technical Field

[0001] The invention belongs to the technical field of biomedical signal processing, and in particular relates to a method for automatically calculating a respiratory rate for processing respiratory sounds of an electronic stethoscope. Background Art

[0002] In recent years, with environmental pollution and changes in lifestyle, respiratory diseases have become increasingly prevalent. Therefore, wearable real-time monitoring of the respiratory system and cardiopulmonary activity is of great significance.

[0003] Respiratory rate is an important physiological indicator for real-time monitoring of cardiopulmonary activity. It is mostly monitored by pulmonary function testers. Compared with pulmonary function testers, electronic stethoscopes have the advantages of low price, easy use, and long-term wearable monitoring. The respiratory sound signal collected by the electronic stethoscope is the turbulent vibration caused by the respiratory airflow passing through the respiratory tract and lungs. It can be recorded in the form of sound signals and can reflect the respiratory health status at the first time.

[0004] Designing an acoustic signal processing algorithm is expected to automatically calculate the respiratory rate based on the respiratory sound signal, but the difficulty lies in:

[0005] (1) When recording respiratory sound signals, electronic stethoscopes are particularly susceptible to interference from other sound signals (such as heart sound interference and background noise interference);

[0006] (2) The complexity of the automatic calculation algorithm of respiratory rate should be as low as possible to meet the needs of real-time applications.

[0007] Therefore, we propose an automatic calculation method of respiratory rate for electronic stethoscope respiratory sound processing to solve the above problems. Summary of the invention

[0008] The object of the present invention is to provide a method for automatically calculating the respiratory rate for electronic stethoscope respiratory sound processing in view of the above problems.

[0009] To achieve the above object, the present invention provides the following technical solution: a method for automatically calculating respiratory rate for electronic stethoscope respiratory sound processing, comprising the following steps:

[0010] S1. Use an electronic stethoscope to obtain a period of time T s (T s ≥15 seconds), sampling rate is f s The breathing sound signal can be expressed as a length of T s ·f s The vector x;

[0011] S2. Perform short-time Fourier transform on the acquired respiratory sound signal vector x, with a window length of N w , thus obtaining the time-frequency spectrum matrix S, whose dimension is Nf ×N t , where N f =N w / 2+1,

[0012] S3. Calculate the amplitude |S| of the time-frequency spectrum matrix S, and average the components in |S| with a frequency range of 200Hz to 325Hz in the frequency dimension to obtain a matrix with a length of N. t The time domain vector of the intermediate frequency component

[0013] S4. Time domain vector of intermediate frequency component Find the homomorphic envelope s t , and its calculation formula is

[0014] S5. Find the time domain homomorphic envelope of the intermediate frequency component t is greater than the preset threshold E th All peaks of , assuming there are P peaks that meet the conditions, their indexes are recorded as N1, N2, ..., N P ;

[0015] S6. If P≤1, it is considered that there are no two or more breathing cycles in the signal segment, and the breathing rate cannot be calculated. Return to S1 to continue to obtain the breathing sound signal; if P≥2, the breathing rate calculation result can be given. The calculation method is: using the k-1th and kth adjacent peaks that meet the conditions, we can get t k =N k N w / (2f s ) is the respiration rate at the moment RespRate k =60 / [(N k -N k-1 )N w / (2f s )].

[0016] In the above-mentioned method for automatically calculating the respiratory rate for processing electronic stethoscope respiratory sounds, the step S4 L (·) is a forward and backward Butterworth low-pass filter, and δ>0 is a preset minimum value.

[0017] Compared with the prior art, the present invention provides a method for automatically calculating the respiratory rate for electronic stethoscope respiratory sound processing, which has the following beneficial effects:

[0018] This automatic calculation method for respiratory rate in electronic stethoscope respiratory sound processing can automatically exclude major interference components including heart sounds by calculating the time-domain homomorphic envelope of the intermediate frequency component, greatly improving the detection specificity of the respiratory sound signal component, and thus improving the calculation accuracy of the respiratory rate; respiratory cycle detection and respiratory rate calculation can be achieved by simply finding a peak that meets the threshold condition on the time-domain homomorphic envelope of the intermediate frequency component, without the need for any highly complex pattern recognition method, and is therefore particularly suitable for use in remote online auscultation systems.

[0019] In summary: the algorithm of the present invention has low complexity and can automatically exclude major interference components including heart sounds, providing an accurate and effective technical means for intelligent auscultation. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic flow chart of a method for automatically calculating respiratory rate for processing respiratory sounds using an electronic stethoscope proposed by the present invention;

[0021] Figure 2 This is a schematic diagram of an original respiratory sound signal waveform, an extracted intermediate frequency component time domain homomorphic envelope waveform, and a final respiratory rate calculation result in an embodiment of a method for automatically calculating respiratory rate for electronic stethoscope respiratory sound processing proposed by the present invention. DETAILED DESCRIPTION

[0022] The following examples are for illustrative purposes only and are not intended to limit the scope of the present invention.

[0023] See also Figure 1-2 , a method for automatically calculating respiratory rate for electronic stethoscope respiratory sound processing, characterized in that it comprises the following steps:

[0024] S1. Use an electronic stethoscope to obtain a period of time T s (The duration of this sample data is T s >90 seconds), sampling rate is f s =8KHz breathing sound signal, which can be expressed as a length of T s ·f s The vector x;

[0025] S2. Perform short-time Fourier transform on the acquired respiratory sound signal vector x, with a window length of N w = 256, using the Hamming window, to obtain the time-frequency spectrum matrix S, whose dimension is N f ×N t , where N f =N w / 2+1=129,

[0026] S3. Calculate the amplitude |S| of the time-frequency spectrum matrix S, and average the components in |S| with a frequency range of 200Hz to 325Hz in the frequency dimension to obtain a matrix with a length of N. t The time domain vector of the intermediate frequency component

[0027] S4. Time domain vector of intermediate frequency component Find the homomorphic envelope s t , and its calculation formula is where f L (·) is a forward-backward 4th-order Butterworth low-pass filter with an upper cutoff frequency of 0.015 and a preset minimum value of δ = 10 -6 .

[0028] S5. Find the time domain homomorphic envelope of the intermediate frequency component t is greater than the preset threshold E th All peak values, the preset threshold E th By t The average of the first 90% of the values ​​is given. In this example, P = 18 peaks that meet the conditions are found, and their indexes are recorded as N1, N2, ..., N P ;

[0029] S6. Since P ≥ 2, the respiratory rate calculation result can be given. The calculation method is: using the k-1th and kth adjacent peaks that meet the conditions, we can get t k =N k N w / (2f s ) is the respiration rate at the moment RespRate k =60 / [(N k -N k-1 )N w / (2f s )], and finally the real-time respiratory rate of P-1=17 moments is given, Figure 2 The original breathing sound signal waveform, the extracted intermediate frequency component time domain homomorphic envelope waveform and the final breathing rate calculation result in this embodiment are given.

[0030] By calculating the time-domain homomorphic envelope of the intermediate frequency component, the main interference components including heart sounds can be automatically excluded, which greatly improves the detection specificity of the respiratory sound signal component and thus improves the calculation accuracy of the respiratory rate. Respiratory cycle detection and respiratory rate calculation can be achieved by simply finding the peak that meets the threshold condition on the time-domain homomorphic envelope of the intermediate frequency component, without the need for any highly complex pattern recognition methods. Therefore, it is particularly suitable for use in remote online auscultation systems.

[0031] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for automatically calculating respiratory rate for electronic stethoscope respiratory sound processing, characterized in that: The steps include: S1. Use an electronic stethoscope to obtain a period of time T s (T s ≥15 seconds), sampling rate is f s The breathing sound signal can be expressed as a length of T s ·f s The vector x; S2. Perform short-time Fourier transform on the acquired respiratory sound signal vector x, with a window length of N w , thus obtaining the time-frequency spectrum matrix S, whose dimension is N f ×N t , where N f =N w / 2+1, S3. Calculate the amplitude |S| of the time-frequency spectrum matrix S, and average the components in |S| with a frequency range of 200Hz to 325Hz in the frequency dimension to obtain a matrix with a length of N. t The time domain vector of the intermediate frequency component S4. Time domain vector of intermediate frequency component Find the homomorphic envelope s t , and its calculation formula is S5. Find the time domain homomorphic envelope of the intermediate frequency component t is greater than the preset threshold E th All peaks of , assuming there are P peaks that meet the conditions, their indexes are recorded as N1, N2, ..., N P ; S6. If P≤1, it is considered that there are no two or more breathing cycles in the signal segment, and the breathing rate cannot be calculated. Return to S1 to continue to obtain the breathing sound signal; if P≥2, the breathing rate calculation result can be given. The calculation method is: using the k-1th and kth adjacent peaks that meet the conditions, we can get t k =N k N w / (2f s ) is the respiration rate at the moment RespRate k =60 / [(N k -N k-1 )N w / (2f s )].

2. The method for automatically calculating respiratory rate for electronic stethoscope respiratory sound processing according to claim 1, characterized in that: The S4 L (·) is a forward and backward Butterworth low-pass filter, and δ>0 is a preset minimum value.

Images