Research Of Heart Sound Signal Detection System Based On Wavelet And HHT

The activity of heart directly affects the pulse and blood pressure in the four vital signs of life,and contains rich life information.Heart sound is the vibrational information of the heart mechanical activity reflected by the sound wave transmission,which contains a lot of physiological and pathological information of the heart.The detection of heart sound signals is an important means of obtaining information about the heart's mechanical activity.Based on the principle of heart sound signal generation mechanism and the principle of acoustic stethoscope transmission,combined with the time-frequency characteristics of heart sound signal,this paper proposes a research program of heart sound signal detection system based on correlation wavelet entropy threshold de-noising and Hilbert-Huang Transform(HHT)analysis.It will provide the foundation for establishing the heart sound file with the characteristic parameters of heart sound signal as the main content and having the reference value of clinical diagnosis,and studying the big data analysis related to heart sound in the future.The main contents of this paper are as follows:1.The heart sound generation mechanism is introduced,combined with the time-frequency characteristics of heart sounds,and the characteristic parameters of heart sound signals are determined.The sources of noise and interference during the collection of heart sound signal were analyzed.Further,the general scheme of the heart sound signal detection system which mainly includes the heart sound signal collecting device,the heart sound signal processing method and the analysis method is proposed.2.On the basis of analyzing the acoustic principle of acoustic stethoscope,and according to the time-frequency characteristics of the heart sound signal,the heart sound signal is picked up by combining the acoustic stethoscope hearing head,the sound sensor,the hardware and software parts of the heart sound signal collecting device are designed.3.Considering the characteristics of heart sound signal,noise and interference,combined with the theory of correlation and wavelet entropy,the correlation wavelet transform entropy threshold denoising method is used to denoise the heart sound signal.Contrast experiments with soft and hard threshold denoising methods were performed on the heart sound signal with not obvious Third Heart Sound(S3)and not obvious Fourth Heart Sound(S4),the heart sound signal of obvious S3 and not obvious S4,and the heart sound signal of obvious S3 and obvious S4.Further,the feasibility of denoising based on correlation wavelet entropy threshold is verified.4.For the problem of envelope extraction and instantaneous frequency acquisition of heart sound signal,the HHT analysis method is introduced to analyze the heart sound signal to obtain the instantaneous frequency and the envelope of heart sound signal,so as to extract the time-frequency feature parameters of the heart sound signal.Three kinds of de-noised heart sound signals were taken as the analysis object to conduct HHT analysis experiments,and the characteristic parameters of the heart sound signal in the time-frequency domain were extracted.5.The heart sound acquisition experiment was designed by using the designed heart sound collecting device.The heart sound signals of 10 subjects were randomly selected for processing and analysis,and the characteristic parameters of the heart sound signals were extracted to verify the feasibility of heart sound signal detection system based on wavelet and HHT.