Fetal Heart Rate Detection Algorithm And Implementation Base On Blind Source Separation

Cardiac care is an important means to ensure the safety of fetal and maternal,which also can realizes the national policy of eugenic and superior nurture.For monitoring the fetus in the womb,we can see the status of the fetal’s growing situation,so as to reduce the rate of birth defects and fetal death caused by the distress of hypoxia or ischemia.Heart rate of electronic monitoring is one of the most extensive fetal monitoring methods,for fast detection and analysis of the heart rate has great difficulties,which also become the research direction in recent years.This paper summarizes a kinds of algorithms for heart rate detection, and proposes an new algorithm based on blind source separation to detect instantaneous heart rate of Doppler cardiac sound. The algorithm unites the distribution information of the fetal heart sound signal in time domain and frequency domain, then use the prior knowledge of cardiac sound signal’s periodicity to select several frequencies which on behalf of the signal features.Through independent component analysis,let different independent signals separation,after that,select the component which is the most relevant with the original signal,then do relevant calculation for the most matching component and detect the peak,last calculate the instantaneous heart rate cardiac sound. The paper works include the following aspects:First,introduces and summarizes some main method of detecting the fetal heart rate;put forward the main idea of the algorithm in this article studies;introduces some relevant theoretical knowledge, including short-time Fourier transform and blind source separation and independent component analysis, etc.Secondly,make more introduce for the Doppler cardiac sound detection algorithm base on blind source separation in details.Basic steps are as follows:First of all, do de-noising pretreatment for signal collected, then do short time Fourier transform for the original signals to get the time-frequency diagram, use the prior knowledge of cardiac sound signal’s periodicity to select several frequencies which on behalf of the signal’s features in the time-frequency diagram.select each feature spectrum as an observation signal, and do FastICA operation for the observation signal.after that let different independent signals separation. Select the component which has the most relevant with the original signal,then do relevant calculation for the most matching component and detect the peak,last calculate the instantaneous heart rate cardiac sound. After all of these,do analysis on multiple samples of different data about the reliability of algorithm though MATLAB simulation.Thirdly,design and realize the Doppler cardiac sound instantaneous heart rate detection system based on blind source separation,The system consists of hardware parts and PC software parts,where hardware part includes:ultrasonic Doppler cardiac sound signal acquisition module, AD conversion module, system control module, control module achieve to the task coordination and data encapsulation,Software part mainly realize signal processing function, cardiac sound detection of instantaneous heart rate, and hardware are also including in this part of the data communication and signal separation print management functions.