| The antenatal examination is an important part of Perinatal Medicine,one of the its aims is the fetal health monitoring, and it is the main meansof prepotency and giving birth healthily. There are millions upon millionsof pregnant women each year in China, The need of fetal health monitoringis too large to meet，because it is worked mainly in the hospital,and it hasmany disadvantages, it cannot meet the real-time, dynamic, and remotemonitoring needs of fetal health monitoring. This paper analyzes the theoryand characteristics of fetal heart sound，then gives out the detection systemof fetal heart sound based on the the model of " Equipment，Cloud andClient ", which with a capture device, the APP in the smart mobile deviceand the Cloud, to collect the fetal heart sound, denoising, featurerecognition and the Maternal health management based on the trends offetal heart rate. The system not only meets family-safe fetal healthmonitoring, but also provide ongoing health care services and relatedsolutions for maternal nutrition, exercise and health, health management,which is important in social value.The main contents of this paper are as follows:(1) By comparing with the ultrasound, FECG and other technologies of fetal health monitoring, then the paper points out the advantages of thefetal heart sound detection technology, summarizes the research progressin this field, and indicates that the main difficulties of fetal heart sounddetection technology are the signal collection and denoising methods.(2) To summary the mechanism, conduction path, frequency-domainand time-domain characteristics of fetal heart sound, and indicating thatfetal heart sound is one of biomedical signals, which is weak,low-frequency (main band20-200Hz),nonlinear and non-stationary.Because of these characteristics and much noise, it is so difficulty toprocess the fetal heart sound.(3) Work out a fetal heart sound detection system of "Equipment，Cloud and Client" model, including the peripheral and the composition andfunction of APP.(4) Work out a prototype experimental system for collecting fetal heartsound based on LabVIEW and myDAQ, and the audio ports of myDAQused to collect the signal is proving that the it is possible to collect andcommunicate data with the audio port of smart mobile terminal.(5) Fetal heart sound processing (mainly the signal denoising) is thefocus research of this paper. Its innovation is the first to process the fetalheart sound with Hilbert-Huang Transform. This method eliminates thebaseline drift and noise at once, to compare with wavelet transform,mathematical morphology and the Fourier transform, the SNR improve obviously, the MSE is the minimum, although the time complexity is thebiggest, it can satisfy the practical application.Processing the fetal heart sound, firstly, the preprocessing with a cutofffrequency of200Hz low-pass filter and the resampling. Secondly, the signaldecomposition based on EMD of Hilbert-Huang transform, then denoisingsome selected target components with wavelet soft threshold adaptive noisecancellation algorithm, finally get the clean fetal heart sound by combiningthe target components. In the EMD, using a mask signal to eliminate themode mixing problem, using mirrorizing extension method to eliminate theend effect, and the stopping rule is referencing from the research of Rilling.(6) To extract the signal envelope of the clean fetal heart sound basedon Hilbert transform, then make feature recognition based on the thresholdmethod. In this paper the accuracy rate of calculating the average fetal heartrate is98.7%. |
PDF Full Text Request