Study On Fetal Heart Rate's Detection And Non-linear Analysis

Fetal heart rate's detection is one of the most important projects of fetal monitoring. Changes in fetal heart rate are closely related to fetal well-being, oxygen status and dynamic status. Perinatal period, especially in late pregnancy and parturition process, monitor fetal heart rate, timely detection of fetal hypoxia and distress in uterus, take appropriate measure to protect the health of the newborn.Accurate extraction of fetal heart rate is the prerequisite of determine the health status of fetus. Ultrasound doppler heart rate detection with high sensitivity, non-invasive and easy to operate, etc., has been widely used. However, due to the human body is a complex exercise environment, especially in the perinatal period, fetal heart rate is relatively weak. In addition, maternal environmental interference, obtain ultrasound doppler fetal heart echo signal component more complex and interference, difficult to directly read the value of fetal heart rate, therefore necessary to deal with de-noising.In this paper, translation invariant wavelet de-noising method is used to denoising for ultrasound doppler fetal heart echo signal, give suitable decomposition level, wavelet base, threshold function and threshold approach. Compare the denoising effect between traditional wavelet threshold de-noising method and translation invariant wavelet de-noising method (the traditional wavelet threshold de-noising method can effectively eliminate the noise in signals, but at the same time may produce Psuedo-Gibbs phenomena). The research results show that this method can effectively suppress signal noise and Pseudo-Gibbs phenomenon, distinguish useful signal which is neglected by the noise, and obtain better visual effect.After effective detect fetal heart rate, the further work analyze the fetal heart rate signals. This paper's main method is non-linear analytical method, the non- linear parameters such as correlation dimension, largest Lyapunov exponent and Kolmogrov entropy are mainly discussed, analyze its power spectral characteristics and complexity metrics-complexity and approximate entropy. The research results show that it's power spectrum is a continuous spectrum with a pinnacle, this shows that the fetal heart rate signal is not a simple noise; its correlation dimension is the fraction, largest Lyapunov exponent and Kolmogrov entropy are positive, these show that the fetal heart rate signal have non-linear characteristics; complexity is between 0 and 1, approximate entropy is obviously differ from noise, these show that the fetal heart rate signal seems complicated but the system has a certain regularity. All these indicate that the fetal heart rate system is a system of non-linear factors; fetal heart rate signal has a certain non-linear factors.