基于光电容积脉搏波的人体生理指数检测算法研究与实现

发布时间：2018-06-27 09:53

本文选题：光电容积脉搏波描记法 + 生理指数　；参考：《东北大学》2011年硕士论文

【摘要】：光电容积脉搏波描记法(PPG)是借助于光电技术手段,在活体组织中检测血液容积变化的一种无创检测方法。由于血液容积在人体心脏的周期性脉动作用下会发生周期性的脉动变化,所以通过光接收器的光强也随之呈脉动性的变化,将获得的光信号转化为电信号即是光电容积脉搏波。容积脉搏波中包含很多有关于心血管系统的重要信息,但是目前为止,利用光电手段得到的人体生理指数只有脉率和血氧饱和度,在其他生理指数的测量中或者仅仅停留在理论阶段,或者不能对其进行标定。在实际测量人体PPG信号时,由于仪器中电子元件所产生的噪声、电磁场的干扰、环境光的变化以及人体的呼吸和运动等因素的影响,得到的响应常常包含很多的噪声和基线漂移。针对上述问题,首先设计了基于小波变换的低通滤波器和基于经验模态分解(EMD)的去除基线漂移算法,对容积脉搏波进行滤波去噪处理,去除包括工频在内的高频干扰和低频基线漂移；然后提出了基于微分阈值法、梯形面积法和指数加权移动平均(EWMA)的特征参数值提取算法,提取PPG波形的特征参数,并利用这些特征参数计算出人体脉率、脉率变异性、动脉血压、血氧饱和度、心搏出量、心输出量、外周阻力、动脉顺应性、心搏指数、心脏指数等人体生理指数,同时本文还提出了基于滑动窗口的奇异点阈值剔除算法,结合人体生理机能将超出生理范围内的奇异点剔除；最后对动脉血压和波形特征参数进行相关性和回归分析,分别构造出收缩压和舒张压与波形特征参数的回归方程。实验表明,利用本文所提出的算法可以有效的提取光电容积脉搏波的波形特征参数,计算出的人体生理指数与其它仪器检测的结果基本吻合,可以应用到无创光电监测仪中。此外,利用回归分析构造回归方程计算人体动脉血压为血压的无创测量提供了种新的手段和方法。
[Abstract]:PPG is a noninvasive method for detecting the changes of blood volume in living tissues by means of photoelectric technique. Because the blood volume changes periodically under the action of the periodic pulsation of the human heart, the intensity of the light through the light receiver also fluctuates with it. The obtained light signal is converted into an electrical signal, that is, the photovolumic pulse wave. Volume pulse waves contain a lot of important information about the cardiovascular system, but so far, only pulse rates and oxygen saturation are the only physiological indices obtained by optoelectronic means. Other physiological indices are measured only at the theoretical level or can not be calibrated. In the actual measurement of human PPG signal, due to the noise produced by electronic components in the instrument, the interference of electromagnetic field, the change of environmental light and the breathing and movement of human body, etc. The resulting response often contains a lot of noise and baseline drift. To solve the above problems, a low-pass filter based on wavelet transform and a baseline drift removal algorithm based on empirical mode decomposition (EMD) are designed, and the volumetric pulse wave is filtered and de-noised. After removing high frequency interference and low frequency baseline drift including power frequency, a method based on differential threshold method, trapezoidal area method and exponentially weighted moving average (EWMA) is proposed to extract the characteristic parameters of PPG waveform. And using these characteristic parameters to calculate human body physiological indexes such as pulse rate, pulse rate variability, arterial blood pressure, blood oxygen saturation, cardiac output, peripheral resistance, arterial compliance, cardiac beat index, cardiac index, etc. At the same time, this paper also proposed a sliding window based singular point threshold elimination algorithm, combined with human physiological functions to remove the singular points beyond the physiological range. Finally, the correlation and regression analysis of arterial blood pressure and waveform characteristic parameters are carried out. The regression equations of systolic and diastolic blood pressure and waveform characteristic parameters are constructed respectively. The experimental results show that the proposed algorithm can effectively extract the waveform characteristic parameters of photovolumic pulse wave, and the calculated body physiological index is in good agreement with the results of other instruments, and can be applied to the non-invasive photoelectric monitor. In addition, using regression analysis to construct regression equation to calculate human arterial blood pressure provides a new method for noninvasive measurement of blood pressure.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：R331;TP274

【参考文献】