小波变换结合快速傅里叶变换从PPG中提取呼吸率

Respiratory rate extracted from photoplethysmography signals by using wavelet transform and fast Fourier transform

从光电容积脉搏波描记法(PPG)信号中提取呼吸率是一种简便、高效、成本低的呼吸检测方法。本文用多道生理记录仪同时采集由温度传感器和透射式光电脉搏传感器获得的人体呼吸波和PPG信号,应用小波变换对PPG信号进行9层分解,将第9层细节信号和第8层细节信号分解得到的近似信号重建后相加得到呼吸波,然后用改进的快速傅里叶变换频率估计方法从该呼吸波信号中提取呼吸率。用该法从30个PPG样本中提取呼吸率,并将所提取的呼吸率与温度传感器获得的呼吸率用Bland-Altman法进行对比,得到了两者具有良好一致性的结论。

The respiratory rate was extracted from the signal of photoplethysmography （PPG）, which was a simple, high efficient and low-cost method for respiratory monitoring. Respiratory wave and PPG signal, which were respectively acquired by temperature sensor and transmitted optical-electronic pulse sensor, were recorded by multi-channel physiological recorder simultaneously. PPG signal underwent the decomposition with nine-level by using wavelet transform, and the detail signals of ninth-level and eighth-level were decomposed to obtain approximate signals. The obtained signals were rebuilt, and the sum of the rebuilted signals was respiratory wave. An improved fast Fourier transform frequency estimation method was adopted to extract respiratory rate from the respiratory wave. Respiration rate was extracted from 30 PPG samples by using the method. Bland-Altman method was used to compare the extracted respiratory rate with the reference rate obtained by temperature sensor. The comparison results showed the two respiratory rates had a good consistency.