一种近红外无创血糖检测中的脉搏波预处理方法

Pulse wave pretreatment method applied to near-infrared blood glucose non-invasive measurement

目的:研究人体光电容积脉搏波提取动态光谱的方法,实现人体血糖浓度的无创伤检测。方法:采用自行研制的光电容积脉搏波采集系统采集人体脉搏波。针对采集到的原始信号中由于呼吸和系统噪声等影响产生的基线漂移和高频噪声,提出基于经验模态分解(empirical mode decomposition,EMD)结合软阈值去噪及三次样条插值法进行去噪的方法(即EMD-Spline法),并用信号的频谱进行算法效果评价。结果:该方法有效去除了原始脉搏波信号中的基线漂移和高频噪声。结论:该方法提高了血糖浓度测定的准确性,将促进近红外无创检测人体血糖浓度精度的提升。

Objective To explore a dynamic spectrum method based on photoelectric volume pulse wave to detect human blood glucose concentration noninvasively.Methods A self-developed pulse wave acquisition system was used to collect human pulse wave.The baseline draft and high-frequency due to respiratory and system noises were eliminated by the empirical mode decomposition（EMD） combined with soft threshold denoising or EMD-Spline algorithm,and algorithm efficiency evaluation was carried out with signal spectrum.Results This method eliminated the original baseline drift and high frequency noise of pulse wave signal at the same time.Conclusion The method improves the precision of the final determination of blood glucose concentration and the accuracy of near-infrared spectroscopy in non-invasive detection of human blood glucose.