脉搏信号中有效信号识别与特征提取方法研究

Effective signal recognition and feature extraction of pulse signal

脉搏信号是重要的人体生理信号,但采集过程中会存在一定的干扰信号。针对人体脉搏波采集后的有效信号识别与特征提取的问题,提出一种基于时间序列描述的信号识别方法,首先将脉搏时间序列进行分割,每个分割段采用斜率符号化进行表示,通过段与段之间的相似性判断出信号的有用段和干扰段。再根据得到的有用段信号,提出一种基于滑窗的特征提取方法,寻找脉搏信号中的峰值、谷值,同时调整滑窗宽度,还能够进行重搏波波峰的检测。经实验验证,所提出的识别与特征提取方法准确率高且抗干扰性强。

The pulse signal is an important human physiological signal , but there is some interference signal in the collection process. In this paper,feasible and effective measurement method based on the time series was proposed and implemented to solve the problem of effective signal recognition. Firstly, this method divided time series and used multiple slopes to show the each sub-section of time series. Judging by the similarity between two paragraphs, it could tell the useful and interference signal. Secondly, another method based on sliding window was proposed to solve the problem of feature extraction. The experiment show that the method is effective and accurate.