脉搏波信号特征点识别与预处理方法研究

发布时间：2018-05-16 09:46

本文选题：脉搏波 + 数学形态学　；参考：《生物医学工程学杂志》2015年01期

【摘要】：脉搏波信号特征包含人体心血管的生理病理信息,准确识别其特征点对分析人体心血管的健康状况有重要意义。脉搏波特征点因人而异,且易受干扰,因而获取高信噪比的完整信号是准确识别特征点的前提。本文基于数学形态学,设计了一种组合滤波器对信号作预处理,可有效抑制基线漂移并去除高频噪声。利用特征点与小波变换过零点的位置对应关系,识别预处理后信号的特征点,并提出用微分法来校正特征点的位置偏移。采用三个参数可调的高斯函数重构出四种典型的脉搏波进行对照实验,证明了本文方法识别脉搏波特征点的准确性。
[Abstract]:The characteristics of pulse wave signal contain physiological and pathological information of human cardiovascular system. It is important to accurately identify the characteristic points for analyzing the health status of human cardiovascular system. The characteristic points of pulse wave vary from person to person and are easily interfered, so obtaining complete signal with high signal-to-noise ratio is the premise of identifying feature points accurately. Based on mathematical morphology, a combined filter is designed to preprocess the signal, which can effectively suppress baseline drift and remove high frequency noise. Based on the relationship between the position of feature points and the zero-crossing points of wavelet transform, the feature points of pre-processed signals are identified, and the differential method is proposed to correct the position shift of feature points. Four typical pulse waves were reconstructed by Gao Si function with three adjustable parameters. The results show that the method proposed in this paper is accurate in identifying the characteristic points of pulse waves.
【作者单位】：南方医科大学生物医学工程学院;广州医科大学附属肿瘤医院;
【基金】：广东省科技计划基金项目资助(2012A032200013)
【分类号】：R318.08

【参考文献】