面向可穿戴设备的脉搏波基线漂移去除算法  

A Pulse Wave Baseline Drifting Removal Algorithm for Wearable Devices

作　　者：许金林[1] 李晓风[1] 李皙茹 元沐南[1] 

机构地区：[1]中国科学院合肥物质科学研究院,安徽合肥230031

出　　处：《计算机技术与发展》2017年第11期14-18,共5页Computer Technology and Development

基　　金：国家科技支撑计划课题(2013BAH14F01);国家自然科学基金资助项目(61301059)

摘　　要：为满足可穿戴便携测量实时性分析的要求,在广义数学形态学滤波算法的基础上,提出了一种简化的PPG(photoplethysmography)基线漂移去除算法。该算法在应用广义形态滤波器算法与简化的形态滤波算法对红外发射管与环境光学传感器所采集人体的PPG信号进行处理的基础上,对校正后的信号进行相似度计算,然后应用静态波峰识别算法分别进行心率值计算。实验结果表明,分别采用广义形态滤波器算法与简化的形态滤波算法处理后的PPG信号相似度高达88.83%,标准心率值相关度分别为98.61%和98.68%;简化算法处理后的校正信号与广义形态滤波后的信号相比,漂移基线校正能力接近,运算适应性更强,计算量减少了4倍,为可穿戴设备实时分析提供更好的软件支撑。In order to meet the requirements of wearable portable real-time measurement analysis, a simplified approach for baseline drift removal in photoplethysmogmphy （PPG） is proposed based on the mathematical morphology. The PPG collected by the infrared emission tube and environmental optics sensor is processed by generalized morphological filter algorithm and simplified morphological filtering al- gorithm and on the basis of that, the similarity between them is calculated. Then the static peak recognition algorithm is applied for the calculation of value of the heartbeat respectively. The experimental results show that the similarity of two filtered PPG is up to 88.83 % and the goodness of fit between the gold heart rate and heartbeat obtained from the two algorithms is 98.61% and 98.68% respectively. Compared with generalized morphological filter algorithm,it has similar ability to filter data but can contribute to quicker 4 times than o- riginal morphology method,which makes it better applied in wearable device in dally healthy real-time analysis.

关 键 词：PPG 可穿戴测量 实时分析 形态学 基线漂移 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]