【作者】 王玉静；

【导师】 宋立新；

【作者基本信息】 哈尔滨理工大学 ， 信号与信息处理， 2007， 硕士

【摘要】 心电信号(ECG)是诊断心脏疾病的主要依据。对心电信号进行分析处理,以及时准确获取病人生理活动可靠的内部特征信息,是生物医学研究者所面临的重要问题。心电信号是一种微弱低频的非平稳信号,具有随机性和噪声背景强等特点。对其进行分析过程中需要反映其局部变化的特征。时频分析方法能够很好地反映信号时频局部变化的特征,决定了其在心电信号分析领域的重要性。人体心电信号一般比较微弱,幅度仅在10μV~4 mV,极易受环境中各种干扰的影响,心电信号分析前期的去噪处理工作显得至关重要;R波在心电信号特征波形中占有非常重要的地位,它的检测定位为我们提取心电信号的其他参数提供最基本的依据;心率变异性信号中含有人体自主神经活动的重要信息,提取其瞬时特征信息,可以为病理分析及疾病诊断提供可靠的定量指标。本文主要基于小波变换方法以及Hilbert-Huang变换方法对心电信号进行了包括去噪、R波检测以及心率变异性分析的研究。并通过大量的实验结果对比分析,验证了它们各自的优缺点。小波变换的多分辨率分析的特点,能很好地反映信号的局部特征。但是小波基的选择问题,以及小波分解的非单频率分量问题给研究工作者带来一定困难。Hilbert-Huang变换方法的自适应性以及时频聚集等特点体现了其应用于心电信号分析的优势。更多还原

【Abstract】 Recording of electrocardiograms (ECG) signal is an important basis for heart disease diagnosis. For getting accurate intrinsic characteristic message of physiology activity of patient in time, ECG signal analysis and processing becomes an important problem facing medicinal researchers. ECG signal is a weak low frequency non-stationary signal, and possessing random character and noisy setting strong character. So we need to reflect local vary character during analyzing process. Time-frequency analysis method can reflect signal time-frequency local alter, and occupies important action in ECG signal analysis field.ECG signal usually is weaker, amplitude scope is only from 10μV to 4 mV, so it can be influenced by various interfere easily, so signal de-noising processing is very crucial; R wave takes up very vital position in ECG signal characteristic waves, R wave accurate position is the basic for extracting else parameters of ECG signal; heart rate variability signal contains important message of body autonomic nervous system, extracting instantaneous character message, can afford us reliable quantitative index for pathology analysis and disease diagnosis.This thesis mainly analyzing ECG signal, based on wavelet transform method and Hilbert-Huang transform method, containing signal de-noising problem, R wave detection problem and heat rate variability signal analysis research. Through large experiment results comparing analysis, proves their each excellences and flaws. Multi-resolution analysis character of wavelet transform can well reflect signal local characteristic. But the selection problem of wave base function and not single frequency component problem of wavelet decomposition are very difficult to researchers. Hilbert-Huang transform method possesses self-adaptation and time-frequency centralizing, show its advantage in ECG signal analysis filed.更多还原