心房颤动信号处理及社区心脏病人监护系统研制

发布时间：2018-07-17 21:23

【摘要】：目前，房颤（Atrial Fibrillation，AF）信号提取算法大部分是针对多导联心电图（Electrocardiogram，ECG），而多导联采集在心电监护中移动性不强，单导联监护系统由于灵活方便，将成为未来“可移动化”的房颤监护系统的发展趋势。本文主要研究了单导联房颤信号提取、房颤类型分类和开发了远程心电自动监护诊断系统。从单导联提取房颤信号的现有方法精度受心电波形形态以及噪声的影响较为严重，方法的鲁棒性较差。鉴于此，本文提出了一种新的单导联房颤信号提取方法。新方法利用房颤信号在时间上的非平稳性，把单导联ECG进行扩维（分段），然后再利用盲源提取算法进行房颤信号的提取。实验结果表明，新方法能够很好地从单导联ECG中提取房颤信号，计算时间较短，具有应用于实时无线房颤监护系统的前景。在提取房颤信号的基础上，为了加深对房颤自发终止机制的进一步理解，改善对持续性房颤的治疗，本文还对阵发性和持续性房颤的分类进行了研究。本文采用主成分分析从单导联心电信号中提取出房颤信号，选择和计算出房颤信号的特征，，最后用分类器对阵发性和持续性房颤进行了分类。本文首次用复杂度去表征了房颤波波动复杂度的特征。实验结果表明，预测的总正确率是90%。在1000次随机性实验中，最高分类正确率可达到92%，平均正确率为77.12%。该方法可以用来很好地分类两种房颤，对房颤自发性终止的预判有一定的指导意义。最后，本文研究和开发了远程心电自动监护诊断系统，其主要内容包括心电Q、R、S、P、T特征波的检测、数据波形显示、客户端和服务端的心电数据的实时通信、数据库管理以及病人的病例管理系统等。这项工作在社区化医疗系统中具有一定应用价值。
[Abstract]:At present , most of the extraction algorithms of atrial fibrillation ( AF ) signal are directed to multi - lead electrocardiogram ( ECG ) , while multi - lead monitoring is not very mobile in ECG monitoring . Because of its flexibility and convenience , the single - lead monitoring system will become the development trend of the future " mobile " atrial fibrillation monitoring system . This paper mainly studies the extraction of single - lead AF signal , the classification of AF type and the development of remote ECG monitoring and diagnosis system .

This paper presents a new method for extracting atrial fibrillation signal from single lead ECG . The results show that the new method can extract the atrial fibrillation signal from single lead ECG well , and the calculation time is short , which has the prospect of applying to real - time wireless AF monitoring system .

In order to deepen the understanding of AF self - termination mechanism and improve the treatment of persistent AF , this paper studies the classification of paroxysmal and persistent AF by analyzing the characteristics of atrial fibrillation signal by means of principal component analysis . The results show that the total accuracy rate is 92 % and the average accuracy is 77.12 % . The method can be used to classify the two kinds of atrial fibrillation well , and has certain guiding significance for the pre - determination of spontaneous termination of atrial fibrillation .

Finally , the remote ECG monitoring and diagnosis system is studied and developed in this paper . The main contents include the detection of ECG Q , R , S , P , T characteristic wave , data waveform display , real - time communication of ECG data of client and server , database management and patient ' s case management system . This work has certain application value in community medical system .
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R318.6

【参考文献】