【作者】 张佑春；

【导师】 王平；

【作者基本信息】 西华大学 ， 电力电子与电力传动， 2009， 硕士

【摘要】 心音信号是人体最重要的生理信号之一,富含心脏各个部分如心房、心室及心血管功能状态的大量病理信息,是临床评估心脏功能状态的最基本参数。在心血管疾病尚未发展到足以产生临床及病理改变之前,心音中出现的杂音和畸变就是重要的信息。脉搏信号同样携带有丰富的心血管健康状况信息,因此,记录心音和脉搏图对于心血管监护具有重要的参考价值。本文在分析国内外有关研究现状的基础上,对多源信息融合理论作了较为详细的介绍,分析几种融合方法的优缺点,并最终选用了在处理不确定信息方面具有明显优势的Dempster-Shafer证据理论。考虑到监护系统一般要求具有良好的实时处理能力,而监护系统的数据量又非常大,且要以最快的速度来到达检测目标,此时D-S证据理论就显得有些力不从心,为了弥补理论的不足而影响系统的实时性,我们借助了具有良好的非线性映射能力和自学习能力的BP神经网络。在监护系统软件设计方面,以图形化虚拟仪器开发软件LabVIEW 8.2为平台,开发出了友好的监护系统界面,主要有实时采集界面、信号分析界面和记录管理界面;系统软件设计主要实现心音、脉搏信号的采集、显示、信号处理、特征提取、存储、回放等功能。在监护系统的数据存储方面,本文采用数据库技术,研究开发了病员管理数据库。考虑到监护系统对数据库实时访问的要求,本文选用LabSQL实现对数据库的访问,数据库选用广泛使用的小型桌面关系型数据库管理系统——Microsoft Access 2003。本次研究综合运用了多源信息融合技术、神经网络技术、数字信号处理技术、数据库技术,等等。本文设计的监护系统通过数字化心音、脉搏波等传感器采集与心血管健康相关的心音、脉搏波等信号,经过计算机的处理与分析获取心音、脉搏等特征参数,利用智能化的信息融合方法观察人体的心血管功能是否正常。整个系统本身具有为无创伤、实时性好等优点。更多还原

【Abstract】 Heart sound signal is one of the most primary physiological signals.It contains a great deal of pathological information which belongs to each parts of heart,such as atria,ventricle and cardiovascular functional status.It is also an important essential parameter used to evaluate heart functional status. Before the development of cardiovascular disease having not yet been sufficient to produce clinical and pathological changes, the appearance of noise and distortion in phonocardiogram indicates some important information.There is abundant cardiovascular health information in pulse signal,so it has significant reference value for cardiovascular monitoring to record phonocardiogram and pulse diagram.At the basis of analyzing the situation of related research at home and aroad,this thesis has made more detailed description of multi-source information fusion theory. What’s more,it has analyzed the advantages and disadvantages of several fusion methods. Dempster-Shafer evidence theory has been selected fmally,which has obvious advantages in dealing with uncertain information.Considering the weakness of D-S evidence theory and the requirement of monitoring system for real-time processing capability, reaching target dection as soon as possible,we resort to the BP neural network which has a good nonlinear mapping ability and self-learning ability.On the software design of monitoring system side,three friendly panels of monitoring system have been completed,including real-time sampling panel,signal analysis panel and record management panel at the platform of graphic virtual instrument development software LabVIEW 8.2. The software design mainly realizes heart sound and pulse sampling,display,processing,characteristics extracting,storage, playback,etc.On the data storage of monitoring system side,this thesis designs patients management database by use of database technology.In view of monitoring system visiting database real-timely, LabSQL is chosen to access database developed by Microsoft Access 2003,which is a small desktop relational database management system.This study makes good use of multi-source information fusion technology, neural network technology,digital signal processing technology,database technology,ect.The designed monitoring system acquires heart sound signal and pulse signal from digital heart sound sensor and pulse sensor,which reflect cardiovascular health.The characteristic parameters are obtained by virtue of computer processing and analysis. The use of intelligent information fusion method is used to observe whether the body’s cardiovascular function is normal or not. The entire system itself has non-invasive, real-time advantages and so on.更多还原