一种基于级联无损声管的心音产生模型的研究及应用

发布时间：2018-03-04 21:32

本文选题：心音　切入点：心音产生模型　出处：《南京邮电大学》2017年硕士论文　论文类型：学位论文

【摘要】：心音是指在心动周期中,由于心脏肌肉的收缩和舒张,心脏瓣膜的开启与关闭,血流冲击心室内壁和动脉血管等因素引起的一系列机械振动,经过体表组织传到胸壁表面的音频信号。心音信号蕴含了心脏内部结构与发声过程的关系,心音的特征与心脏内部形状特征也存在着一定的内在联系。因此,研究心音产生模型的主要目的是通过对心音产生模型的研究,提出一种能够体现心音产生过程的模型,有助于进一步理解心音产生机理;将心音模型应用在对异常心音的研究中,为诊断异常心音并对其进行病理分析提供了可靠的理论基础。本文提出了一种基于级联无损声管的心音产生模型,并将心音产生模型应用于分类识别和异常心音分析领域。首先,从声学角度出发,结合行波理论,提出了一种基于级联无损声管的心音产生模型,它包含S1级联无损声管模型和S2级联无损声管模型。将心室及动脉血管抽象为一系列长度相同,内半径不等的多级声管,通过控制声管内半径的大小去模拟一个心动周期中心房、心室以及动脉血管的舒张与收缩状态,并以共振峰频率作为目标参数来仿真心音的产生过程。该模型具有以下三个特征:第一,心音是一种声学信号,本文从声学角度出发,以共振峰频率作为模型的目标参数直接体现心音的声学特性;第二,将心脏内部的心室和动脉血管抽象为长度相等,内半径不等的多级声管,将心音产生模型与心脏内部的形状及尺寸特征联系起来,凸显了心音产生模型与心音产生机理之间的联系,对于探索心音产生原理具有积极的作用;第三,模型参数为9级内半径{Rm}(m取1,2,3...,9),模型简单直观,模型参数易于控制,目标参数为S1和S2的共振峰频率,具有维度低、物理意义明确等特点。其次,利用该心音产生模型对异常心音进行了仿真,对比异常心音与正常心音的模型参数,对异常心音进行病理分析,分析结果与临床医学的表现基本一致。最后,基于共振峰理论,提出了以S1和S2共振峰频率作为特征的心音分类识别方法。该特征能够很好的表征心音的能量分布特征,将其作为心音的一种识别特征具有较高的识别率。本文提出的心音产生模型不仅可以模拟正常心音,而且可以模拟异常心音,仿真结果表明,该模型对于从声学角度了解心音的产生过程,探寻异常心音的病理原因,拓展心音信号的应用范围,具有积极的意义。
[Abstract]:Cardiac sound is a series of mechanical vibration caused by the contraction and relaxation of heart muscle, the opening and closing of heart valve, and the impact of blood flow on the inner wall of ventricle and artery during the cardiac cycle. The audio signal transmitted through the surface of the body surface to the surface of the chest wall. The heart sound signal contains the relationship between the internal structure of the heart and the sound process, and the characteristics of the heart sound and the internal shape of the heart are also related to each other. The main purpose of the study of heart sound production model is to put forward a kind of model which can reflect the process of heart sound production through the study of heart sound production model, which is helpful to further understand the mechanism of heart sound production. The application of heart sound model in the study of abnormal heart sound provides a reliable theoretical basis for diagnosing abnormal heart sound and analyzing it pathologically. In this paper, a heart sound production model based on cascaded lossless sound tubes is proposed. The heart sound production model is applied to classification recognition and abnormal heart sound analysis. Firstly, a heart sound generation model based on cascaded lossless acoustic tubes is proposed from the acoustic point of view and combined with traveling wave theory. It consists of S1 cascade lossless acoustic tube model and S2 cascade lossless sound tube model. The ventricular and arterial vessels are abstracted into a series of multistage acoustic tubes of the same length and varying inner radius. By controlling the radius of the acoustic tube to simulate the relaxation and contraction of the central atrium, ventricle, and arterial vessels in a cardiac cycle, The model has the following three characteristics: first, the heart sound is an acoustic signal. The resonance peak frequency is taken as the target parameter of the model to directly reflect the acoustic characteristics of the heart sound. Secondly, the ventricular and arterial vessels inside the heart are abstracted into multistage acoustic tubes with equal length and different internal radii. The relationship between the heart sound production model and the shape and size characteristics of the heart is highlighted, which plays an active role in the exploration of the heart sound production principle. Third, the relationship between the heart sound production model and the heart sound generation mechanism is highlighted. The model parameter {RM} / m is 1 / 2 / 3..0.The model is simple and intuitive, the model parameter is easy to control, the target parameter is the resonance peak frequency of S _ 1 and S _ 2, the dimension is low, the physical meaning is clear and so on. The abnormal heart sound is simulated by using the model, and the model parameters of abnormal heart sound and normal heart sound are compared, and the pathological analysis of abnormal heart sound is carried out. The results are basically consistent with the clinical medical performance. Finally, based on the theory of resonance peak, A method of heart sound recognition based on S _ 1 and S _ 2 resonance peak frequency is proposed, which can well represent the energy distribution of heart sound. The model proposed in this paper can not only simulate normal heart sounds, but also can simulate abnormal heart sounds. The model has positive significance for understanding the process of producing heart sound from the angle of acoustics, exploring the pathological cause of abnormal heart sound, and expanding the range of application of heart sound signal.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R540.4;TN911.7

【参考文献】