基于非线性分形理论的心音信号分析研究
发布时间：2019-06-21 09:53
【摘要】：心音信号是一种重要的生理信号，其中包含着心脏系统各种生理和病理信息。因为心脏是一个复杂的非线性动力系统，决定了由其振动产生的心音信号也是一种非线性信号，具有内在的复杂性。而现有的心音信号分析方法多是建立在心音信号线性模型基础之上，不足以研究本质上为非线性的生命活动。分形理论作为非线性科学的一个重要分支，可以很好的揭示出非线性过程内在所具有的特殊规律。本论文试图从非线性角度，将分形理论应用到心音信号的分析研究，从本质上更深入的探索心音信号的内在特征规律。同时将分形理论的定性分析和定量分析结合起来，利用计盒维数算法估算人群心音信号的分形维数值，并将其应用到心脏功能评估，以一个全新角度实现对心音信号的分析研究。 论文首先分析了心音信号的时频域特点，并对心音信号和杂音信号进行短时傅里叶变换和Welch功率谱估计，研究结果表明正常心音信号和心杂音成分的频率区间不同。而且心音信号的时域和频域特征都有一定的规律性，为后续对心音信号进行非线性分析奠定基础。 然后利用小波阈值去噪方法对心音信号进行预处理，通过分析噪声在小波分解下的特性和小波函数的性质，基于重构因子方法选取coif3小波去除心音信号噪声，并通过实验结果对比证明coif3小波对心音信号的去噪效果较好。 将分形理论应用到心音信号研究的前提需保证心音信号具有分形特征，本文利用分形布朗运动模型和心音信号模型进行对比分析，定性证明了心音信号具有自相似性和无标度性这两个重要的分形特征，为后续利用分形理论研究心音信号提供依据。接着论文介绍了计盒维数、关联维数、信息维数、Hausdorff维数和相似维数五种分形维数估算方法，由于计盒维数算法原理简单且计算量小，本课题最终选择该算法对人群心音信号进行研究分析。 最后论文分别对健康孕妇(153例)和妊娠期疾病孕妇(64例)、健康孕妇(153例)和健康未孕妇女(161例)及普通大学生(65例)和运动员(97例)三组人群心音信号样本的计盒维数值进行调研统计，结果显示健康孕妇分形维数值(1.27±0.29)相比健康未孕妇女(1.31±0.41)较低，健康孕妇分形维数值(1.27±0.29）相比妊娠期疾病孕妇（1.38±0.05）较低，第三组样本运动员运动后的分形维数值(1.30±0.065)与静息状态(1.39±0.037)相比降低，普通大学生运动后的分形维数值(1.32±0.054)与静息状态(1.41±0.065)相比降低，均具有统计学差异，而静息状态运动员与普通大学生的分形维数值相对比没有显著性差异，且运动后运动员与普通大学生的分形维数值相对比也没有显著性差异。并将计盒维数值与D/S比值、S1/S2比值、HR三种指标进行对照研究，，分析结果说明应激情况下会调用人体的心力储备和心率储备，使D/S比值下降、S1/S2比值升高、HR变快，通过这些指标可以对人群心脏功能进行综合评估。 研究结果初步表明，分形维数值可定量描述心音信号的复杂程度，而且不同人群心音信号的分形维数值具有一定的统计学差异。探索性的将分形维数值作为一个评估人群心脏功能的指标，让医生和科研人员了解应激状态下心脏功能的变化，为临床提供有价值的信息，为更细致的保护心脏提供依据。
[Abstract]:The heart sound signal is an important physiological signal, which contains various physiological and pathological information of the heart system. Because the heart is a complex nonlinear dynamical system, it is determined that the heart sound signal produced by its vibration is also a non-linear signal, which has inherent complexity. The existing method of heart sound signal analysis is based on the linear model of the heart sound signal, which is not enough to study the vital activity which is non-linear in nature. As an important branch of non-linear science, the fractal theory can well reveal the special rules inherent in the nonlinear process. In this paper, from the non-linear angle, the fractal theory is applied to the analysis and study of the heart sound signal, and the intrinsic characteristic of the heart sound signal is further explored from the essence. At the same time, the qualitative analysis and quantitative analysis of the fractal theory are combined, the fractal dimension value of the human heart sound signal is estimated by using the box-counting dimension algorithm, and the fractal dimension value of the human heart sound signal is estimated by using the box-counting dimension algorithm, and the method is applied to the cardiac function evaluation, and the analysis and research of the heart sound signal are realized by a new angle. In this paper, the characteristics of the time and frequency domain of the heart sound signal are analyzed, and the short time Fourier transform and the Welch power spectrum estimation are carried out on the heart sound signal and the noise signal. The results show that the frequency range of the normal heart sound signal and the cardiac murmur component is not and the time domain and the frequency domain characteristic of the heart sound signal have certain regularity, which lays the foundation for the follow-up of the non-linear analysis of the heart sound signal In this paper, the wavelet threshold de-noising method is used to pre-process the heart sound signal, and by analyzing the characteristics of the noise and the properties of the wavelet function under the wavelet decomposition, the coif3 wavelet is selected to remove the heart sound signal based on the reconstruction factor method. The noise of the coif3 wavelet to the heart sound signal is proved by the comparison of the experimental results. It is better to apply the fractal theory to the study of the heart sound signal. It is necessary to ensure that the heart sound signal has the fractal feature. This paper makes a comparative analysis with the fractal Brownian motion model and the heart sound signal model, and it is proved that the heart sound signal has both self-similarity and non-scaling. The fractal feature of the study of the heart sound in the follow-up of the fractal theory In this paper, five fractal dimension estimation methods of the dimension of the box, the correlation dimension, the information dimension, the Hausdorff dimension and the similarity dimension are introduced. The box dimension of heart sound signal samples from three groups of healthy pregnant women (153 cases) and pregnant women (64 cases), healthy pregnant women (153 cases) and non-pregnant women (161 cases) and general college students (65 cases) and athletes (97 cases) were analyzed. The results showed that the fractal dimension of healthy pregnant women (1.27-0.29) was lower than that of pregnant women (1.31-0.41) and that of pregnant women (1.38%) compared with that of pregnant women (1.27-0.29). The fractal dimension (1.30-0.065) after the movement of the third group of sample athletes was lower than that of resting state (1.39-0.037). The fractal dimension (1.32-0.054) after the normal college students' movement was lower than that of resting state (1.41-0.065). There is a statistical difference, while the relative ratio of the rest state athletes to the fractal dimension of the normal college students is no significant difference, and the relative ratio of the post-sports athletes to the fractal dimension of the normal college students is also There was no significant difference in the ratio of D/ S, S1/ S2 and HR, and the results of the analysis indicated that the heart and heart rate reserve of the human body was called under the stress condition, and the ratio of D/ S was decreased and the ratio of S1/ S2 was decreased. Increased HR, through which the heart work of the population can be achieved. The results of the study show that the fractal dimension value can quantitatively describe the complexity of the heart sound signal, and the fractal dimension value of the heart sound signal of different people There is a certain statistical difference. The exploratory fractal dimension value is used as an index for evaluating the heart function of the population, so that the doctors and the researchers can understand the changes of the heart function in the state of the heart, provide valuable information for clinic, and provide more detail
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R318.0
[Abstract]:The heart sound signal is an important physiological signal, which contains various physiological and pathological information of the heart system. Because the heart is a complex nonlinear dynamical system, it is determined that the heart sound signal produced by its vibration is also a non-linear signal, which has inherent complexity. The existing method of heart sound signal analysis is based on the linear model of the heart sound signal, which is not enough to study the vital activity which is non-linear in nature. As an important branch of non-linear science, the fractal theory can well reveal the special rules inherent in the nonlinear process. In this paper, from the non-linear angle, the fractal theory is applied to the analysis and study of the heart sound signal, and the intrinsic characteristic of the heart sound signal is further explored from the essence. At the same time, the qualitative analysis and quantitative analysis of the fractal theory are combined, the fractal dimension value of the human heart sound signal is estimated by using the box-counting dimension algorithm, and the fractal dimension value of the human heart sound signal is estimated by using the box-counting dimension algorithm, and the method is applied to the cardiac function evaluation, and the analysis and research of the heart sound signal are realized by a new angle. In this paper, the characteristics of the time and frequency domain of the heart sound signal are analyzed, and the short time Fourier transform and the Welch power spectrum estimation are carried out on the heart sound signal and the noise signal. The results show that the frequency range of the normal heart sound signal and the cardiac murmur component is not and the time domain and the frequency domain characteristic of the heart sound signal have certain regularity, which lays the foundation for the follow-up of the non-linear analysis of the heart sound signal In this paper, the wavelet threshold de-noising method is used to pre-process the heart sound signal, and by analyzing the characteristics of the noise and the properties of the wavelet function under the wavelet decomposition, the coif3 wavelet is selected to remove the heart sound signal based on the reconstruction factor method. The noise of the coif3 wavelet to the heart sound signal is proved by the comparison of the experimental results. It is better to apply the fractal theory to the study of the heart sound signal. It is necessary to ensure that the heart sound signal has the fractal feature. This paper makes a comparative analysis with the fractal Brownian motion model and the heart sound signal model, and it is proved that the heart sound signal has both self-similarity and non-scaling. The fractal feature of the study of the heart sound in the follow-up of the fractal theory In this paper, five fractal dimension estimation methods of the dimension of the box, the correlation dimension, the information dimension, the Hausdorff dimension and the similarity dimension are introduced. The box dimension of heart sound signal samples from three groups of healthy pregnant women (153 cases) and pregnant women (64 cases), healthy pregnant women (153 cases) and non-pregnant women (161 cases) and general college students (65 cases) and athletes (97 cases) were analyzed. The results showed that the fractal dimension of healthy pregnant women (1.27-0.29) was lower than that of pregnant women (1.31-0.41) and that of pregnant women (1.38%) compared with that of pregnant women (1.27-0.29). The fractal dimension (1.30-0.065) after the movement of the third group of sample athletes was lower than that of resting state (1.39-0.037). The fractal dimension (1.32-0.054) after the normal college students' movement was lower than that of resting state (1.41-0.065). There is a statistical difference, while the relative ratio of the rest state athletes to the fractal dimension of the normal college students is no significant difference, and the relative ratio of the post-sports athletes to the fractal dimension of the normal college students is also There was no significant difference in the ratio of D/ S, S1/ S2 and HR, and the results of the analysis indicated that the heart and heart rate reserve of the human body was called under the stress condition, and the ratio of D/ S was decreased and the ratio of S1/ S2 was decreased. Increased HR, through which the heart work of the population can be achieved. The results of the study show that the fractal dimension value can quantitatively describe the complexity of the heart sound signal, and the fractal dimension value of the heart sound signal of different people There is a certain statistical difference. The exploratory fractal dimension value is used as an index for evaluating the heart function of the population, so that the doctors and the researchers can understand the changes of the heart function in the state of the heart, provide valuable information for clinic, and provide more detail
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R318.0
【参考文献】
相关期刊论文 前10条
1 曹汉强,朱光喜,李旭涛,夏文芳;多重分形及其在地形特征分析中的应用[J];北京航空航天大学学报;2004年12期
2 方向林,郭兴明,彭承琳,詹直夫,肖守中;运动和呼吸对心率及心音幅值影响的探讨[J];北京生物医学工程;2002年02期
3 姚丽慧;高井祥;王坚;;动态变形信号小波阈值降噪模型研究[J];测绘工程;2011年05期
4 王琳;郭兴明;成敏;郭玮珍;肖守中;廖克龙;;孕期和产程心脏储备无创性检测和评估方法的初步研究[J];第三军医大学学报;2006年04期
5 陈果;;非线性时间序列的动力学混沌特征自动提取技术[J];航空动力学报;2007年01期
6 孙博文,张本祥;上证指数的分形结构及盒维数测量的研究[J];哈尔滨理工大学学报;2001年06期
7 骆懿;赵治栋;;一种便携式心音信号记录仪研究[J];杭州电子科技大学学报;2011年04期
8 许可;孟o
本文编号：2503987
本文链接：https://www.wllwen.com/yixuelunwen/swyx/2503987.html
