心电信号处理关键技术的研究与实现

发布时间：2018-12-14 22:55

【摘要】：据调查显示,心血管疾病引起的死亡人数占每年全世界死亡人数的三分之一且发病率正逐年增加,群体不断年轻化,已经成为严重危害人体健康的第一大疾病。由于心血管疾病发病具有突发性和不规律特性,提前预防和及时发现是治疗心血管疾病的关键步骤。心肌缺血、心绞痛等疾病在心电图上表现为心律失常波形,因此面向心律失常的心电信号处理算法和心电监护系统的研究对于挽救生命有着重要的意义。针对目前心律失常信号处理、分析和智能诊断算法中存在的不足,本文就心电信号的预处理(去噪)、波形检测及异常波形分类等关键技术进行研究,并针对算法应用的心电监护与分析诊断系统也做了相应的探讨。具体内容如下所示：(1)针对ECG信号中存在工频、基线漂移和肌电干扰等的噪声问题,本文设计了集改进的Notch滤波器、自适应形态滤波结构和平稳小波阈值去噪为一体的心电预处理算法。通过仿真实验,验证本文预处理算法相对其他经典滤波算法而言滤波效果更好。(2)针对异常心电波形检测不准问题,本文提出了基于双正交样条小波滤波器组和LS估计的心电P-QRS波形检测算法,实现对QRS波与P波的位置及宽度的检测。该算法通过构造双正交样条小波对心电信号进行二进小波分解,在第三和第四尺度上检测R波和P波峰值对应的模极大值对,实现R波和P波峰值的定位,并利用LS估计的方法确定QRS波和P波的宽度。仿真结果表明,本文的算法对R波和P波的的检测准确率较对比算法有所提高。(3)针对心电信号中心律失常波形分类问题,本文提出了改进的KNG-FCM心律失常分类算法。利用改进K均值、高斯核函数和粒度原理对传统FCM算法改进,改善了算法的鲁棒性和对噪声的敏感程度。仿真结果表明,本文提出的算法分类准确率能达到98.86%,较对比算法有所提高。(4)针对现有心电监护方案的不足,提出了由动态心电监护仪、心电智能分析软件组成的心电监护系统,可实现动态心电的远程监测和智能分析诊断。
[Abstract]:According to the investigation, the death toll caused by cardiovascular disease accounts for 1/3 deaths in the world every year and the incidence rate is increasing year by year. The population is getting younger and has become the first major disease that seriously endangers human health. Because of the sudden and irregular characteristics of cardiovascular disease, prevention and timely detection are the key steps in the treatment of cardiovascular disease. Myocardial ischemia, angina pectoris and other diseases show arrhythmia waveform on electrocardiogram, so the study of ECG signal processing algorithm and ECG monitoring system for arrhythmia is of great significance for saving lives. In view of the shortcomings of the current algorithms of arrhythmia signal processing, analysis and intelligent diagnosis, this paper studies the key technologies of ECG signal pretreatment (denoising), waveform detection and abnormal waveform classification, etc. The ECG monitoring and analysis diagnosis system which is applied by the algorithm is also discussed. The main contents are as follows: (1) aiming at the noise problems of power frequency, baseline drift and electromyography interference in ECG signal, an improved Notch filter is designed in this paper. An adaptive morphological filter structure and stationary wavelet threshold denoising algorithm for ECG preprocessing. The simulation results show that the preprocessing algorithm is better than other classical filtering algorithms. (2) aiming at the problem of irregular ECG waveform detection, This paper presents an ECG P-QRS waveform detection algorithm based on biorthogonal spline wavelet filter banks and LS estimation to detect the position and width of QRS and P waves. In this algorithm, biorthogonal spline wavelets are constructed to decompose ECG signals, and at the third and fourth scales, the mode-maximum pairs corresponding to the peaks of R and P waves are detected to locate the peaks of R and P waves. The width of QRS wave and P wave is determined by LS estimation. The simulation results show that the detection accuracy of R wave and P wave is higher than that of the contrast algorithm. (3) aiming at the problem of arrhythmia waveform classification in ECG signal, an improved KNG-FCM arrhythmia classification algorithm is proposed in this paper. The improved K-means, Gao Si kernel function and granularity principle are used to improve the robustness and noise sensitivity of the traditional FCM algorithm. The simulation results show that the classification accuracy of the proposed algorithm can reach 98.86, which is better than the comparison algorithm. (4) aiming at the shortcomings of the existing ECG monitoring scheme, a dynamic ECG monitor is proposed. The ECG monitoring system composed of ECG intelligent analysis software can realize the remote monitoring and intelligent analysis diagnosis of dynamic ECG.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.7;R540.41

【相似文献】