摘要
分析了基于小波分析的4种不同的去噪方法,并在其中寻找最适宜在实际中应用的心电信号中肌电干扰的去除方法。4种去噪方法分别采用软、硬及改进等3种阈值函数,通过MATLAB对MIT-BIH数据库中所提供的心电信号进行实验分析,根据去噪效果及所需时间对比结果判断最适宜的去噪方法。离散小波变换阈值法在采用3种阈值函数时去噪效果均较差,平移不变量小波阈值法本身运算量过大,平稳小波变换阈值法与提升小波变换阈值法在采用改进阈值函数时去噪效果好且所需时间相对较少。采用改进阈值函数的平稳小波变换阈值法与采用改进阈值函数的提升小波变换阈值法为4种方法中最适宜在心电信号肌电干扰去除中应用的方法。
Four different kinds of denoising methods based on wavelet analysis were analysed in this paper, and in order to look for the most suitable methods among them to remove the EMG interference from ECG signals in the practical application. These methods adopted the soft, the hard and the improved threshold functions respectively. Use MATLAB to do experiments with the ECG signals provided by the MIT - BIH database, and select the most suitable methods according to the comparison results of the de- noising effects and the time needed. The discrete wavelet transform threshold method gets a poor denois- ing effect when using the three kinds of threshold functions, the translation invariant wavelet threshold method needs much computation time itself, and the stationary wavelet transform threshold method and the lifting wavelet transform threshold method get a good denoising effect and need a little time relatively when using the improved threshold function. The stationary wavelet transform threshold method combined with the improved threshold function and the lifting wavelet transform threshold method combined with the improved threshold function are the most suitable methods to remove the EMG interference among the four kinds of methods.
出处
《工业仪表与自动化装置》
2015年第2期12-17,共6页
Industrial Instrumentation & Automation
基金
教育部博士点基金(20106201110003)
关键词
小波分析
心电信号
肌电干扰
阈值函数
wavelet analysis
ECG signals
EMG interference
threshold function
