基于小波熵的心电信号去噪处理  

Denosing Processing of ECG Signal Based on Wavelet Entropy

作　　者：侯宏花[1] 桂志国[1] 

机构地区：[1]中北大学仪器科学与动态测试教育部重点实验室,信息与通信工程学院,太原030051

出　　处：《中国生物医学工程学报》2010年第1期22-28,34,共8页Chinese Journal of Biomedical Engineering

基　　金：山西省自然科学基金(20051021);山西省高等学校科技项目(20081024)

摘　　要：实测的心电信号不可避免地存在一些强干扰和噪声,如何在强背景干扰和噪声下准确提取出有用的心电信号,是心脏病智能诊断的一个重要内容。提出一种新的基于小波熵的弱心电信号去噪方法,先将信号小波分解,再对不同分解尺度上的高频系数进行小波熵阈值的量化处理,然后利用最高一层小波分解的低频系数分量和经过阈值处理的不同尺度的高频小波系数分量,组成进行信号重构所需要的系数分量进行重构,将严重的干扰和噪声去掉,实现有效信号的提取。最后分别利用临床的实测心电数据和M IT/B IH心电数据库信号进行验证,并针对不同噪声类型和不同信噪比情况进行分析。结果表明,该方法简单有效,尤其对于高频噪声效果更优,且适于实际应用。There are inevitably strong disturbance and noise in the measured ECG signal. How to extract the ECG wave accurately with strong background disturbance and noise is an important part in the intelligent diagnosis of heart disease. In this paper, a new method based on wavelet entropy was presented. Firstly, the signal was disposed of wavelet decomposition. Secondly, the high-frequency wavelet decomposition coefficients of different scale was processed with wavelet entropy thresholds and quantification. Then the low-frequency coefficients of the highest level wavelet decomposition and the high-frequency wavelet coefficients of the different scales after the threshold processing were composed into the wavelet coefficients component for reconstruction. Then the serious disturbance and noise were removed and the signal was extracted. The proposed method was verified using the clinical data and the data from MIT-BIH arrhythmia database. The different noise types with different SNR was analyzed. The results indicated that this method was simple, effective, accurate, especially for removing high-frequency noise.

关 键 词：心电信号 小波熵 小波变换 去噪处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]