Research On ECG Signal Pretreatment And Characteristic Points Detection Based On Wavelet Transform

The heart disease endangers human health and affects people’s lives,so people have been devoted to analysis and study on it. People who diagnose heart disease mainly depend on the amplitude and phase of a series of waveforms of ECG signal,because the heart information mainly concentrates in these areas.But, ECG signal is often accompanied by the interference noises in the process of data acquisition, and it is not conducive to analyze it that these noises make the signal become blurred easily.The research is conducted at removing the noise and detecting the characteristic points in the ECG signal in this paper.In the noise removal from the ECG signal,the Mallat algorithm was used to remove the baseline wander noise with sym8according to the denoising experiment using various wavelet functions.When removing the powerline interference and myoelectricity interference noises, the Mallat algorithm and stationary wavelet transform were used to do the denoising experiment analysis combined with sym8and the threshold denoising method, which adopted the match of six kinds of thresholds and three kinds of threshold functions,and the stationary wavelet transform got the best denoising effect when using the improved VisuShrink threshold and the hard threshold function match and suppressed the Pseudo-Gibbs phenomenon effectively.In the detection of ECG signal characteristic points,the time-invariant stationary wavelet transform was used to decompose the ECG signal into five layers with the first derivative of the smooth function-quadratic B spline wavelet.The R wave detection was achieved according to the modulus maximum pairs generated by the R wave located on the cd4,and the Q wave、S wave、the starting point and end point of QRS waves were detected according to the modulus maximum pairs generated by the Q wave and S wave located on the cd2. The peak point、starting point and end point of P wave and T wave were detected according to the modulus maximum pairs generated by the P wave and T wave on both sides of the modulus maximum pairs generated by the R wave on the cd5. The detection of ST segment was made behind the detection of QRS waves、P wave and T wave, the ST segment starts from the end point of QRS waves and ends to the starting point of T wave, the starting point of the ST segment was used to judge the depression and elevation compared with the base point,the form of the depression and elevation was judged by the slope of the ST segment and the extent was judged by the characteristic point located in the middle position of the ST segment.In the detection of characteristic points,some measures which prevent some points being checked by mistake and make the position-correction as needed were made to improve the detection precision,and the experiment showed that these measures were necessary.