Denoising Of The Ecg Qrs Detection Algorithm

The research of ECG automated analysis and diagnosis system has been focus in signal processing field in the past years. It will be a great fruit to put the search into practice with a reliable performance,not only in medical field, but also in modern signal processing field and artificial intelligence field. Rapid and accurate detection of the QRS complex is a key link in ECG automated analysis and diagnosis system, and at the same time eliminating effectively all kinds of noise interferences from ECG signal is the premise of accurate detection of various waves with features in ECG signal. Consequently this dissertation carries on the study in the key technolgies of two main aspects which concerns about ECG signal de-noising, QRS complex detection and fiducial point location.ECG signal is a complex non-stationary random signals which is sensitivity to the noise. In this part the traditional digital filter de-noising method and the de-noising algorithm based on wavelet transform are focused on. In the classical digital filtering method, according to spectral characteristics of ECG signal and noise the equiripple Chebyshev FIR and zero-phase butterworth IIR digital filters are designed. By the simulation experiment which applies two kinds of filter banks to the actual ECG signal de-noising, the effectiveness and performance of two filter banks in de-noising were verified and compared. In part of Wavelet transform-based denoising algorithms, this paper mainly discusses the wavelet modulus maxima denoising method, wavelet threshold de-noising method and translation invariant wavelet de-noising algorithm. For the defect of traditional soft and hard thresholding function a new threshold function is put forward and through numerical experiments the validity and superiority of this new threshold function is proved. Finally an improved de-noising algorithm is proposed by combining the new threshold function and translation invariant wavelet de-noising method. Through de-noising experiments using the MIT-BIH database shows that this new algorithm has a better de-noising performance for ECG signal compared to the wavelet threshold de-noising method while the expense is cost of computing speed.About QRS detection, at first the paper discussed in depth the method based on the wavelet transform. The QRS detection mothod base on the dual-orthogonal quadratic B-spline wavelet transform is researched in focus. The experiment validated the detection accuracy and location precision of this algorithm is quite high while it also has strong anti-interference ability and can accurately locate the start and end of the QRS complex. Then, for a sharp decline in detection rate of this algorithm when ECG signal was seriously interfered with noises, the assembled QRS detection program was proposed based on wavelet soft-threshold de-noising pre-processing. This assembled method can greatly increased the detection accuracy and location precision of QRS complex. Finally, this paper discusses QRS detection algorithm based on EMD which is a new signal processing method. And a new QRS detection algorithm is proposed based on EMD method and the Marr wavelet transform. The new method integrated EMD method and the singularity detection principle based on the Marr wavelet transform modulus maximum. This new algorithm has quite high rate of detection accuracy, fast processing speed and high positioning accuracy. It can provide an accurate analytical tools for ECG automated processing.