Research And Implementation Of ECG Signal Processing Key Technology

According to the survey, the number of deaths caused by cardiovascular diseases accounts for one-third of deaths worldwide each year. The incidence of cardiovascular is increasing year by year and the group becomes younger. It has been the first major disease which harms to human health seriously. Since cardiovascular disease has the sudden and irregular characteristic, early prevention and timely detection is the key step in the treatment of this diseases.The deseases such as myocardial ischemia and angina respond the arrhythmia wave in Dynamic ECG, so the research of ECG signal processing algorithms for arrhythmia and intelligent ECG device have great significance of saving lives.For the shortage of the current arrhythmia ECG signal processing, analysis and intelligent diagnostic algorithms, this thesis does the research of key technology in ECG signal pre-processing (denoising), waveform detection and waveform automatic classification. Specific contents are as follows:(1) For the problems of ECG signal with power-line interference, baseline drift and EMG interference, an pre-processing algorithm is proposed which contains improved Notch filter, adaptive morphological filtering structure and stationary wavelet threshold denoising method. The simulation results show the ECG pre-processing algorithm has a better filtering effect than other classical algorithms.(2) For the abnormal ECG waveform inaccurate detection problem, the ECG P-QRS wave detection algorithm based on biorthogonal spline wavelet filter group and LS estimation is proposed which can detect the location and width of QRS wave and P wave. ECG signal is decomposed by constructing biorthogonal spline wavelet. It can locate the R wave and P wave peak position by detecting the modulus maxima respectively in third and fourth scales. The width of the QRS wave and P wave is determined by using LS estimation method. Simulation results show that accuracy rate of R wave and P wave detection in this algorithm is improved compared with the contrast algorithms.(3) For the classification of arrhythmia wave in ECG signal, this thesis presents an improved KNG-FCM arrhythmia classification algorithm. It uses the improved K-means, Gaussian kernel function and granularity principle to improve the traditional FCM algorithm. This algorithm improves the robustness of the algorithm and sensitivity to noise. Simulation results show that the classification accuracy of proposed algorithm can reach 98.86% which is improved compared to the comparison algorithms.(4) Aiming at the deficiency of existing ECG monitoring scheme, this thesis presents the ECG monitoring system which includes dynamic ECG monitor and intelligent ECG analysis software. It can realize remote ECG monitoring and intelligent diagnosis.