Study On ECG Signal Detection And Analysis Algorithms Using Wavelet Transform

The cardiac disease is one of the main diseases to harm human body now. ECG is an important method for the diagnosis in the cardiac disease in clinic. The exact automatic analysis for ECG signal is the key to the diagnosis in the cardiac disease, and which is also the problem many researchers pay a mount of attention to solve. QRS complexes, P and T waves are important key features of electrocardiogram (ECG), so automatic detection of ECG waves (QRS complexes, P and T wave) is important to cardiac disease analysis and diagnosis. The detection of ECG signal is began form QRS complexes. Calculation and analysis of heart rate and P and T waves are based on the precise detection of QRS complexes, so that it is essential for the ECG automatic detection.Dissertation focuses on the further research on the algorithm of QRS complex detection based on the previous work.Firstly, QRS complexes detection method based on wavelet transform of real value wavelet basic function and golden section search was discussed. The selection of wavelet basic function is an important factor to influence the results of wavelet decomposition. Comparing the mathematic characteristic of eight kinds of wavelet basic function, we selected an optimal wavelet basic function suited to ECG signal. Furthermore, the selection of decomposition scale is another important factor to ECG denosing results. Based on analysis of signal 3-db bandwidth frequency at different scales and its own frequency of QRS complexes, we determinate the appropriate decomposing scale for better denosing. In addition, we analysis the limitations of QRS complexes detection method, when use the modulus maximum for detection only, after using wavelet transform to decompose ECG signal. To conquer the limitation we proposed a new QRS complexes detection method using wavelet transform and golden section search. In order to evaluate the method, we used records from MIT-BIH arrhythmia database to detect the QRS complexes. And the detection rate is 99.6%. The proposed method is better than the method using modulus maximum for detection only.The important key characteristic waves of ECG are not only QRS complexes, but also P and T wave. So in order to improve QRS complexes detection rate and search for a new method for P and T wave detection. We proposed a new method for ECG characteristic points detection based on complex wavelet transform. It is different form the traditional wavelet transform for the wavelet basic function is a complex wavelet function. And we used the MIT/BIH arrhythmia database records to evaluate the proposed method. We found that it's superior than other known methods and the detection rate of QRS complexes is 99.85%. After QRS complexes detection, we proposed a detection method for the onset and offset of QRS complexes and P and T wave. Experiment results verified the efficiency of the proposed method.In order to deepen the study, we discuss the future research work in the final.