Study On The Pretreatment And Waveform Detection Algorithm Of ECG

Cardiovascular disease is one of the main killers to human health, and the ECG diagnosis an important method of diagnosing cardiovascular diseases, so analyzing and diagnosing ECG signal accurately plays a key role in treating the cardiovascular disease. How to accurately and rapidly process and analyze the ECG signal has become one of the hot research topics in current signal processing field, whose study results will promote the development of health care and raise the level of human health effectively.ECG processing and analysis have a wide range of details. This paper mainly focuses on the algorithm study of its two key technologies, i.e. ECG signal pretreatment and ECG waveforms detection and features location.The algorithm study of ECG signal pretreatment: ECG is weak and it's extremely vulnerable to environmental impacts. So, if accurate detection and analysis are wanted, it is necessary to pre-process the signal to the most common interferences in ECG: baseline drift, the frequency interference and EMG interference. The paper puts emphasis on the study of the notch filter to eliminate 50Hz power line. Firstly, the digital filtering techniques are used to carry out ECG pre-treatment, and the designs of notch filters are realized based on IIR and FIR, which eliminate the ECG power-line interference effectively. Taken into account that cut-off frequency of notch filters based on IIR and FIR is fixed, the notch-filter can't play its role when the frequency of power line fluctuats above or below, so the paper also studies on the notch filters based on adaptive algorithm and wavelet-transform threshold algorithm respectively, achieving better results than IIR and FIR notch filters. From two aspects of computing speed and processing accuracy, the simulation results of the four algorithms are compared and analyzed.The algorithm study of ECG waveform detection and characteristics location: waveform detection and characteristics location is the key to ECG diagnosis, which directly impact the accuracy and reliability of diagnosis. QRS wave is the main features of ECG, which is the primary problem of ECG detection, and P wave and T wave detection is of great significance in accurate ECG analysis as well. From two aspects of time domain and transform domain, QRS wave, P wave and T wave of the major ECG features are detected and located in the paper. Firstly, the detection of QRS wave is achieved based on peak time-domain positioning algorithm and wavelet-transform maximum modulus algorithm respectively. Then, according to the detected QRS feature points, combining the relationship of amplitude and slope of the P and T wave characteristics and combining the corresponding relationship between P&T wave and their wavelet decomposition extreme values, P wave and T wave are detected, using the differential threshold algorithm and the wavelet maximum modulus algorithm. Besides, many ECG data units of the standard ECG database are detection simulated and effect assessed.All of the algorithms are simulated on Matlab tool, and the paper has 34 figures, 6 tables and 41 reference articles.