Research On Extraction Of ECG Signal Characteristic Parameters Based On Wavelet Transform

At present,the death of cardiovascular disease accounts for the first reason of the total death of urban and rural residents in China.With the aging of the population and the acceleration of urbanization process,the number of cardiovascular disease will still maintain a rapid growth in the next 10 years.The traditional ECG detection equipment is huge,complex,and needs professional diagnosis.With the development of wearable devices and automatic analysis and diagnosis technology,the research of ECG signal extraction is increasing.In this background,based on wavelet transform,this paper studies the feature parameter extraction algorithm of ECG signal,and applies it to FPGA platform.The main work of this paper is as follows:Firstly,in the aspect of ECG denoising,the Soft-threshold algorithm of wavelet threshold denoising is used to process ECG signal.In the paper,50 Hz sine wave is used to simulate the power frequency interference noise,and 0.5 Hz sine wave is used to simulate the baseline drift to establish the noise model.Then by comparing the characteristics of each wavelet,DB4 wavelet is selected to decompose the noisy ECG signal with 8-level wavelet transform,which realizes the noise reduction filtering of ECG signal,and finally evaluates the noise reduction effect.Secondly,in the aspect of ECG feature parameter extraction,the ECG R-wave detection system is designed based on wavelet transform and adaptive double threshold algorithm.The detection system is divided into two parts: QRS wave extraction module and QRS wave determination module.Firstly,the effects of different schemes are compared,and bior4.4 wavelet is used to decompose ECG signals in 5-level.Then,the Sliding-Window integral algorithm and the adaptive double threshold algorithm are designed.The self-adaptive double threshold makes full use of the advantages of FPGA,such as high integration and fast operation speed,and realizes more accurate QRS real-time detection by using high and low double thresholds.Finally,the whole system is modeled and simulated by MATLAB,and 48 groups of ECG data in MIT-BIH ECG database are used to get the detection rate of QRS wave for normal ECG signal is close to 100%,and the overall detection rate is 98.93%,which meets the design requirements.Thirdly,based on 4ce115f23i7 core chip of cyclone IV series,the QRS real-time detection system is implemented on FPGA.Firstly,the whole QRS detection system ismodularized,which is divided into wavelet transform module,QRS detection module and data transmission module.The main content is to use bior4.4 wavelet as the basis function to transform ECG signal into 5-level wavelet;then the QRS wave extracted is processed into a single wave peak signal,and it is judged;finally,the data is sent to the upper computer by UART transmission module.The whole system is written by Verilog Hardware Description Language,and compiled and synthesized by Quartus II software.Finally,the test system is simulated and tested.The timing simulation of Modelsim is carried out,and the board test and signal analysis by SignalTap are completed.The result is consistent with MATALB simulation.The detection algorithm designed in this paper can be applied to FPGA hardware platform to realize QRS real-time detection.