The Design And Implementation Of ECG Pretreatment System Based On FPGA

Heart disease is a high risk and strong sudden disease, and has become one of the maindiseases which threatens human health. The precise processing and analysis of ECG signalhave important meanings to improve human’s life quality. However ECG signal issusceptible to power frequency interference, muscle power interference and baseline drift.These noises have seriously affected the accurate identify of ECG signal characteristicinformation, and can reduce the diagnosis reliability of heart disease. To ensure the smoothprogress of the analysis and diagnosis of heart disease, the pretreatment of ECG signal isvery necessary, which can restrict related noises in a greatest degree.As the good time-frequency properties, wavelet transform is widely used in ECGsignal pretreatment, and can reduce all kinds of noises effectively. At present, ECG signalpretreatment algorithm based on wavelet transform is mostly implemented by software. Butthe complexity of wavelet transform algorithm is high, which could cause the real-timecapability of software realization of pretreatment is poor. With the large single chipcapacity, higher parallel operation ability, FPGA can provide the possibility for realizingcomplex ECG signal pretreatment algorithm on hardware. At the same time, the realizationof ECG signal processing and analysis hardware algorithm falls in with popular increasingdemand of "home for medical treatment, individual health care" and the ECG monitoringproducts’ development tendency of portability, muti_function, low power consumption.This paper discusses the characteristics of ECG signal and noises deeply, and adoptswavelet threshold denoising theory, using the wavelet lifting features of simple structure andeasy implementation on hardware, puts forward a kind of ECG signal pretreatment hardwarealgorithm based on wavelet transform, and finally complete the FPGA structure design ofthe algorithm. This algorithm overcomes the shortcomings of bad real-time from generalcomputer realize way. The wavelet functions in past wavelet threshold denoising hardwarealgorithms are not appropriate in ECG denoising. This paper adopts db4wavelet, whoseshape is similar with ECG signal, wavelet function is approximation symmetry, supportlength and disappear moment are appropriate, can make the high frequency coefficientsafter decomposition smaller, and can avoid distortion during reconstruction, and can distinguish signal and noises better. Soft threshold processing algorithm way could cause thedeviation between estimated value and signal. This paper uses weighted threshold treatmentway based on soft threshold, which can retain the useful information of P、T waves in agreatest degree.This paper designs ECG signal pretreatment FPGA realizing system based on waveletlifting transform using MATLAB, DSP Builder, QuartusII and VHDL. The system mainlyincludes five modules: ECG signal producing module, lifting wavelet decompositionmodule, coefficient processing module, lifting wavelet reconstruction module and timesequence control module. The paper simulates the function of each module in ECG signalpretreatment system based on CycloneII EP2C35F672C8core chip, and completes the testand verify the performance of the system. Experimental results show that the design of ECGsignal pretreatment system based on FPGA can get good pretreatment effect, and could meetthe real-time ECG signal pretreatment requirement. The application prospect of the systemis extensive.