According to the request of automatic analysis and depressing high frequency interference of the ECG signals, this paper applies low-pass filter to preprocess ECG signals, and proposes a QRS complex detection method based on wavelet transform, which takes advantage of Marr wavelet to decompose and filter the ECG signals with Mallat algorithm, using the relationship between wavelet transform and signal singularity to detect QRS complex with amplitude threshold method in scale 3, and to detect P wave and R wave in scale 4. Meanwhile, compositive detection method is used for re-detection, thus to improving the detection accuracy ratio. At last, records from ECG database of MIT/BIH which is widely accepted in the world are used to test the algorithm. And the result shows that correction detecting ratio under this algorithm has been more than 99.8 percent. The detection method in this paper is simple and running fast, and is easy to be realized in the real-time detecting system using for clinical diagnosis.
[1]
W.J. Tompkins,et al.
Neural-network-based adaptive matched filtering for QRS detection
,
1992,
IEEE Transactions on Biomedical Engineering.
[2]
P.E. Trahanias,et al.
An approach to QRS complex detection using mathematical morphology
,
1993,
IEEE Transactions on Biomedical Engineering.
[3]
C. Li,et al.
Detection of ECG characteristic points using wavelet transforms.
,
1995,
IEEE transactions on bio-medical engineering.
[4]
Huang Yuxi,et al.
Using Biorthogonal Wavelet Filter Bank for ECG Detection and Reconstruction
,
2001
.
[5]
Stéphane Mallat,et al.
Singularity detection and processing with wavelets
,
1992,
IEEE Trans. Inf. Theory.
[6]
J. S. Sahambi,et al.
Using Wavelet Transforms for ECG Characterization
,
1997
.