The Study On Signal De-noising Algorithm Based On Wavelet Threshold

The wavelet transform theory is the crystallization of the common struggle ofmathematics, physics and other disciplines. Wavelet transform theory and analysis oftoday’s most powerful tools are widely used in various fields. In signal processing,because of its good localization characteristics in time and frequency domain,allowing for more detailed analysis on signal. Whether signal detection, estimation,identification, and other signal processing is required in the context of signalde-noising. Therefore, signal de-noising algorithm is especially important.This article on wavelet threshold algorithm and its application in signal de-noisingwas studied and explored. Wavelet basic theory, the statistical characteristics ofnon-Gaussian noise and principles and Wavelet threshold de-noising thresholdequations were first introduced in the paper. For non-Gaussian noise thresholdde-noising threshold equation of inadequacies in improvement of threshold equationis improved. With selection of wavelet and threshold in MATLAB and carried out in adifferent threshold function de-noising effect analysis and comparisons.Improved threshold equation is then applied to the weak life signal-de-noising ofECG. Faced with power frequency interference, muscle artifact and baseline wanderintegration of complex Gaussian noise, the simulation results show that improvedthreshold de-noising slightly inadequate in terms of removing baseline wander.Therefore, the lifting Wavelet threshold de-noising algorithm which combines liftingWavelet transform and threshold functions was introduced. Signal-to-noise ratio gainin de-noising process as a basis for selection of wavelet basis function, combined withthe Brige-Massart policy based on Adaptive threshold processing of ECG. Thensimulation results were analysed to draw conclusions.Finally, re-apply it in non-Gaussian noise background signal de-noising resultsshow that lifting wavelet threshold algorithm in the composite Gaussian noise and a class of non-Gaussian noise de-noising algorithm can obtain better de-noising effectsignal.