A seismic data processing method based on modular maximum of wavelet transforms

The singularity indexes of signal and noise are analyzed. According to the theory that modular maximum values of wavelet transforms of the functions with different singularity indexes have different spreading characteristics, local maximum values of noise are analyzed from adjacent modular minimum and the smothing is conducted over the pass-regions, at which local maximum of noise is located. Then noise-extracted signal is reconstructed by inverse transform. This method is used to extract noise from sinusoidal signal, synthetical seismogram and oilfield seismic data respectively. The result shows it is simple and effective.