A New Seismic Data De-Noising Method Based on Wavelet Transform

A main task of geophysical exploration is to remove random noises in seismic data processing to improve the signal-to-noise ratio. Recently wavelet theory is applied widely to remove random noises in seismic data processing. But conventional wavelet threshold de-noising method does not utilize the correlations of seismic data to remove random noises. So a new de-noising method is proposed in this paper. This new de-noising method combines time-frequency correlation analysis with threshold filter in wavelet domain. The paper explains in much detail how to use time-frequency correlation analysis to analyze correlations of seismic data, i.e., to analyze wavelet coefficients of multi-scales; after correlation analysis, these wavelet coefficients are reconstructed; in this way, most random noises can be removed. Then conventional wavelet threshold de-noising method is used to remove more noises. The results of theoretical model and practical data processing show that the method presented by the paper can remove most random noises and effectively improve S/N ratio of seismic data.