On the denoising method of prestack seismic data in wavelet domain

In this paper, the modeling of the random noise in prestack seismic data is discussed. The statistic character of the local wavelet power spectrum of the random noise is studied by qualitative analysis and Monte Carlo simulation using physical wavelet. A conclusion is drawn that the distribution of the local wavelet power spectrum of the noise is χ₂². On the basis of the conclusion, a method of confirming the space of the useful signal energy distribution in wavelet domain is given and the noise can be restrained by reconstruction of the signal from the space. The validity of the method is proved by model and field seismic data.