
Full text loading...
The resolution of the seismic signal can be remarkably enhanced by nonstationary deconvolution because it considers the attenuation of the source signature travelling in the strata. But traditional nonstationary deconvolution extracts wavelet by smoothing the Gabor spectrum dependant of supposing the reflectivity is white.In practice, the reflectivity is nonwhite. This paper deduced the relationship of the wavelet and reflectivity series with the seismic signal in the logarithmic time-frequency spectrum whether reflectivity is white or not. And this relationship is used to obtain the wavelet from the seismic signal in the logarithmic time-frequency spectrum in the nonstationary deconvolution. After testing on the model data and practical data, the method has been proved that it can obtain the wavelet from seismic signal better than the traditional nonstationary deconvolution. Then the new method is applied to the 3-dimension seismic data and obtained more information about the reservoir predictions.