Seismic spectral decomposition using short-time fractional Fourier transform spectrograms

Abstract

A time-frequency method with high resolution is introduced for seismic spectral decomposition. The classical short-time Fourier transform spectrogram has low resolution, limiting its usefulness in application such as stratigraphic visualization and gas reservoir detection for thin beds. This work expands the traditional short-time Fourier transform spectrogram to the fractional domain, leading to the proposed short-time fractional Fourier transform spectrogram (FrFTSG). The short-time Fourier transform spectrogram can be calculated using the 2-D Fourier transform of the product between the ambiguity functions (AF) of the signal and the window function, and using the fractional Fourier transform (FrFT) as the rotation operator of the window function, the window function could be rotated to the direction with the auto-term of the signal in the ambiguity domain. The FrFTSG could be obtained by the 2-D Fourier transform of the product between the ambiguity functions of the rotated window function and the signal. Analyses of synthetic test model and actual seismic data show that the method proposed in this study can characterize the signal with a higher accuracy. This allows more accurate characterization of reservoir location. The method proposed in this paper provides a novel technique for seismic spectral decomposition.