Abstract: The generalized S-transform (GST) applied to seismic data interpretation is achieved by means of replacing the basic wavelet in S-transform (ST) by the seismic wavelet extracted from seismic data. It is shown from the test of thin layer model that the thin layer greater than λ/8 can be detected by GST with the super time-frequency localization spectrum of ST. Moreover, the self-adapted time- frequency window of GST and the arbitrary input data length allow to deal with the limitations of discrete Fourier transform time window, and the reservoir imaging by energy spectrum can be realized based on the spectrum decomposition theory of Fourier transformation with good results by comparison between the two. This method has been applied to many areas such as Tarim, Junggar and Indonesia, successfully.

Key words: S-transform, generalized ST, spectrum decomposition, reservoir imaging