Application of Time-frequency Attributes Based on Generalized S-transform for Thin Bed Indication

During the last decade, spectral decomposition technique has proven to be an excellent tool to describe thin beds associated with channel sands, alluvial fans, and the like. However, the traditional spectral decomposition method based on short time Fourier transform, is difficult to acquire the accurate time-frequency spectrum for non-stationary seismic signals. In addition, it generates dozens of unwieldy iso-frequency cubes, not amenable for interpretation. We propose two new time-frequency attributes based on generalized S-transform and apply them to an oilfield of western China to characterize the Triassic reservoir distribution of fluvial deposits. Compared with the traditional short time Fourier transform, the generalized S transform results in more accurate time-frequency spectrum, and generate two attributes which can directly map reservoir thickness.