Time-frequency analysis of seismic data for the characterization of geologic structures via synchroextracting transform

Time-frequency (TF) analysis based method is one of the most powerful tools in seismic data processing and interpretation. To delineate the subsurface geological structures clearly, the energy concentration of the seismic TF representation (TFR) is as high as possible. The traditional TF methods, such as the short-time Fourier transform (STFT) and wavelet transform (WT), have been widely applied for seismic TF analysis. However, they suffer from the diffused TFR because of the Heisenberg uncertainty principle. To achieve a higher energy-concentrated TF result, the synchrosqueezing transform (SST) was proposed. The SST method has been successfully used for seismic processing. In this paper, we introduce a novel approach for seismic spectrum analysis based on the synchroextracting transform (SET), which is an extension of the Fourier-based SST (FSST). The SET extracts the TF information only located at the instantaneous frequency (IF) trajectory of the analyzed signal and removes the most smeared TF energy, which leads to a highly sharpened TFR. We also put forward a theorem to prove that the SET can get the exact IF of a linear chirp signal. All the results of the synthetic signal and real seismic data demonstrate the validity and effectiveness of the proposed method.