Second-order time-squeezing transform for seismic time-frequency analysis

Abstract

Spectral decomposition is a cornerstone in seismic signal processing and interpretation. As a novel decomposition method, synchrosqueezing transform (SST) is attractive for high-resolution time-frequency analysis of seismic signals. To highlight geological structures with high precision, a second-order time-squeezing transform (STST) is proposed by reassigning the short-time Fourier transform (STFT) coefficients along the time axis rather than along the frequency axis as in STFT-based SST (FSST). This method deduces the relationship between the signal group delay and the STFT coefficients from a second-order frequency-domain signal model. Then, a real-valued filter is constructed and convolved with the STFT coefficients to establish a strong correlation between the time-frequency information, and a novel group delay estimator is further derived. Finally, the highly robust and stable time-frequency energy-concentration of seismic signals is achieved by reassigning the STFT coefficients. A simulated seismic trace and field data are employed to verify the effectiveness of the proposed method. It is concluded that it has great potential as a time-frequency analysis technique for seismic signal processing.