AN EFFICIENT ATTENUATION COMPENSATION METHOD USING THE SYNCHROSQUEEZING TRANSFORM

摘要

Attenuation of seismic waves, which decreases the amplitude and distorts the phase, also usually results in low resolution of seismic data. In this study, inverse Q filtering is introduced to compensate the attenuation of seismic waves using the synchrosqueezing transform (SST), which condenses the spectrum energy along the frequency axis and provides highly localized time-frequency representations. To perform a stable inverse Q filtering, a denoising filtering and reliable Q values are needed. At first, a spare SST-domain filtering is utilized to remove the random noise based on a low-rank and spare decomposition-based method. Then, a reliable Q-factor estimation method using the peak frequency shift method in SST domain is applied in the implementation of inverse Q filtering scheme. At last, we reformulate amplitude correction as an inverse problem with a l(2)-norm regularization term in the SST domain to prevent the noise bursting because of the inherently unstable process of amplitude correction. Therefore, we propose a complete three-stage work flow to denoise, estimate Q factor and compensate attenuation using the SST. Tests with synthetic examples and real data demonstrate the robustness and effectiveness of this proposed method.