随着复杂断块、岩性等隐蔽油气藏勘探的不断深入,保幅型非稳态的提高分辨率地震处理技术受到人们的关注。在传统谱模拟反褶积的基础上,提出了一种基于S变换的时变谱模拟反褶积方法。在S变换域通过最小二乘拟合、低通滤波及多道统计加权来获得适应地震记录非平稳特征的时变子波振幅谱,并在反褶积算子设计时考虑信噪比的影响,进而实现时变谱模拟反褶积处理。模型试算和实际资料处理验证了方法的正确性与有效性。
With the advance of subtle reservoirs exploration,such as complex fault block reservoir and lithologic hydrocarbon reservoir,amplitude-preserved and non-stationary processing technologies that enhance seismic resolution have been focused.Based on the conventional spectral modeling deconvolution,time-varying spectral modeling deconvolution on S transform is presented.In S transform domain,we obtained the time-varying spectral of wavelet amplitude by the least squares fitting,low-pass filtering and multi-channel statistical weighting,which adapts to non-stationary seismic records.Take the effect of SNR into consideration,we designed deconvolution operator and realized time-varying spectral modeling deconvolution.Model test and the actual seismic data processing proved the correctness and effectiveness of the method.