Enhancing the resolution of non-stationary seismic data using improved time–frequency spectral modelling
Open Access
- 16 February 2016
- journal article
- research article
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 205 (1) , 203-219
Abstract
Maximizing vertical resolution is a key objective in seismic data processing. Early deconvolution and spectral balancing algorithms assumed that the seismic source wavelet was temporally invariant, or stationary. In practice, seismic scattering and attenuation give rise to non-stationary seismic source wavelets. To address this issue, most conventional time-varying deconvolution wavelet shaping and spectral modelling techniques using the stationary polynomial fitting assume the wavelet to be locally stationary within a small number of overlapping analysis windows while the fitting coefficients are invariant with all the frequencies. In this paper, we show an improvement obtained by modelling smoothly varying spectra of the seismic wavelet using non-stationary polynomial fitting in the time–frequency domain. We first decompose each seismic trace using a generalized S-transform that provides a good time–frequency distribution for the estimation of the time-varying wavelet spectra. We then model the slowly varying source wavelet spectrum at each time sample by a smooth low-order polynomial. Finally, we spectrally balance the modelled wavelet to flatten the seismic response, thereby increasing vertical resolution. We calibrate the algorithm on a simple synthetic and then apply it to a 3-D land survey acquired in western China, showing the value on both vertical slices through seismic amplitude and attribute time slices. Our new algorithm significantly improves the vertical resolution of the seismic signal, while not increasing the noise.Keywords
This publication has 47 references indexed in Scilit:
- Seismic data decomposition into spectral components using regularized nonstationary autoregressionGeophysics, 2013
- Seislet transform and seislet frameGeophysics, 2010
- Adaptive multiple subtraction using regularized nonstationary regressionGeophysics, 2008
- Shaping regularization in geophysical-estimation problemsGeophysics, 2007
- Short-time fourier transform: two fundamental properties and an optimal implementationIEEE Transactions on Signal Processing, 2003
- Instantaneous spectral analysis: Detection of low-frequency shadows associated with hydrocarbonsThe Leading Edge, 2003
- Frequency-time decomposition of seismic data using wavelet-based methodsGeophysics, 1995
- Similarity and the inverse Q filter; some simple algorithms for inverse Q filteringGeophysics, 1992
- Predictive deconvolution and the zero-phase sourceGeophysics, 1984
- Dispersive body wavesJournal of Geophysical Research, 1962