Improved generalized S-transform deconvolution for non-stationary seismic data

Improving the vertical resolution is one of significant tasks for seismic data processing. Most traditional resolution-enhancement techniques assume that wavelet time-invariant. However, varies with wave propagation in subsurface. To solve this issue, a new spectral-modeling method proposed to extract time-varying using improved generalized S-transform (IGST) and higher-order Fourier series. The IGST based on modified time-window function can effectively improve time-frequency (t-f) spectrum. high-order series used fit logarithm t-f spectrum achieve high-precision wavelet. composed four steps implementation. Firstly, decomposed by converted domain. Secondly, modeled at each time sample Thirdly, boxcar smoothing smooth Hilbert transform. Finally, spectrally balance flatten response. Synthetic field examples demonstrate feasibility improving signal-to-noise ratio enhancing resolution.