A Nonstationary Sparsity Deoncolution Method and Its Application

Deconvolution is important and recurrent in seismic processing. In fact, it is well acknowledged that traditional statistical deconvolution is based on the assumption of white reflectivity and minimum-phase wavelet while practical researches indicate that wavelet is mixed-phase and reflectivity series rarely behaves according to Gaussian distribution. As an effective way to get rid of these assumptions, sparsity deconvolution estimate reflectivity by employing the constraint of Cauchy distribution. However, as a phenomenon that caused signal distortion in seismic exploration, nonstationarity is frequently overlooked. In this paper, to accurately estimate reflectivity, we initially use an improved Gabor deconvolution method to eliminate nonstationary, and then utilize the lateral constrained sparsity deconvolution method to estimate reflectivity. Ultimately, the application on carbonate reservoir shows more accurate and detailed information, successfully demonstrate the effectiveness and robustness.