Kinematic wavefield attributes based multidimensional prestack data reconstruction

Abstract

The 4D/5D interpolation and regularization methods effectively improve the quality of seismic imaging. In addition to the Fourier domain interpolation method, 5D interpolation based on the common reflection surface (CRS) method has attracted more and more attention due to simplicity of its implementation and effectiveness of performance. However, the main challenge of this method is the heavy calculation in parameter estimation. To overcome this limitation, we introduce a kinematic wavefield attributes based prestack data interpolation and regularization method. This method uses gradient structure tensor (GST) and quadratic structure tensor (QST) methods to extract kinematic wavefield attributes (local slopes and curvatures) and use them for fast 3D zero-offset (ZO) CRS parameter estimation. The derived parameters are then used for 3D CRS based prestack interpolation and regularization. The proof of concept is demonstrated on datasets acquired by TopSeis. The corresponding results show that the improved efficiency of the GST/QST based method in kinematic wavefield attribute extraction and 3D ZO CRS parameter estimation. Moreover, the interpolated and regularized TopSeis prestack data derived from the subsequent 3D ZO CRS proves the simplicity and effectiveness of this method.