Time-varying residual phase estimation and correction in seismic data

The residual phase estimation and correction of a post-stack seismic section is important and necessary. The remaining non-stationary phase detection of data without the use of well logs information can be made using statistical methods.
Kurtosis maximization approach by constant-phase rotation is the most popular post-stack statistical methods that can reveal the non-stationary phase of seismic data. Kurtosis criterion is a fourth-order statistics that preserves the wavelet phase information and plays an important role in seismic data interpretation. In this paper, we change the problem of regularized kurtosis maximization for non-stationary phase estimation that is presented by van der Baan and Fomel (2009) to reduce significantly the computational volume while maintaining the quality of our results. The proposed approach due to lower computational volume, and also, the fewer number of free parameters is more efficient than similar regularization approaches. Therefore, for large-scale data analysis, it is easier to use. The effectiveness of the proposed technique for identifying and correcting the non-stationary residual phase of the seismic signals are shown on both synthetic and field data.