Attenuating Seismic Random Noise by Multichannel Singular Spectrum Analysis

Abstract

Suppression of random noise can significantly improve the quality of seismic imaging. In this study we used a method, which reduces multichannel random noise and time of computations while protecting seismic structures. The method is based on multichannel singular spectrum analysis (MSSA). Usually attenuating multichannel random noise via rank reduction comes up with large block Hankel matrices and a large amount of computations. We applied a randomized singular value decomposition (RSVD) method (Oropeza and Sacchi, 2010) to estimate rank reduced matrices. We evaluated the performance of the method on synthetic and real seismic data.