Compressive sensing principles applied for seismic data acquisition and processing in time and space

Abstract

Summary Compressive sensing (CS) principles provide new perspectives for non-uniform surface sampling that significantly reduces acquisition cost and cycle time compared to current conventional practices in seismic exploration. CS in time i.e., blended acquisition of seismic data can be achieved by the combination of simultaneous sources and continuous recording. CS in space can be considered during seismic acquisition in which non-uniform undersampling of source and/or receiver locations are utilized such that coherent aliases are depicted as incoherent random noise. In this paper, we first propose a spatial compression scheme based on the CS theory, in which a non-uniform undersampled survey geometry is obtained by minimizing the mutual coherence (MC) of a sampling operator. Moreover, considering a source blending scheme, we present a joint deblending and wavefield reconstruction by sparsity-promoting inversion algorithm. Examples using real seismic data indicate the ability of obtaining high resolution and accurate images from solutions given by the joint deblending and wavefield reconstruction.