Application of compressive sensing to seismic data reconstruction

There is a contradiction between seismic data processing and remote transmission of field seismic data, which is their different requirements about the data volume. More precisely, the former requires that the data quantity be as large as possible, while the latter requires the opposite. In order to solve this problem, a new method, compressed sensing (CS) technology, based on curvelet transform, is introduced into seismic data processing to reconstruct the actual incomplete data. The results show the advantage of curvelet transform compared with FFT in CS method. However, when it is used to actual data processing, surface wave must be removed first and the effect of the number of missing traces should be considered simultaneously. Finally, a reasonable reconstructed result is achieved, with clear texture and natural connection, illustrating the applicability and validity of the CS method.