摘要
By analyzing the advantages and disadvantages of Curvelet transform in seismic processing,we propose a non-dyadic Curvelet transform for seismic noise elimination.There are two improvements:(1) The non-adaptive dyadic scale and angular decomposition has been updated to an adaptive non-dyadic controllable scale and angular decomposition,which provides a better representation of the primary characteristics of seismic data in terms of orientation and frequency.(2) The anisotropic analysis is introduced to the lowermost scale to improve the accuracy of low-frequency components processing.The synthetic and field data examples are used to validate the effectiveness of the proposed data analysis method.