ABSTRACTTube wave suppression and the separation of elastic up-going and down-going waves are the most important steps in vertical seismic profile (VSP) data processing. Different methods have been developed to suppress tube waves and separate wave fields. In this work, we use a pyramidal dual-tree directional filter bank as a multi-directional and multi-scale method. We then design two different adaptive algorithms to suppress tube waves and aliased tube waves, and to separate elastic up-going and down-going waves. In the first proposed algorithm, subscales containing tube waves and aliased tube waves were detected adaptively through energy calculation in the 2D Fourier domain, through which the filtered VSP were achieved by zeroing these subscales. When applied to field VSP data with strong tube waves, the proposed filter can attenuate both the tube wave and aliased tube wave successfully, with results outperforming those of f–k filtering, singular value decomposition and median filtering. Similarly, following tube wave and aliased tube wave suppression in the second designed algorithm, subscales with elastic up-going waves were detected fully. By reconstructing these subscales and subtracting them from the original VSP data, the wave fields were separated successfully. The performance of both algorithms in the presence of random noise was investigated and none of the algorithms were found to be highly sensitive.
Read full abstract