Fourier methods of seismic data regularisation

Abstract:

Acquisition geometries of seismic data invariably lead to an irregular sampling in space of midpoint, offset and azimuth, however many seismic processing theories and algorithms work best if not exclusively on regularly sampled data. Here, three methods of Fourier regularisation are implemented and contrasted each with their own benefits and drawbacks. Standard theory is extended to the use of discrete spheroidal prolate sequence multitapers to help to diminish spectral contamination caused by an irregular input grid. Early results show that the use of multitapers alone have not yet proved to adequately reconstruct the input data, however the use of these to augment other types of Fourier regularisation have shown promise in increasing accuracy and reducing the number of required iterations.