Many multi-channel seismic algorithms assume that the input seismic data have a regular trace spacing. Unfortunately, this is not always so because of incomplete or noisy acquisition. This leads to problems with procedures such as DMO and prestack migration. In this paper, a method based on the least-squares f-x prediction filter is de-veloped to interpolate the missing traces to regularize the data set. This filter is applied to a synthetic data set and a field data set from Pine Creek, Alberta. It is shown in syn-thetic test that the filter is successfully predicted the conflicted dips and interpolated as much as one-third of the data. The result is also shown to be superior than the conven-tional f-x deconvolution. The field example shows that the missing trace interpolation can lead to considerable improvement in the final processed results.
View full article as PDF (0.96 Mb)