Interpolation is used to improve the spatial coverage of the data in cases where irregularities exist due to edits and coverage holes from acquisition or obstructions. It is performed in common-offset cubes, utilizing multidimensional information: line, CMP, time, offset and azimuth, whereby azimuth information is preserved. This method is based on the work of Hindriks (2000) and is implemented as an inversion problem in the F-Kx-Ky domain. The unknown is a vector containing the data of the new trace locations, which will be solved from the data vector containing the original data. The regularization can be implemented in many domains, but to preserve the azimuth information, we select to implement it in the shot gather domain.
Hindriks, K., and A. J. W. Duijndam, 2000, Reconstruction of 3-D seismic signals irregularly sampled along two spatial coordinates: Geophysics, 65, 253–263.