Interpolating GPR data using anti-alias singular spectrum analysis (SSA) method
Ground Penetrating Radar data are often acquired along profiles employing bistatic equipment with a fixed distance between the transmitter (Tx) and receiver (Rx) antennae. Even in cases where more than two antennae are used, the number of channels tends to be relatively small, resulting in either a limited number of offsets or gathers with inadequate far offsets. Estimating stacking velocity and performing migration from this type of datasets are difficult. In this paper, we present techniques to interpolate both aliased and non-aliased datasets in the offset domain and the common- midpoint domain. The latter permits us to increase the fold of the survey and consequently improve the process of velocity analysis and migration. We assess the reconstruction efficiency of the interpolator using both synthetic and real data to different degrees of decimating. In both cases, the unaliased version of both datasets provides an accurate solution for a careful comparative analysis. At the end of this work, we make a further comparison between the resulting migrated and stacked sections for both the original and reconstructed datasets in order to highlight the efficiency of the interpolation algorithms.