INTRODUCTION

At present, the majority of ground penetrating radar (GPR) work involves the collection, processing, and interpretation of constant-offset data. However, the increased availability of multi-channel GPR systems is making multi-offset data collection increasingly popular. Advantages of working with multi-offset GPR data include the improved estimation of subsurface velocities and imaging of reflectors, and also the ability to perform amplitude versus offset (AVO) or angle (AVA) analysis Baker (1998); Fisher et al. (1992a); Greaves et al. (1996). These advantages allow for better estimation of sedimentary facies and subsurface properties.

Because of the many similarities between the GPR and seismic methods, numerous exploration seismic imaging techniques have been directly transferred to the radar community. One family of imaging methods commonly used in seismic exploration, especially in areas of complex geology, is based on wave-equation wavefield extrapolation. Methods in this family employ a one-way scalar wave equation to extrapolate a wavefield recorded at the surface through a subsurface velocity model. In theory, this procedure generates the wavefield that would have been recorded had the instruments been located at a surface deeper within the earth. Images of geologic structure are then constructed through the evaluation of a physical imaging condition at each subsurface model point.

Here, we consider the application of shot-profile migration, a prestack imaging algorithm belonging to the family of scalar wave-equation methods, to 2-D multi-offset GPR data. This method is similar to the zero-offset, survey-sinking migration method proposed by Sena et al. (2003), which involves using split-step Fourier operators to extrapolate GPR wavefields. However, the algorithm presented here is designed for application to multi-offset data in a shot-profile configuration, and allows for the easy formation of angle-dependent images that are suitable for migration velocity analysis (MVA) and AVA studies.