Target-oriented beamlet migration based on Gabor-Daubechies frame decomposition

We develop beamlet propagation and imaging using Gabor-Daubechies (G-D) frame decomposition based on local perturbation theory and apply it to target-oriented prestack depth migration. The method is formulated with local background velocities and local perturbations in wavefield extrapolation. The localized propagators and phase-correction operators are obtained analytically or semianalytically by one-way operator decomposition and screen approximation in the beamlet and space-beamlet mixed domain. Beamlet wavefields have superior localization properties in both local space and direction (wavenumber) over Gaussian beams in the sense that localizations are not limited within short propagation distances in either homogeneous or heterogeneous media. Comparisons of the prestack depth-migrated images for the 2D SEG-EAGE salt model and the Marmousi model indicate that, for seismic-wave propagation and imaging in complex structural environments, the G-D beamlet propagator has higher accuracy and better wide-angle properties than do global propagators. Target-oriented prestack G-D beamlet migration is performed by means of local-angledomain imaging and controlled superposition of commondip-angle images based on the local directivity features of the target structures. This considers the spatial and direction localizations of beamlets. As a numerical example, we process the 2D SEG-EAGE salt-model prestack data. The results show that the proposed migration method has considerable advantages in suppressing noise and enhancing structural features. Image quality for subsalt structures, especially for steep faults, is improved through structure-based superposition of common-angle images. This demonstrates the potential and capability of beamlet migration in target-oriented seismic imaging.