Applications of Non-Rigid Matching to 3-D Converted-Wave Imaging
Tony D. Johns
Processing limitations in the time domain to accurately ray trace, or meticulously model, the full-waveform expression of converted-wave moveout in the presence of lateral heterogeneity and polar anisotropy, often culminate in imaging discrepancies between different azimuth sectors of a 3-D converted wave (P-Sv) seismic data volume. To compensate for lateral or temporal divergence of converted-wave imaging as a function of azimuth, a two-tiered workflow for applying a non-rigid matching (NRM) algorithm is applied to combine two distinct 3-D azimuth sectors of a P-Sv pre-stack time migration (PrSTM) dataset to form a final enhanced 3-D P-Sv volume with superior stack response and continuity. The method allows for the crossline artifacts from the effect of azimuthal anisotropy on the converted wave moveout to be almost completely removed. Furthermore, the severe acquisition footprint, from insufficient crossline aperture as a result of a sub-optimum survey design, is effectively mitigated. Seismic data examples of 3-D inlines, crosslines and time slices taken from a typical 3-D/4C survey acquired from the Gulf of Mexico and processed in 2006, demonstrate the compelling benefits of the NRM application and the robustness of the developed workflow. Furthermore, the output NRM displacement attributes (voxel time-shift values) are found to possess a qualitative value which are not only powerful indicators of azimuthal anisotropy, but also through calibration, may yield valuable information on the magnitude of shear splitting and principal directions of polarization.
AAPG Search and Discovery Article #90077©2008 GEO 2008 Middle East Conference and Exhibition, Manama, Bahrain