Filtered dense h picking for non-hyperbolic moveout

Abstract

The focusing quality of far-offset volumes is strongly dependent on the density of the Non-Hyperbolic Normal Move Out (NHNMO) parameter estimation. The automatic picking of the residual time-shift due to the non-hyperbolic behavior of the reflection curve attempts to achieve this goal. The anellipticity h field containing many errors related to mis-picks and artifacts is replaced by anelliptic velocity, which is related to the asymptote of the reflection curve. This new NHNMO parameter field, thanks to the short spread curvature weighting, suggests a more reliable attribute for geostatistical filtering. Standard geostatistical techniques such as factorial kriging are successfully applied to perform the filtering of the anelliptic velocity as well as the filtering of NMO velocities. The real data example shows that filtering of anelliptic velocity instead of anellipticity h itself increases the reliability of effective h field, therefore improving the focusing of the far-offset data.