Optimum focusing from filtered dense velocities picking

Abstract

Over the past few years, efficient methods have been developed to estimate densely sampled stacking velocity fields with the aim to improve the S/N ratio, the spatial resolution and the frequency content of the stack. As a by-product, this densely sampled attribute cube also becomes open to interpretation. However, the raw estimates of the automatically derived velocities are often too noisy for immediate use. Fortunately, due to their dense nature we can make use of efficient geostatistical techniques as Factorial Kriging to perform quality control and to remove noise. We will show on a real data example that these techniques can have a clear impact on the NMO stack result.