Page 395 - Acquisition and Processing of Marine Seismic Data

386 7. SUPPRESSION OF MULTIPLE REFLECTIONS

FIG. 7.20 Schematic illustration of Radon velocity filter application. (A) A CDP gather with primary (red) and multiple
(dashed blue) reflections. (B) NMO correction with the velocities of primary reflections leaves a certain amount of residual
moveout (shown by Δt 1 , Δt 2 , and Δt 3 ) on the reflection hyperbolas of multiples, while the primaries are flattened.
(C) CDP gather in the τ-p domain after Radon transform and the mute zone (blue shaded area) involving the multiple ampli-
tudes. (D) CDP gather after muting and inverse Radon transform, and (E) original CDP gather after inverse NMO correction
with no multiple amplitudes.

Fig. 7.20B into the τ-p domain, the primary Fig. 7.21A illustrates an input NMO corrected
amplitudes (P) are aligned onto the zero residual CDP gather in which the primary reflections
moveout axis, whereas multiple amplitudes (M) are flattened while the multiples are still hyper-
are still hyperbolic and are mapped onto the pos- bolic and have a certain amount of residual
itive panel in the τ-p domain (Fig. 7.20C). The moveout. Its τ-p domain representation is shown
higher the residual moveout of the multiples, in Fig. 7.21B, computed for a residual moveout
the more distant their amplitudes from the zero range from 100 to 200 ms. The amplitudes of
moveout axis in τ-p domain. When the area of the primary reflections between approximately
multiple amplitudes indicated by shaded area 250 and 500 ms are located along the zero move-
inFig.7.20Caremutedout,aCDPgatherwithout out axis, whereas the amplitudes of the multi-
multiples is obtained after an inverse Radon ples between 500 and 1000 ms are mapped in
transform (Fig. 7.20D). This gather is still in the positive moveout panel. This differentiation
NMO corrected form, and if required, the origi- in the amplitudes based on their residual move-
nal CDP is reconstructed after an inverse NMO out times allows us to easily discriminate the
correction using primary reflection velocities multiple and primary amplitudes in the τ-p
(Fig. 7.20E). domain. As in Fig. 7.21C, we can mute out the
Fig. 7.21 shows an example application of amplitudes of the multiples to obtain multiple-
muting in the τ-p domain on a real CDP gather. free CDPs. Alternatively, one can also mute