next up previous print clean
Next: Discussion-Conclusion Up: Multiple attenuation results Previous: Inside the salt boundaries

Stacked sections

I now compare the stacked sections for the three techniques in Figure [*]. These figures are clipped to the same value for direct comparison. It is clear from Figure [*] that the best multiple attenuation result is obtained with the pattern-based technique with 3-D PEFs. Then comes the pattern-based method with 2-D PEFs, then the adaptive subtraction and the HRT. The HRT is simply unable to correctly attenuate multiples in the complex area defined by the salt boundaries. However, it does a fairly good job outside the salt boundaries. The quality of the adaptive subtraction method is hampered by modeling uncertainties, especially where the diffracted multiples are the strongest, i.e., between four and five seconds.

It is pleasing to see that the pattern-based method, while preserving the primaries, removes the diffracted multiples very well. The multiple attenuation result with 3-D filters is even more striking: the diffracted multiples are almost completely gone (see at four seconds). Although the model of the multiples is not perfect, there is enough resemblance between the ``real'' diffracted multiples and the modeled ones across dimensions to obtain a satisfying attenuation.

As a final comparison, I show in Figure [*] the stacked sections of the removed multiples for the three methods. It shows that the pattern-based approach with 3-D filters removed more multiples than any other method.

 
compstack-multatt
compstack-multatt
Figure 6
Stacked sections of the estimated signal. (a) Adaptive subtraction result. (b) HRT result. (c) Pattern-based method result with 2-D PEFs. (d) Pattern-based method result with 3-D PEFs.
[*] view burn build edit restore

 
compstack-diff
compstack-diff
Figure 7
Stacked sections of the removed multiples. (a) Adaptive subtraction result. (b) HRT result. (c) Pattern-based method result with 2-D PEFs. (d) Pattern-based method result with 3-D PEFs.
[*] view burn build edit restore


next up previous print clean
Next: Discussion-Conclusion Up: Multiple attenuation results Previous: Inside the salt boundaries
Stanford Exploration Project
7/8/2003