next up previous [pdf]

Next: Acknowledgements Up: Ayeni et al.: Inversion Previous: discussion and conclusions

future directions

A more robust minimization such as L1 will likely yield better quality results than those shown in this paper. We anticipate that a combination of L1-minimization (including both data and model re-weighting) with our preconditioners will lead to a robust inversion scheme for simultaneous source time lapse data sets.

marm-nocoded-mig4d marm-encoded-mig4d marm-encoded-inv4d
marm-nocoded-mig4d,marm-encoded-mig4d,marm-encoded-inv4d
Figure 5.
(a) Single source migrated time-lapse image. Simultaneous source (b) migrated and (c) inverted time-lapse images. Note the cross-term artifacts in (b).[CR]
[pdf] [pdf] [pdf] [png] [png] [png]

marm-encoded-mig-rep-lab marm-encoded-inv-rep-lab
marm-encoded-mig-rep-lab,marm-encoded-inv-rep-lab
Figure 6.
Weighted RMS difference between the baseline and monitor images obtained via (a) migration and (b) inversion. Note that inversion attenuates the non-repeatbaility RMS energy and preserves preserving the production-related change.[CR]
[pdf] [pdf] [png] [png]


next up previous [pdf]

Next: Acknowledgements Up: Ayeni et al.: Inversion Previous: discussion and conclusions

2009-09-25