Abstract
Conventional L2 norm based full waveform inversion usually has more than one local minima and is also sensitive to noise in the observed data. The Earth is at least attenuative, and thus, a pixel-to-pixel fitting of the waveform is not feasible in practice. We propose a normalized nonzero-lag crosscorrelation based elastic full waveform inversion algorithm, which intends to maximize the similarity of the predicted and observed data. The proposed objective function emphasizes that the matching the phases of the predicted and observed data is more immune to the simplified physics we often use to represent the medium. The normalization term can compensate the energy loss in the far offset and avoid the estimates being biased by extreme values in the observed data. We introduce a polynomial-type weighting function and suggest an approach to determine the optimal time lag. A modified elastic Marmousi model is used to verify the effectiveness of the proposed method.
| Original language | English (US) |
|---|---|
| Title of host publication | 80th EAGE Conference and Exhibition 2018 |
| Publisher | EAGE Publications BV |
| ISBN (Print) | 9789462822542 |
| DOIs | |
| State | Published - Oct 16 2018 |
Bibliographical note
KAUST Repository Item: Exported on 2020-10-01Acknowledgements: We thank Yike Liu, Bingbing Sun, Juwon Oh and Yunseok Choi for their helpful discussions. We thank KAUST for its support and specifically the seismic wave analysis group members for their valuable insights. For computer time, this research used the resources of the Supercomputing Laboratory at King Abdullah University of Science & Technology (KAUST) in Thuwal, Saudi Arabia.