Abstract
The conventional full waveform inversion (FWI) is often suffered from the cycle skipping problem. In order to solve this problem, we introduced the wavelet transform to the FWI, and combined with the least squares filter in the wavelet domain. It can effectively reduce the influence of the cycle skipping problem in the inversion procedure, and improve the stability of the FWI. The least square filter has the higher accuracy in the wavelet domain than in the time domain. By using this feature we can narrow the phase difference between the predicted data and observed data, and construct a new objective function to make the inversion procedure steadily converge to the global minimum. Meanwhile, due to the multiscale characteristic of the wavelet transform, the data can be divided into different frequency bands. We could run the FWI in a multiscale way. The results from the synthetic example demonstrates that the multiscale adaptive FWI based on the wavelet transform is much less dependent on the initial model and low-frequency data. The method can also effectively reduce the cycle skipping problem and more robust than the conventional FWI.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call