Amplitude Increment Coding Based Full Waveform Inversion
S. Dong, L. Han, Y. Hu, P. Zhang and B. Mao
Event name: 81st EAGE Conference and Exhibition 2019
Session: Full Waveform Inversion I
Publication date: 03 June 2019
Info: Extended abstract, PDF ( 903.05Kb )
Price: € 20
The conventional full waveform inversion(FWI) suffers from the cycle skipping problem which is mainly caused by the observed data lack of low frequency components. And the noises in the observed data also has impact on FWI results. We proposed a amplitude increment coding method to help FWI avoid cycle skipping problem. It introduce the amplitude increment encoding matrix of the observed data and the calculated data, and by comparing the matrices to find the positions of the elements in the two matrices that are not equal, and then set the value of these positions in the calculated data to zero,so the portion of data which causes cycle skipping is eliminated. Moreover, we use the global-correlation misfit function for iteration to weaken the influence of the positions in the calculated data which are set to zero. The numerical examples show that FWI based on amplitude increment coding method generate better results when the observed data lack low-frequency components,and this method has strong noise immunity.