Cross-correlation weighted reverse time migration with wavefield decomposition based on optical flow vector

Due to the wavefield continuation based on the two-way wave equation, the imaging results of Reverse Time Migration (RTM) obtained by cross-correlation imaging conditions contain a lot of low-wavenumber noise. Although the method of RTM with wavefield decomposition based on Poynting vector can effectively suppress these noises, it also has some problems, such as unstable calculation of Poynting vector, low accuracy of wavefield decomposition, and difficult determination to the weight of each imaging section after wavefield decomposition. Based on the optical flow vector method, this paper implements the high-precision wavefield decomposition for the source and receiver wavefields, and generates up-left, up-right, down-left and down-right going wavefields. Then 16 sections will be obtained by cross-correlation of every two wavefields. On this basis, the imaging results of the 16 sections are respectively correlated with the reference section, and the cross-correlation values are used as weights to the imaging results of each section. Finally, the cross-correlation weighted RTM with wavefield decomposition based on optical flow vector is implemented. The test of theoretical model and field data show that compared with the conventional RTM with wavefield decomposition based on Poynting vector, the method proposed in this paper can achieve wavefield decomposition more accurately and stably both for the source and receiver wavefields. Moreover, the correlation weighted imaging method adopted in our approach avoids the defects of artificial selection or only assigns empirical weight to each imaging section after wavefield decomposition, and effectively improves the imaging accuracy of RTM.