Plane-wave Prestack LSRTM

In order to further improve the quality of the image, the plane-wave LSRTM algorithm (equation ) is applied to these 31 plane-wave gathers and the image after 30 iterations is shown in Figure (c). Compared to the plane-wave RTM image, the LSRTM image contains fewer artifacts and shows better resolution. In the zoom view of the red boxes (Figure ), the horizontal reflectors in the LSRTM image are of higher resolution and are characterized by better balancing of amplitudes, which provides better delineation of the normal faults compared to the RTM image. In the deep part of the section (blue boxes, Figure ), the LSRTM image shows similar advantages to better illuminate the faults, in spite of the fact that these reflectors become more wiggly in the LSRTM image.

The common image gathers can be extracted from the prestack images and they suggest that the migration velocity is not very accurate in the deep part and the reflectors are undermigrated (Figure  and ). In this example, the LSRTM algorithm can only marginally improve the quality of the CIGs. The vertical artifacts at both edges of any one CIG are removed, and some of the reflector amplitudes are enhanced, so that they appear to be more continuous. In spite of the errors in the velocity model, the convergence of plane-wave LSRTM is stable and robust (solid line with squares in Figure ). In this example, the plane-wave LSRTM still shows better convergence than the conventional LSM.