An efficient and effective common reflection surface stacking approach using local similarity and plane-wave flattening

•An efficient and effective approximate CRS stacking approach has been proposed.•The proposed approach can obtain high-resolution ZO image by two step stackings.•Similarity-weighted stacking is used to stack along the offset direction.•Plane-wave flattening is used to stack in the midpoint direction.•A synthetic and two field data examples show successful performance.

We propose an efficient and effective approximate common reflection surface (CRS) stacking approach to obtain a clean zero-offset (ZO) seismic image. Since the basic purpose of CRS stacking is to first stack along offset direction and then smooth along midpoint direction, we propose to first use local similarity as the weight to stack pre-stack seismic data along offset dimension and then to use plane-wave flattened events as the integral surface along midpoint dimension for stacking. Compared with the classic CRS stacking approach, we only need to calculate the local slope and to scan the normal moveout (NMO) velocity in the proposed two-step approximate CRS approach, thus it is more efficient and applicable to real seismic data processing. Both synthetic and field data examples demonstrate a successful performance of the proposed approach.