Efficient RTM angle gathers using source directions

In general, the pre-imaging condition methods for generating angle-domain common-image gathers are computationally expensive. Algorithms that require full or partial wavefield decomposition at every image point prior to imaging are considered prohibitive for their application to full wide-azimuth surveys. Direction-vector-based methods are an attractive alternative to reduce the cost, but are often troubled by complicated wavefields that exist in geologically complex areas. In this paper, we present a method that reduces the cost of most preimaging condition methods and in particular, direction-vectorbased methods while improving their quality considerably. We present the detailed workflow and substantiate our claims with the help of synthetic and real data examples. In 3D, it is also very important to choose opening angles and azimuths such that they uniformly sample the surface of the unit sphere around the image point. However, this is usually not the case, and we discuss a possible scheme to achieve that.