Velocity modelling using a generalized linear inversion technique

Abstract

Many post slack velocity modelling methods currently used in the seismic processing industry involve a top-down ray tracing procedure. In other words, rays are traced from the surface to determine the spatial locations of the next horizon of interest and the interval velocity, the procedure then being repeated as many times as there are layers of interest. A difficulty with these methods is that the error in the determination of the interval velocity is cumulative with increasing depth. In an attempt to minimize this difficulty generalized linear inversion techniques are being used (Twomey 1977; Cooke & Schneider 1983; Lines & Treitel 1984). It is this technique which is used in our approach to velocity modelling. Specifically we wish to obtain a two-dimensional interval velocity model from data available on the unmigrated stacked section.