Tomographic inversion using shape and velocity parameters

The structure of a section of earth is sometimes well represented by a small collection of lithological units of constant or linearly varying velocity. Even without knowing what the velocities are, the knowledge that they are constant or linearly varying within discrete units is a substantial constraint to the inversion problem. By setting the shape of lithological units and the velocity constants within the units as the free parameters, it is possible to restrict attention at every iteraction of this inversion process to models that retain the prior description. The focus is then on recovering the shape of the lithological units and a small number of velocity parameters rather than the velocity at every point in the panel. A synthetic example is used to illustrate the technique.