An iterative approach to seismic skeletonization

Abstract

Abstract In this paper pattern recognition techniques are applied to the automatic tracking of seismic reflectors. We call this procedure seismic skeletonization, since it reduces a seismic section into a line drawing corresponding to seismic reflectors. The procedure uses an iterative approach. Strong reflectors are tracked first; weaker reflectors are tracked at later iterations. At any iteration, the tracking is confined to the areas delineated by the reflectors already tracked in a previous iteration. Reflectors are tracked by correlating neighboring seismic traces. An algorithm called the binary consistency-checking (BCC) scheme is used to correlate three traces at a time. The combined use of the iterative approach and the BCC scheme is efficient, emulating human interpretation and incorporating 'geologic trend' into the results.