Structural Oriented Signal Decomposition Using Adaptive Frequency Distributions

Signal decomposition is often used for delineating turbidite plays, as well as channels and deltas. Post-stack seismic is comprised of a range of frequencies stacked together to create a coherent and concise image. Using signal decomposition, one can decompose the seismic data, singling out desired frequencies, and dulling out the rest. This makes capturing changes of phase, and hence channels, easier. There have been two major features lacking in these attributes, namely an even distribution of frequency occurrences, and loss of information along dipping stratigraphic structures. We propose a new technique that uses an adaptive algorithm, in addition to a local three dimensional trace extraction that filter perpendicular to the stratigraphy. The adaptive algorithm caters to the very nature of seismic segmenting the frequency distribution accordingly. This preserves an even distribution of the sum of frequency occurrences, and hence reduces over-and-under sampling considerably. This results in an improved delineation of the both minor and major channels, a higher imaging resolution and contrast, and a reduction in geological noise. The fact that the trace is filtered perpendicular to the local stratigraphy, this further enhances the resolution, and gives a more accurate result.