Diffraction separation using structure-oriented orthogonal polynomial transform

Diffractions possess the potential to reveal subsurface discontinuities, such as faults, collapsed columns, and the rough edges of the salt body. We have developed a novel diffraction separation method that first flattens the dominant reflections and then identifies the diffraction wavefields through an orthogonal polynomial transform (OPT). Based on the plane-wave assumption and local slope, the reflection events can be predicted and flattened using plane-wave differential equations. To enhance the stability of the flattening process, a sliding window algorithm and three finite-difference forms are adopted. As the dominant coherent event, a reflection often corresponds to low-order coefficients in the OPT. To better preserve the diffraction energy, a time-varying order threshold is used when removing the dominant reflections. Two field data applications are presented to demonstrate the feasibility and effectiveness of the proposed diffraction separation strategy.