Inverse-Scattering Internal Multiple Prediction (ISIMP)

Overview Library

Predict internal multiples from all generating horizons simultaneously

The inverse-scattering internal multiple prediction (ISIMP) algorithm is a data-driven tool that predicts all orders of internal multiples for all horizons simultaneously, without need of prior subsurface information. Originally developed at the University of Houston by the Mission-Oriented Seismic Research Project (M-OSRP) consortium, it is a wave-theory-based method employing multidimensional prediction in the pseudodepth frequency domain. The algorithm

  • predicts internal multiples with converted waves, head waves, diving waves, and diffractions
  • does not require picked horizons
  • eliminates the need for depth images or velocity models.

Normal incidence (1D) and angle-of-incidence (1.5D) prediction modes are available and both can be applied on prestack or poststack data according to the survey application. ISIMP requires data with surface multiples eliminated (e.g., by using 3D GSMP general surface multiple prediction) and high signal-to-noise ratio. The 1.5D version also requires offset gather regularization and zero-offset reconstruction prior to prediction.

Request More Information