• 060 Ma
    Late Paleocene

    March 2007

  • 050 Ma
    Early Eocene

    March 2007

  • 040 Ma
    Middle Eocene

    March 2007

  • 030 Ma
    Early Ollgocene

    March 2007

  • 020 Ma
    Early Miocene

    March 2007

  • 010 Ma
    Late Miocene

    March 2007

  • 000 Ma
    Present Day

    March 2007

imode your observatory
to a transparent earth

IMODE 202: Automated Imaging of Multi-Scattering Data

Class Information
  • Instructor: Luc Ikelle
  • Textbook
  • TBA
  • Regular Price: $850 US
  • Student Price: $450 US
  • Duration: 2 days
  • Prerequisite: IMODE 102, IMODE 103
  • Computer Requirements: Unix/Matlab or Windows/Matlab

Upcoming Class Information

Date: TBA
Location: TBA
MODE 202 is pending scheduling. We will post the relevant information once a date has been picked. Attendees can register online for our other available classes.

Detailed Overview

The key processes in the current imaging of multiscattering data include:

  • removing from the data the direct waves and free-surface events and leaving those with no reflection at the free surface (single-scattering and multiple-scattering events)
  • removing multiple-scattering events and leaving events with only one reflection in the subsurface (single-scattering events)
  • locating the scattering points and reflectors inside the medium using single-scattering events only

This last step in the current imaging process is known as single-scattering imaging. In other words, the prerequisites for single-scattering imaging are that the free-surface events and multiscattering events have been removed from the data.

When the formulation of the single-scattering imaging is based on the Born approximation (which is also known as the single-scattering approximation), as is classically the case, we assume that the medium is decomposed into two parts: the reference medium and the perturbation. The perturbations are the scattering points and reflectors inside the medium which are the cause of single-scattering data. The reference medium is the part of the medium which allows us to model the wave propagation from the source of waves to the scattering points and from the scattering points to the receivers. The single-scattering imaging assumes that the reference medium is known, and it focuses only on locating the scattering points and reflectors inside the medium. In other words, in addition to removing free-surface events and multiple-scattering events from the multiscattering data, we also have to know (or estimate) the reference medium before performing the single-scattering imaging. Fulfilling these prerequisites of single-scattering imaging is the most challenging aspect of the current imaging process; in the case of seismic data, for example, fulfilling these prerequisites requires months and months of investigation and human verification. So the view that the difficulties associated with the current multiscattering imaging will be overcome by the ever-increasing computational speed and storage is, at the very least, incomplete. We need to rethink and even reformulate this problem while taking advantage of computational advances.

There are two possible ways of approaching the automatization of multiscattering data. One approach is to automate each of the three current steps in the imaging processes described earlier, and then to use some control systems to link these three steps. The other approach is to image all the events in the multiscattering data, including free-surface events and multiple-scattering events, simultaneously. This course describes recent advances in these approaches. In the first part, we describe automated algorithms for removing free-surface events, for removing multiple-scattering events, and for single-scattering imaging. In the second part, we describe algorithms which can be used to directly image multiscattering data.

This course is meant for researchers, scientists, and engineers involved in the analysis and interpretation of multiscattering data. This naturally includes seismologists, opticians, quantum physicists, etc. The only background required of the participants is a good knowledge of advanced calculus, linear algebra, and the basic notions of elasticity (Hooke's law and Newton's equations). Prior exposure to the scattering theory is not needed, since we develop all of the materials from the first principles.

Our numerical illustrations in this course are limited to seismic scattering experiments and to seismic data. As we would like to make this course accessible to researchers, scientists, and engineers involved in the analysis and interpretation of multiscattering data from other disciplines, we will start with some basic background on seismic scattering problems.

Course Outline

  • Attenuation of Free-Surface and Multiple-Scattering Events
  • Single-Scattering Imaging
  • Multi-Scattering Imaging: Diagrammatica
  • Multi-Scattering Imaging: Focusing
  • Multi-Scattering Imaging: ICA
  • Seismic Interpretation

© 2015 iMode Energy Research. All Rights Reserved.

Houston Web Design - Graphics by Cindy