2003 RESEARCH UTAM ACCOMPLISHMENTS AND 2004 GOALS
- Migration Deconvolution
- RESULT: Gaussian beam Migration Deconvolution
as accurate as wave equation MD, but much cheaper.
Noticeably more accurate than ray-based MD.
- IMPLICATION: Noticeable improvement in MD performance for data collected over
- PROBLEMS: None.
- 2004 GOALS: 3D Gaussian Beam MD
- Multiple Prediction and Migration
- RESULT: Several approaches. 1). Extension of Wiggin's
extrapolation-prediction of free-surface-related
multiples seems to work well on synthetic and field data.
2). VSP multiple migration extends subsurface illumination and
is somewhat insensitive to static+velocity errors.
3). Semblance analysis and least squares migration filtering of multiples
in Unocal data.
- IMPLICATION: Several new approaches to either eliminating
multiples or imaging multiples.
- PROBLEMS: Immaturity in testing of field data.
- 2004 GOALS: Full 2-D tests with field data. Multiple migration of 3D VSP data.
- Seismic Interferometry
- RESULT: Several results: 1). Derivation of modeling, adjoint, asymptotic inverse, and resolution equations
for seismic interferometry. 2). Convincing tests show migration of
VSP crosscorrelograms yields images with
wider subsurface coverage. These images are roughly immune to static problems
and somewhat insensitive to migration velocity errors.
3). Seismic guide stars help focus reflection energy below basalt.
- IMPLICATIONS: Better way to image VSP data is staring us in the face.
Possible means to help focusing below salt or basalt.
Mathematical foundations of seismic interferometry developed.
- PROBLEMS: Immature testing phase for seismic guide stars.
- 2004 GOALS: Theory of statistical seismic interferometry.
Possibly 3D VSP crosscorrelogram migration. Rigorous testing of seismic
guide star idea with realistic synthetics and possibly field data.
- Primary Only Imaging Condition + Radon Filtering of Multiples
- RESULT: POIC+Radon filtering of SMAART data multiples.
- IMPLICATIONS: Can eliminate some near-offset multiples.
- PROBLEMS: Not all near-offset multiples eliminated.
- 2004 GOALS: Replace ray tracing with Gaussian Beam modeling. Replace
Radon transform with sparse Radon transform.
- First Arrival Waveform Imaging and Refraction Migration
- RESULT: Several results. 1). New efficient Fresnel Tomographic inversion
of 1st arrival and most energetic arrival.
2). Numerical test of tomogram sensitivity to source-receiver configuration.
3). VLF+refraction tomography > one method in getting accurate subsurface
Efficient and robust alternative to waveform inversion, possibly.
- PROBLEMS: No synthetic data tests. Waveform inversion of land data failed to
provide any better results than traveltime tomography.
- 2004 GOALS: Extensive synthetic tests with Fresnel Tomography.
- Target-oriented Least Squares Migration.
Prestack LSM imaging of target-oriented area in UNOCAL poststack data
showed improvement in resolution and efficient convergence.
However, severe noise was introduced into the section.
- IMPLICATION: Cheap LSM?
- PROBLEMS: Noise is a showstopper. Must use regularization and
preconditioners. Pay attention to the velocity model as well.
- 2004 GOALS: 2D and 3D Target Oriented LSM with regularization+preconditoning.
- Segmenting and Labeling Rocks in Photos
RESULT: Morphological processing+watershed transform allow labeling+segmentation
of rocks in photos.
- IMPLICATION: Computerized Optical Stratigraphy.
Subtle clues about rock history obtained from photos.
- PROBLEMS: Not a perfect segmentation record, about 80 percent accurate.
Need user intervention.
- 2004 GOALS: Southern Utah photos of stratigraphic sections analyzed.
- Efficient Steep-dip Anisotropic Migration
RESULT: Efficient anisotropic migration operator
developed for steep dips and VTI, TI, HTI and orthorhombic media.
- IMPLICATION: As fast as standard 45 degree dip migration operators.
- PROBLEMS: 2D so far.
- 2004 GOALS: 3D anisotropic migration operator.