Extraction automatique d'information géologique à partir d'images sismiques tridimensionnelles

Abstract : Three-dimensional seismic images are the main source of information for the geomodeling of oil reservoirs. We propose here several algorithms, implemented in the gOcad geomodeling software, to automate their long and difficult interpretation. First, a continuous approach of the seismic signal enables us to implement a library of seismic attributes. Two methods of combination of these attributes are then studied: one based on the P.C.A. and moving centers, to compute seismic facies; the other, entirely new, based on interpolation maps, to compute meta-attributes. Lastly, two new techniques of automatic picking of seismic horizons are proposed, with editing tools. Faults picking is also approached under a bibliographical angle.
Document type :
Theses
Complete list of metadatas

https://hal.univ-lorraine.fr/tel-01749693
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 12:23:48 PM
Last modification on : Wednesday, July 25, 2018 - 9:59:49 AM
Long-term archiving on : Friday, September 14, 2018 - 10:53:17 AM

File

INPL_T_2004_LABRUNYE_E.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01749693, version 1

Citation

Emmanuel Labrunye. Extraction automatique d'information géologique à partir d'images sismiques tridimensionnelles. Sciences de la Terre. Institut National Polytechnique de Lorraine, 2004. Français. ⟨NNT : 2004INPL090N⟩. ⟨tel-01749693⟩

Share

Metrics

Record views

16

Files downloads

13