Tensor-Based Method For Representation, Analysis, and Reconstruction of Seismic Data

• US 20130090857A1
• Filed: 08/17/2012
• Published: 04/11/2013
• Est. Priority Date: 10/06/2011
• Status: Active Grant

• Current Assignee: Dimitrov, Pavel, Imhof, Matthias, Kaufhold, John P.
• Original Assignee: Dimitrov, Pavel, Imhof, Matthias, Kaufhold, John P.
• Inventors: Imhof, Matthias, Dimitrov, Pavel, Kaufhold, John P.
• Granted Patent: US 9,176,247 B2
• Time in Patent Office: Days
• US Class Current: 702/14
• CPC Class Codes: G01V 1/30 Analysis G01V1/50 takes precedence
• CPC Class Codes: G01V 1/301 for determining seismic cross-sections or geostructures
• CPC Class Codes: G01V 1/345 Visualisation of seismic data or attributes, e.g. in 3D cubes
• CPC Class Codes: G01V 2210/63 Seismic attributes, e.g. amplitude, polarity, instant phase

Abstract

Method for representing seismic data as a spatially varying, second-order tensor field (23). The spatial relationships expressed in these tensors are exploited to link, classify, or separate neighborhoods (22); or to infer global or relational properties among them, thereby suggesting geobodies despite a noisy background. The tensors may be decomposed into their fundamentals that may either be used directly as derivative datasets or attributes, or may be used to facilitate linkage, classification, or separation of neighborhoods or analysis of linkage patterns. Decomposition may be by eigenvalues, with the eigenvalues used to define attributes called ballness, plateness and stickness (24). Alternatively, connections may be made between points in the data where the tensor has been computed, called tokens (21), based on tensor voting and polarity. Or, the connections may be based on a distance measure between tokens.

Claims

1. A computer-implemented method for analyzing a volume of seismic or seismic attribute data (the “

seismic data volume”

) to discover spatial location and shape of geologic features of a subsurface region, comprising;

making a second-order tensor representation of the seismic data, and using a computer to perform decomposition of the tensor representation to form seismic attributes;

wherein the tensor decomposition is eigenvalue decomposition, the eigenvalues being referred to as λ

₁, λ

₂, and λ

₃, and wherein the seismic attributes comprise at least one of ballness (λ

₃), plateness (λ

₂-λ

₃) and stickness (λ

₁-λ

₂) and their orientations; and

using the seismic attributes to interpret the seismic data volume for geologic features.

2. The method of claim 1, wherein the seismic data volume is a discontinuity volume, and the tensor representation uses a second-order, centered-moment tensor.

12. The method of claim 1, wherein the second-order tensor is a structure tensor or a moment tensor.

3. A computer-implemented method for analyzing a volume of seismic or seismic attribute data, either bring referred to as “

the seismic data volume,”

to discover spatial location and shape of geologic features of a subsurface region, comprising;

making a tensor representation of the seismic data, and then grouping the data by a tensor voting method, wherein components of a tensor are computed, using a computer, to represent each of a plurality of selected spatial points, called tokens, in the seismic data volume;

making connections between tokens based on the tensor voting but also on a binary polarity attribute; and

interpreting the connected tokens for geologic features.

4. The method of claim 3, wherein the binary polarity attribute is formed by giving each token a + or −

sign according to whether it corresponds to a peak or trough of a seismic data trace, and wherein the binary polarity attribute is used during or after the tensor voting process to support or cast doubt upon a proposed linkage between tokens, or before the tensor voting process to group the tensors before voting.

5. The method of claim 4, wherein a linkage between tokens of the same sign resulting from the tensor voting process is interpreted as indicating that their corresponding spatial locations belong to a common subsurface geological layer, and a linkage between tokens of opposite sign from the tensor voting process is interpreted as either signifying termination of a layer or a fault interrupting a layer.

6. The method of claim 3, wherein the tensor for each token is computed from seismic data associated with a selected spatial neighborhood of that token.

7. A computer-implemented method for analyzing a volume of seismic or seismic attribute data, either being referred to as “

the seismic data volume,”

to discover spatial location and shape of geologic features of a subsurface region, comprising;

making a tensor representation of the seismic data, and then grouping the data by a tensor method, wherein components of a tensor are computed, using a computer, to represent each of a plurality of selected spatial points, called tokens, in the seismic data volume, resulting in each token having an associated tensor representation of seismic data located around the token'"'"'s location;

making connections between the tokens based on a distance measure between the tokens and their associated tensor representations of seismic data; and

interpreting the connected tokens for geologic features.

8. The method of claim 7, wherein the distance measure is used in determining one or more nearest neighbors for a token and, from that, connectivity between tokens.

9. The method of claim 7, wherein the selected distance measure is the Mahalanobis measure.

10. The method of claim 7, wherein the tensor for each token is computed from seismic data associated with a selected spatial neighborhood of that token.

11. The method of claim 10, wherein a selected spatial neighborhood is based on the selected distance measure.