Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Earth science
Reexamination Certificate
1999-08-06
2001-08-21
McElheny, Jr., Donald E. (Department: 2862)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Earth science
Reexamination Certificate
active
06278949
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This method is concerned with automated segmentation of geophysical data for the purpose of providing a model of the shapes and properties of earth layers beneath a region of interest.
2. Discussion of Related Art
Geophysical exploration of the earth's subsurface is commonly done by illuminating the subsurface earth layers using suitable radiation. The radiation may be acoustic, electromagnetic, radioactive or of any other nature now known or unknown. During a survey, a radiation source, positioned at or near the surface of the earth, visits a plurality of survey stations arranged in a grid-like pattern over a region of interest. The source is associated with an array of spaced-apart receivers that are responsive to the radiated energy after that energy has passed through and has been filtered and reflected by the subsurface earth layers. At each survey station, the source is activated to allow the respective receivers to record a snap-shot image of a portion of the subsurface in the light of chosen radiation. Each resulting image is digitally recorded on an archival storage medium in the form of a plurality of time-scale traces, usually one trace per receiver group location. Each trace exhibits a series of transients or waveforms in terms of trace amplitude vs. source-to-receiver travel time. On the multi-trace recordings the respective waveforms of the series may be representative of the stratigraphic sequence of a volume of the earth in the vicinity of each survey station, plus unwanted glare or noise.
It is usual practice to arrange the field procedure to allow redundant overlap of adjacent records. Preliminary routine data processing contemplates filtering out the noise, correcting the data for angularity, spherical spreading and sharpening the wavelet imagery using various well-known stacking and multiple-reflection abatement and migration routines.
From the survey, with particular reference to well-known seismic surveying practices by way of example but not by way of limitation, the processed records are merged to produce a 3-dimensional (3-D) image of a volume of the earth in the selected region. To that end, events (locations of a specified phase of waveform) are tracked continuously across adjacent traces and between records by simple inspection, cross-correlation or other well-known means. Each event, representative of a particular geologic horizon or earth layer, is tracked from trace-to-trace and station-to-station to build the desired model of the subsurface earth layers. The desired model may be purely structural to define the limits of a mineral or hydrocarbon deposit. Concomitantly, using attribute methods, some physical characteristic such as the lithology or porosity of a particular rock layer may be mapped.
The interpreted results of a survey may be shown as a topographic map of one or more specific earth layer interfaces or as a sequence of one or more horizontal frames across a data volume, referred to as time slices, to provide subcrop maps.
In a large survey region, it is prohibitively laborious for a human interpreter to follow a plurality of seismic horizons over thousands of stations. The task is preferably relegated to a computer. Several known computer-aided methods of geological interpretation of geophysical data are now presented by way of review. Such methods fall into two categories based on non-uniform- and uniform-time sampling of seismic attributes.
For the first category, Paulson et al., published a paper on horizon tracking entitled Automatic Seismic Reflection Picking, Geophysics v.33, 1968, p431-40. The method first picks the time positions of either all amplitude maxima or amplitude minima on a seismic trace (i.e. a seismic record at one spatial location). This step creates an array of unequally spaced time positions. Second, it uses maximum normalized cross-correlation criterion to match seismic waveform in a time window centered at picked time on one with that in a possibly time-shifted window centered at picked time on a neighboring trace. The continuous trajectory between reflection times in matched windows on contiguous traces is defined as a seismic horizon. Finally, the method associates with each picked time an attribute called grade, which is the product of the length of a trajectory and mean maximum or minimum amplitude along the trajectory. This method was applied to 2-D seismic data since 3-D data did not exist at the time.
Hildebrand et al. in U.S. Pat. No. 5,251,184 issued Oct. 5, 1993, entitled Method and Apparatus for Finding Horizons in 3-D Data, disclose a process that applies one-trace forward and backward search within specified time window and an amplitude tolerance attribute to track horizons in 3-D data. This method is a modification, of Paulson et al. method, where (a) it iterates between forward and backward trace directions among four orthogonally disposed neighbor traces to verify connectivity among picks, (b) it provides an interactive graphical means of introducing a seed point from where the horizon tracking expands laterally, (c) it keeps a history of trajectory expansion starting from the seed point, (d) it enables the user to edit erroneous trajectories, and (e) it displays a tracked cross-section on a graphical screen. Since the seed input and trajectory editing require human intervention, the horizon tracking with this method sets a practical limit on the number of horizons that may be mapped in a given time.
Having tracked one or more horizons over the region, it is of interest to estimate the dip and strike (or azimuth). Dalley et al., in an article entitled Dip and Azimuth Displays for 3-D Seismic Interpretation, in First Break, v. 7, 1989, p86-95, discuss a method for estimating the dip and strike of an automatically tracked or interpreted horizon. The dip and azimuth are spatial (lateral) attributes associated with the tracked horizon time.
The horizon tracking algorithms have problems. First, the tracked horizon (s) sometimes fail to close a loop even on a locally continuous surface. Problems related to loop closure on continuous surfaces include cycle-skipping due to aliasing in the presence of steep dips and inability to extrapolate the horizon across areas of poor record quality. Second, the Hildebrand method requires a seed point to pick a horizon. That is, a human must tell the computer where to start. In this method, since only one attribute—the maximum or minimum amplitude—is used, noise leads to tracking of pseudo horizons especially in areas of poor record quality. Third, both Hildebrand and Dalley methods provide only sparse information confined to particular interpreted horizons out of an entire volume of possible horizons. Fourth, although all horizon tracking based methods provide structural information, i.e., bulk geometrical shapes of horizons, yet they lack stratigraphic and lithologic information, i.e. detailed subdivisions between boundaries of structural layers and their properties such as porosity or fluids filling the pores. Finally, these methods do not provide any information in regions where horizons cannot be tracked.
For the second category, Taner et al., in a paper entitled Complex Seismic Trace Analysis, Geophysics, v.44, 1979, p1041-63, quantified the character of a seismic reflection through the use of Hilbert Transform to calculate temporal attributes: Instantaneous Envelope, Instantaneous Phase and Instantaneous Frequency. Instantaneous Envelope provides the lithologic component and Instantaneous Phase the structural component of stratigraphy. The first advantage of uniform time sampling based method is that it does not require horizon tracking since attributes are estimated at equal increments of time, unlike the first category methods where attributes are calculated only at unequally spaced horizon picks. The second advantage is that it provides an attribute value even in those regions where a horizon from first category methods does not exist. The disadvantage of Taner's attribute method is that th
Marnock Marvin J.
McElheny Jr. Donald E.
LandOfFree
Method for multi-attribute identification of structure and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for multi-attribute identification of structure and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for multi-attribute identification of structure and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2439824