# US 4,397,006 A · Cross trace coherent noise filtering for seismograms

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 4,397,006 A
- Document ID: US4397006A
- Status: Granted
- Kind: A
- Application number: 06/221,865
- Filed: 1980-12-31
- Granted: 1983-08-02
- Earliest claimed: 1980-12-31
- Assignees: Mobil Oil Corporation
- Inventors: James N. Galbraith, Jr.
- Web page: https://patentgrep.com/patent/US4397006A

## CPC

- G01V 1/366: {by correlation of seismic signals}
- G01V 2210/3246: Coherent noise, e.g. spatially coherent or predictable
- G01V 2210/52: Move-out correction
- G01V: GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS (means for indicating the location of accidentally buried, e.g. snow-buried, persons A63B29/02)

## Abstract

In seismic exploration windows from a pair of contiguous traces are used in a process which determines the coherent noise on a basis trace and removes this coherent noise from a prediction trace. The operator of a digital filter is determined by autocorrelating the basis trace, cross-correlating the window of the basis trace, with the window of the prediction trace, and selecting the filter operator which produces the least error between the cross-correlation function and the product of the autocorrelation function and the operator.

## Claims

### Claim 1

In seismic exploration wherein sets of seismic traces represent the reflections of seismic energy detected at locations along a line of exploration, the method of filtering coherent noise which has a different moveout from trace to trace than said reflections comprising:
selecting sampling windows from a pair of contiguous traces in said set, one being a basis trace and the other a prediction trace;
determining the coefficients of a digital filter which correlates the coherent noise in the window of said basis trace with the coherent noise in the window of said prediction trace;
filtering said basis trace with said digital filter to produce an estimate of said coherent noise;
removing said estimate of said coherent noise from said prediction trace; and
repeating all of the foregoing steps for other pairs of contiguous traces in said set to remove said coherent noise from said traces.

### Claim 2 (depends on 1)

The method recited in claim 1 wherein the step of determining the coefficients of a filter comprises:
autocorrelating the window of the basis trace;
cross-correlating the window of said basis trace with the window of said prediction trace; and
selecting the coefficients which produce the least error between the cross-correlation function and the product of the autocorrelation function and the coefficients.

### Claim 3 (depends on 1)

The method recited in claim 1 wherein the step of selecting the window from said prediction trace comprises:
cross-correlating the window of said basis trace with said prediction trace through all sampling intervals in which said coherent noise may appear;
determining the coefficients of a digital filter which correlates the coherent noise in said window of said basis trace with the coherent noise in a window of said prediction trace;
determining the least squared error between the coherent noise in said basis trace and the noise predicted from said filter coefficients;
repeating the foregoing two steps for different sampling windows of said prediction trace through all of said sampling intervals; and
selecting the minimum least squared error to identify the sampling window to be selected for said prediction trace.