CROSS‐CORRELATION FILTERING

Abstract

Correlation analysis techniques may be applied to seismic data already subjected to standard recording and analysis procedure in an effort to extract additional information, or to raw data as an alternative filtering method. These techniques involve determination of certain parameters which provide a quantitative measure of the correlation between two sets of data. Among the most useful of these parameters are the auto‐ and cross‐correlation coefficients and functions long used by statisticians in time series analysis and recently applied to filtering and prediction problems in the field of communications. This paper discusses some applications of correlation analysis in interpretation of seismograms. The use of cross‐correlation analysis to identify weak reflections masked by high noise is illustrated for several problems. Equivalence of correlation analysis procedures to filtering operations is stressed. Special analog computing equipment facilitating computation of correlation coefficients and power spectra directly from oscillograms or graphs is described. A brief discussion of modern optimum filter theory is presented.