Multi-attribute processing techniques for the enhancement and interpretation of seismic data

Abstract

Summary form only given. Seismic data are collected, displayed, and interpreted in the time-distance domain (t-x). Local attributes of seismic data can be grouped into conventional single trace attributes (x=constant). The extraction of multitrace attributes is based on a computer-efficient implementation of localized slant stacking (beamforming) and median filtering. Image processing techniques are then applied to support the interpretation of migrated reflection seismic data whereby a seismic section is treated as a two-dimensional image. Local multitrace attributes have been used in a fast and robust coherency enhancement process for noisy seismic data. In a related application, multitrace attributes have provided the required independent data for successful multispectral image enhancement of seismic data. Multiattribute displays are well suited for the structural interpretation of migrated seismic data: this technique can be used for imaging of steeply dipping structures, analyzing uniformities and possible lithological boundaries, and highlighting focusing of diffracted energy and basin bounding faults.<>