Seismic Diffraction Imaging for Improved Coal Structure Detection by a Structure-Oriented Moving Average Error Filter

Instead of dismissing seismic diffractions as interfering noise and ignoring or suppressing them, they can be used to identify small geological discontinuities such as faults, dykes, and other features that act as small scatters in the subsurface. We describe a simple, but fast and effective, structure-oriented, moving average error filter (MAEF) applied to the neighboring traces to extract diffractions from time-stack reflection seismic data or diffractors from migrated seismic data. The filter estimates the reflections with the average values of the neighboring traces along the reflection dip direction, which can be computed by the gradients of seismic data. The difference between the input data (unmigrated or migrated) and the estimated reflections yields the diffractions. By identifying diffractions, small faults and other minor features that are difficult to detect using conventional seismic reflection processing can be enhanced and detected. The effectiveness of the proposed method for diffraction extraction is demonstrated by numerical models and by applying it to a 3-D seismic survey data over a partially mined-out coal deposit.