Structure-oriented smoothing and semblance

Smoothing along structures apparent in seismic images can enhance these structural features while preserving important discontinuities such as faults or channels. Filters appropriate for such smoothing must seamlessly adapt to variations in the orientation and coherence of image features. I describe an implementation of smoothing filters that does this and is both computationally efficient and simple to implement. Structure-oriented filters lead naturally to the computation of structure-oriented semblance, an attribute commonly used to highlight discontinuities in seismic images. Semblance is defined in this paper as simply the ratio of a squared smoothed-image to a smoothed squared-image. This definition of semblance generalizes that commonly used today, because an unlimited variety of smoothing filters can be used to compute the numerator and denominator images in the semblance ratio. The smoothing filters described in this paper yield an especially flexible method for computing structure-oriented semblance.