Smoothing/Edge Filters Overview

Process name “Smoothing/Edge Filters”

Reduce noise in seismic data by smoothing in 3D. Three methods used are Gaussian Blur, Sobel Edge Detection, and Symmetric Nearest Neighbor smoothing.

Smoothing the noise away will make the seismic data look more attractive and easier to interpret. Several 3D spatial filtering tools like ’3D Volumetric Curvature’ will benefit from smoothed data.

This tool is also used to improve the edge effects when going into coherency or Volumetric Curvature. The Sobel Edge Detection filter performs a 3D spatial gradient measurement on the data to emphasize regions that correspond to edges. The Symmetric Nearest Neighbor filter is designed to preserve edges in the data and is a very effective at noise reduction.

The images below show three types of Smoothing Filters.

  • Amplitude
  • Gaussian Blur 9pt
  • SNN (Smooth + Edge) 9pt

The time-slice images below show four different Smoothing Filters.

  • Amplitude
  • Gaussian Blur 9pt
  • Edge Detection
  • Edge Detection + Smooth 9pt