Professor LIU Yike introduces the vector median filter (VMF) to reduce noise in vector fields.

In this paper, he demonstrates similar benefits in applying the VMF to attenuate noise in a dip/azimuth vector field derived from a migrated seismic image. First, he introduces the concepts of the median value, median vector, scalar median filter (SMF), and VMF, along with a fast VMF algorithm. Then, he describes three properties of the SMF and VMF. Finally, he applies the VMF to synthetic and field seismic data to demonstrate that the VMF can effectively reduce noise while preserving discontinuities.

Figure 1. A migrated seismic amplitude section, which is used as the input for computing dip vectors of seismic events at every sampled location. (Image by LIU)

Figure 2. A subset of the dip vectors before applying the VMF. (Image by LIU)

Figure 3. Field data after filtering by mean filter. (Image by LIU)

Figure 4. Field data after filtering by VMF. (Image by LIU)

Yike Liu. Noise reduction by vector median filtering. Geophysics, 2013, 78(3): V79-V86(Download Here)