Coherent noise attenuation using the Wavelet Transform on Radial Basis
Lucas José A. de Almeida, Rafael Rodrigues Manenti, Milton José Porsani · 2015
Summary Seismic data are often contaminated by coherent noise that mask the reflections needed to properly construct a accurate subsurface image. Those noises usually consist of ground roll and swell noise, for land and marine seismic data, respectively. The wavelet transform has been used with moderate success to attenuate such noises due to its property of segregating seismic events of specific orientations in different subbands. However, this approach is highly susceptible to aliased events, which is a common characteristic of coherent noises due to its highly vertical orientation. Also, due to its low angular resolution, the wavelet transform fails to properly segregate low velocity reflections, placing them at the noise sub-band and inadequately attenuating them. In this work, we present a method based on the joint application of the wavelet and radial transforms in order to attenuate coherent noise. We use the radial domain as a framework to overcome the aforementioned limitations of the wavelet transform.