Dip‐adaptive structurally conformable filtering using 2D complex wavelet transform

Reliable seismic inversion or geological interpretation requires post-migration noise attenuation to condition the data. We have developed a new approach, which is data driven, dip-adaptive, structurally conformable, to enhance the signal-noise ratio. The method includes two fundamental ingredients: (A) local dip-orientation analysis and (B) local attribute analysis, using the 2D complex wavelet transform (CWT) to generate the 5D local attribute volume for a 2D input data section. The local structural dip (A) is estimated independently by the semblance-based method. The filtering flow is: (1) perform a forward 2D CWT to the data in both in-line and x-line directions; (2) compute the 3D local structural orientation using semblance analysis; (3) perform filtering in the CWT transform domain; (4) perform the inverse 2D CWT in both in-line and x-line directions. The filtering criterion is to use the estimated local structural dip to separate the signal from the noise in the 5D CWT domain. Those matching the estimated local orientation are considered to be signal, while those unmatched are considered to be noise and subsequently removed before the inverse 2D CWT is applied. Unlike spatial smoothing described in other methods in the literature, our approach does not smear the faults and is very adaptive to those high dipping layers which are challenges to more conventional methods. We illustrate the effectiveness of this new method with a 3D field data example in which the signal to noise ratio is improved significantly; in particular, in regions of conflicting dips.