2D Stationary Wavelet Transform Based Acquisition Footprint Suppression

Acquisition of seismic data over a large 3D survey acquired in Mexico is constrained by both cultural and ecological limitations, resulting in strong acquisition footprint that contaminates the target turbidite reservoir of interest. Due to the acquisition obstacles, the source and receiver grid is quite irregular, such that we cannot suppress acquisition footprint through simple kx-ky filtering of time slices. In this survey, the most vexing components of acquisition footprint are due to leakage of backscattered ground roll into the migration-stack and migration artifacts. 2D Wavelet Transforms provide a spatially varying filter that better adapts to the irregular acquisition geometry. We find that 2D Stationary Wavelet Transform (2D SWT) based filters applied to seismic time slices allow us to suppress both acquisition footprint and random noise., while preserving geologic discontinuities of interest. We decompose each seismic time slice into five levels of wavelet components that represent progressively coarser details. In the shallow section, acquisition footprint is strong and geologic structure is weak. We therefore examine successive levels (or panels) to determine where the acquisition footprint lies. Once identified, we suppress these components in the data reconstruction. We find that 2D SWT filtering on time slices allows us to suppress backscattered ground roll, as well as migration artifacts that leak through the seismic processing. We evaluate the efficacy of this processing through the use of geometric attribute imaging of the turbidite system.