Application of continuous wavelet transforms for swell noise attenuation
Gustavo Catão Alves, André Bulcão, Bruno José Nogueira Dias, Djalma Manoel Soares Filho, Felipe Prado Loureiro · 2019
Ocean-surface related noises known as swell noises are a common occurrence in marine surveys and present a challenge for traditional noise suppression techniques, due to the non-random nature of the events.In order to identify this type of noises, we can use differences between real and modeled data sets.This is particularly useful in the context of full waveform inversion, where both types of data are available.Here, we apply a continuous wavelet transform (CWT) to real and modeled data to attenuate swell noise events by estimating filters from the modeled data using a smooth velocity model and applying those to the real data.We show that filters using the CWT domain can efficiently attenuate swell noise by correctly estimating the time windows and wavelet scales of desired data.