The use of Continuous Wavelet Transform for Ground-Roll attenuation

This work deals with the use of the Continuous Wavelet Transform (CWT) for the attenuation of Ground-Roll (GR) noise in seismic reflection land data. The CWT is a multi-resolution method that yields a time-frequency (scale) representation of a signal. With respect to classical Fourier-based approach, such as the Short Time Fourier Transform (STFT), the CWT is not limited by the choice of a user-defined fixed-duration window function, thus it is not limited by a fixed time-frequency resolution. This property is of particular interest for the analysis of non-stationary signal such as GR. The transformed domain provided by the CWT (the wavelet spectrum) is able to offer an improved representation of the frequency components of a signal, making it easy to identify the GR and to draw a time-frequency filter aimed at removing it. This work also shows that the filter design has a simple implementation and it allows for the application of complex-shape filters that are hardly or not applicable for Fourier-based methods. The proposed CWT filtering is applied to near surface and deep seismic reflection land data yielding quite satisfactory results, which are illustrated in both the shot gather and stack domain.