Comparison of Ground-Roll Attenuation Methods for Near Surface Investigations

In this work four methods for Ground-Roll attenuation in a near surface data case are compared. The methods presented are: simple band-pass filtering, time-offset variant filtering based on the Short-Time Fourier-Transform, f-k filtering and time-offset variant filtering in the wavelet domain computed by means of the Continuous Wavelet Transform (CWT). The experience carried out demonstrates that the wavelet filter performs better than the others methods. This is because the transformed domain provided by the CWT is able to offer an improved representation of the frequency components of the signal making it easy to identify the GR and to draw a time-offset-frequency filter. The outcomes of the methods are assessed on both shot and stack domain.