Premium
Ground‐roll attenuation using a 2D time‐derivative filter
Author(s) -
Melo Paulo E. M.,
Porsani Milton J.,
Silva Michelângelo G.
Publication year - 2009
Publication title -
geophysical prospecting
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.735
H-Index - 79
eISSN - 1365-2478
pISSN - 0016-8025
DOI - 10.1111/j.1365-2478.2008.00740.x
Subject(s) - seismogram , attenuation , geology , filter (signal processing) , second derivative , amplitude , normal moveout , deconvolution , synthetic seismogram , seismology , mathematical analysis , algorithm , acoustics , geodesy , mathematics , optics , computer science , physics , anisotropy , computer vision
We present a new filtering method for the attenuation of ground‐roll. The method is based on the application of a bi‐dimensional filter for obtaining the time‐derivative of the seismograms. Before convolving the filter with the input data matrix, the normal moveout correction is applied to the seismograms with the purpose of flattening the reflections. The method can locally attenuate the amplitude of data of low frequency (in the ground‐roll and stretch normal moveout region) and enhance flat events (reflections). The filtered seismograms can reveal horizontal or sub‐horizontal reflections while vertical or sub‐vertical events, associated with ground‐roll, are attenuated. A regular set of samples around each neighbourhood data sample of the seismogram is used to estimate the time‐derivative. A numerical approximation of the derivative is computed by taking the difference between the interpolated values calculated in both the positive and the negative neighbourhood of the desired position. The coefficients of the 2D time‐derivative filter are obtained by taking the difference between two filters that interpolate at positive and negative times. Numerical results that use real seismic data show that the proposed method is effective and can reveal reflections masked by the ground‐roll. Another benefit of the method is that the stretch mute, normally applied after the normal moveout correction, is unnecessary. The new filtering approach provides results of outstanding quality when compared to results obtained from the conventional FK filtering method.