The Application of Wiener Optimum Deconvolution Filter (WODF) with Frequency Analysis to Suppress Ground-roll in Seismic Data

Abstract

Basically, ground-roll noise causes signal quality information decrement in land seismic data. Ground-roll is the main type of Rayleigh-wave that is characterized with low velocity, low frequency and high amplitude. Separating them by using frequency filtering, like (f,k) filtering, causes signal distortion due to overlapping signal with ground-roll in time and frequency domain. To solve this problem, we proposed a redesigned Optimum Wiener filter method. The basis of the proposed method is estimating ground-roll in contaminated data by using a reference noise trace, sweep signal. By Wiener filtering, the reference noise is then adjusted to match the ground-roll noise in contaminated data. The adjusted reference noise then subtracted from the seismic trace to obtain the signal. In this research, we also give objective analysis to find the best parameter in maximizing the obtained result whereas the frequency interval is likely the most influenced parameter in Wiener filtering. The obtained results show that Wiener filtering can suppress ground-roll without disturbing the signal while the signal is overlap with ground-roll.