Inversion-based t-x Domain Signal-preserving Random Noise Reduction method
Y. Zhao, G. Li, J. Wei, B. Li, J. Wang and M. Wang
Event name: 78th EAGE Conference and Exhibition 2016
Session: Seismic Noise Attenuation
Publication date: 31 May 2016
Info: Extended abstract, PDF ( 1.11Mb )
Price: € 20
Random noise attenuation is a persistent problem in seismic exploration. There are many methods to attenuate random noise, among which the prediction filtering method is one of the most classical methods. However, prediction method is model-inconsistent and reduces the amplitudes of the signals. To overcome the shortcomings of the prediction method, we propose a novel approach to attenuate random noise in t-x domain through an inversion procedure. In this approach, a prediction error filter (PEF) is calculated from the noisy data. Then we use this filter as a constraint to the seismic data in order to get the denoised data by inversion. Since the PEF is calculated from the noisy data, it should be recalculated from the denoised seismic data. The process of calculating signal, then getting a new PEF should be iterated a few times. Besides, how to select the two trade-off parameters efficiently also has been discussed. Both synthetic and field data examples demonstrate excellent performance of the proposed approach.