Seismic data random noise attenuation using LLSP smoothing

Interpreting and studying reflection seismic data containing low levels of noise has become easier and more accurate. Random noise decreases the quality of seismic data considerably, and suppressing it is an important step in seismic data processing. In this study, we introduce a method called local least squares polynomial (LLSP) smoothing based on the Savitzky-Golay filter to diminish seismic random noise. This filter is based on smoothing the data fitting a curve in the least squares method which can eliminate random noise significantly. The LLSP smoothing has two main notable advantages. First, simple mathematics governing it which makes it convenient for data processing. Second, its excellent ability to preserve the signal waveform after application. The proposed method is applied on two synthetic models and real seismic data. For a more accurate investigation of the method's efficiency, the results are compared with the techniques: frequency offset deconvolution filter, wavelet transform, and singular value decomposition. The final results in both synthetic and real data show that the proposed method is as powerful as other well-known techniques for random noise attenuation, also the signal information is preserved during filtering.