A Comparison of Post-stack Results after Filtering of Harmonic Noise Using Two Filter Methods

Abstract

Harmonic noise in seismic data is a single frequency noise caused by power line within the data acquisition systems and generally filtered by classical notch filter. However, the notch filter would attenuate all recorded data around notch frequency at especially 50 Hz (or 60 Hz) and so leads to loss available information that explains fine details in the seismic data. On the other hand, iterative trimmed and truncated mean filter method locally applied in spectral domain (LITTM) is a signal protector filter and is very effective to remove the harmonic noise in the signal. In this study we proposed to compare the quality of migrated sections after removing the harmonic noise by using classical notch and LITTM filters. We tested the applicability and reliability of the LITTM filter in removing the single frequency harmonic noise on synthetic data set and then applied it to a marine data from Turkey. When comparing the outputs from classical notch and LITTM filters, we found that the LITTM filter strikes a balance between the removing harmonic noise and protecting available frequencies. Our experiments show that the LITTM filter can effectively be used to eliminate other spike-like noise within the bandwidth of the data.