Denoising of magnetotelluric signals by polarization analysis in the discrete wavelet domain. (March 2017)

• Record Type: Journal Article
• Authors: Carbonari, R.; D'Auria, L.; Di Maio, R.; Petrillo, Z.
• Is Part Of: Computers & geosciences. Volume 100(2017)
• Page Start: 135
• Page End: 141
• Publication Date: 2017-03
• Subjects: Magnetotelluric signal -- Discrete wavelet transform -- Data denoising -- Polarization analysis; Environmental policy -- Periodicals; 550.5
• DOI: 10.1016/j.cageo.2016.12.011
• Languages: English
• ISSNs: 0098-3004

Abstract

Abstract: Magnetotellurics (MT) is one of the prominent geophysical methods for underground deep exploration and, thus, appropriate for applications to petroleum and geothermal research. However, it is not completely reliable when applied in areas characterized by intense urbanization, as the presence of cultural noise may significantly affect the MT impedance tensor estimates and, consequently, the apparent resistivity values that describe the electrical behaviour of the investigated buried structures. The development of denoising techniques of MT data is thus one of the main objectives to make magnetotellurics reliably even in urban or industrialized environments. In this work we propose an algorithm for filtering of MT data affected by temporally localized noise. It exploits the discrete wavelet transform (DWT) that, thanks to the possibility to operates in both time and frequency domain, allows to detect transient components of the MT signal, likely due to disturbances of anthropic nature. The implemented filter relies on the estimate of the ellipticity of the polarized MT wave. The application of the filter to synthetic and field MT data has proven its ability in detecting and removing cultural noise, thus providing apparent resistivity curves more smoothed than those obtained by using raw signals. Highlights: Discrete Wavelet Transform (DWT) is first proposed for denoising of MT data. A polarization filter based on DWT is developed for cultural noise removal. The choice of suitable thresholds determines the selectivity of the filter. Reliable estimates of resistivity values from filtered MT field data are obtained.