Premium
Rayleigh wave separation using high‐resolution time‐frequency polarization filter
Author(s) -
Kazemnia Kakhki Mohsen,
Mansur Webe J.,
Peters Franciane C.
Publication year - 2020
Publication title -
geophysical prospecting
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.735
H-Index - 79
eISSN - 1365-2478
pISSN - 0016-8025
DOI - 10.1111/1365-2478.12994
Subject(s) - polarization (electrochemistry) , time–frequency analysis , seismogram , frequency domain , polarizing filter , signal processing , time domain , rayleigh wave , computer science , geology , seismic wave , acoustics , dispersive body waves , economic geology , love wave , algorithm , filter (signal processing) , optics , seismology , surface wave , physics , wave propagation , radar , telecommunications , computer vision , longitudinal wave , mechanical wave , chemistry , optical filter , tectonics , telmatology
Obtaining geological information from seismic data motivates researchers to innovate and improve seismic wave processing tools. Polarization‐based methods have received much attention regarding their ability to discriminate between different phases of the seismic wave based on polarity. Combining the intuitive definition of polarity in the frequency domain (monochromatic waves) with the non‐stationary properties provided by time‐domain methods, time‐frequency approaches are attracting widespread interest because they localize the information extracted from the seismic waves in the joint time and frequency domains. Due to the lack of high‐resolution time‐frequency maps, the time‐frequency polarization approach was not able to resolve specific temporal polarity changes in the seismic signal. The main objective of this study was to devise a robust time‐frequency‐based polarization filtering method using high‐resolution polarization attributes obtained directly from the sparse time‐frequency map without using Eigen analysis or analytic signals. The method proposed here utilizes a computationally effective sparsity‐based adaptive S‐transform to obtain a high‐resolution polarization map of an inherently non‐stationary seismogram for the entire frequency content of the signal at different times. The superiority of the proposed method over the S‐transform method was verified using synthetic and real data sets to calculate the polarization attributes in the time‐frequency domain and separate the Rayleigh waves from the seismogram.