Estimation of polarization parameters using time–frequency representations and its application to waves separation - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Signal Processing Année : 2006

Estimation of polarization parameters using time–frequency representations and its application to waves separation

Antoine Roueff
Jerome I. Mars

Résumé

This paper deals with the detection of polarized seismic waves, the estimation of their polarization parameters, and the use of these parameters to apply waves separation. The data, containing several polarized waves together with some noise, are recorded by two-component sensors. After a review presenting the tools classically used to estimate the polarization parameters of a wave in the time domain and in the time–frequency domain, respectively, we present a new methodology to detect polarized waves, and estimate their polarization parameters automatically. The proposed method is based on the segmentation of a time–frequency representation of the data. In addition, after describing the proposed polarization estimation method, we present the oblique polarization filter (OPF) that enables the separation of two polarized waves using their polarization parameters, even if the corresponding patterns partially overlap in the time–frequency plane. The OPF consists in applying phase shifts, rotations, and amplifications in order to project one wave on one single component and the other wave on the other component. Being more efficient than classical polarization estimation methods, our approach greatly increases the separation performances of the OPF. Results are presented both on synthetic and real seismic data.

Dates et versions

hal-00105053 , version 1 (10-10-2006)

Identifiants

Citer

Antoine Roueff, Jocelyn Chanussot, Jerome I. Mars. Estimation of polarization parameters using time–frequency representations and its application to waves separation. Signal Processing, 2006, 86, pp.3714-3731. ⟨10.1016/j.sigpro.2006.03.019⟩. ⟨hal-00105053⟩

Collections

UGA CNRS LIS
90 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More