Polarization Filtering in Time-frequency Domain Based on Generalized S Transform

摘要

We propose an adaptive covariance polarization filtering in time-frequency domain based on the generalized S transform.On one hand,we structure the cross-energy matrix in time-frequency domain using the generalized S transform,and compute instantaneous polarization attributions through eigenanalysis.On the other hand,we design filtering algorithm in time-frequency domain,achieve multi-component seismic polarization filtering.The specialties of this method are that the length of time window of covariance matrix is determined by instantaneous frequency of the multi-component seismic data,it can adaptive with the dominant period of the desired signal.Meanwhile,it calculates polarization parameters at each time-frequency point and no longer need to undertake the interpolation.The processing results of model and real three-component coalfield exploration Seismic data show that this method has the very good visual,high resolution and practicability.