Aplicación de la teoría de ondículas para analizar señales sísmicas en el espacio tiempo-frecuencia

A seismic signal can be represented like the sum of geologic signal and noise. The noise causes problems in the geophysical data processing because it limits the identification of the geologic characteristics. In this sense, it is important to reduce the associated noise to the measurement trying to preserve the useful information. In Geophysical, the deconvolucion is the technique commonly used in the seismic data processing; this technique is based on the theory of filter of Wiener-Levinson and is used to attenuate or to eliminate manifolds or reverberations that appear like noise in the signal. In this work a technique is presented that increases the performance of the Wiener filter in the dominion of the frequency by means of the application of a mixed scheme based on the technique of thresholding of the coefficients of the Transformed of wavelet and the Wiener filtrate in the dominate of wavelet. In order to evaluate the performance of the proposal, the results are compared in the dominion of the frequency with the classic Wiener filtrate.