Seismic data processing via homomorphic filtering

Abstract

Homomorphic deconvolution, a filtering technique, is used to analyze real seismic data. It can separate the wavelet from the impulse train which is formed by convolution. For illustration purposes, four different signals, resulting from regional event, quarry blast, local blast, and a marine explosion, respectively, are considered. The complex cepstrum is also used to detect multiple echoes in a signal. The results show the effectiveness of the method.