Regresar

On using spectral estimation for feature extraction of microseisms at cotopaxi volcano

Abstract:

This paper presents a feature extraction method by using spectral analysis techniques of microseisms, this allows to classify events correctly due to microseisms are the most important sources of information in order to launch an effective early warning in case of an imminent volcanic eruption to safeguard human lifes. We work based on a database contains microseisms representing volcano-tectonic, long period, hybrids, and tremors, which were previously labeled by the Instituto Geofísico de la Escuela Politécnica Nacional, data obtained from Cotopaxi volcano with a broadband seismometer. The results, obtained by analyzing the power spectral density, have shown three main features that allow to discriminate among these types of events with an accuracy of 93%.