Seismic signal discrimination between earthquakes and quarry blasts using fuzzy logic approach

Seismic analysts identify earthquakes signals from those of explosions by visual inspection and calculating some characteristics of seismogram. Such work supposes a great deal of workload for seismic analysts. Therefore, an automatic classification tool reduces dramatically this arduous task, turns classification reliable, removes errors associated to tedious evaluations and changing of personnel. In this present paper we are interested in transforming the analysts' knowledge of classifying seismograms into an automated soft classification system based on fuzzy logic. This is primarily due to its capability of modelling human reasoning and decision-making, managing complexity and controlling computational cost. These capabilities are essential for manipulating high dimensionality and complexity of seismic signal. To conduct effective discrimination, relevant seismogram characteristics are extracted based on human experience. Using these characteristics, a fuzzy classifier is built and tested with real seismic data. The results are found to be encouraging.