Automatic processing of seismic events recorded on a mini-array Signal analysis combined with neural networks

A. Bottero, Y. Cansi, B. Massinon


We present a new method for automatic processing of mini-array records of regional events. It is based on a comprehensive analysis of the cross-correlation functions. This leads to a set of time-delays used to compute the azimuth and velocity of the travelling wave only in case of consisteney of the time-delay set. The second step takes into account the time-frequency representations of these wave parameters to identify each regional wave using a neural network. The resulting standard error on azimuth is 3° and the relative error on distance is less than 20%.


neural networks;mini-array;automatic seismic processing

Full Text:




Published by INGV, Istituto Nazionale di Geofisica e Vulcanologia - ISSN: 2037-416X