Seismic event detection using three-component data

Presents an approach for the detection of seismic events using three-component data obtained from three orthogonally placed seismometers. In the first stage of processing, a detection parameter is computed as a function of the autocorrelation coefficients of the three data channels. This parameter is then input to a detector which uses the concept of comparing a short time average of the parameter against a delayed, long-term average of the parameter to produce a detector output. This algorithm is capable of detecting extremely weak signals of signal-to-noise ratios (SNRs) as low as -6 dB while keeping false alarms down to a minimum. Results obtained by using the algorithm on synthesized data as well as on field-acquired data are presented