Article,
Signal detection using multi-channel seismic data
Bulletin of the Seismological Society of America, 86 (1A): 221--231 (Feb 1, 1996)
Abstract
We outline a simple signal detection approach for multi-channel seismic
data. Our approach is based on the premise that the wave-field spatial
coherence increases when a signal is present. A measure of spatial
coherence is provided by the largest eigenvalue of the multi-channel
data's sample covariance matrix. The primary advantages of this approach
are its speed and simplicity. For three-component data, this approach
provides a more robust statistic than particle motion polarization.
For array data, this approach provides beamforming-like signal detection
results without the need to form beams. This approach allows several
options for the use of three-component array data. Detection statistics
for three-component, vertical-component array, and three different
three-component array approaches are compared to conventional and
minimum-variance vertical-component beamforming. Problems inherent
in principal-component analysis (PCA) in general and PCA of high-frequency
seismic data in particular are also discussed. Multi-channel beamforming
and the differences between principal component and factor analysis
are discussed in the appendix.
