Application of Neutral Networks to Seismic Signal Discrimination
Abstract
This is the first Annual Technical Summary of the West Virginia Institute of Technology Applications of Neural Networks to Seismic Classification project. The first year of research focused on identification and collection of a suitable database, identification of parametric representation of the time series seismic waveforms, and the initial training and testing of neural networks for seismic event classification. It was necessary to utilize seismic events that had a high degree of reliability for accurate training of the neural networks. The seismic waveforms were obtained from the Center for Seismic Studies and were organized into three smaller databases for training and classification purposes. Unprocessed seismograms are not well suited for presentation to a neural network because of the large number of data points required to represent a seismic event in the time domain. Parametric representation of the seismic waveform numerically extracts those features of the waveform that enable accurate event classification. Sonograms and moment feature extraction are two of the several transformations investigated for parametric representation of a seismic event. This parametric representation of the seismic events provides adequate information. Neural networks, Data points, Signal discrimination, Parametric representation, Seismic events.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 15, 1993
- Accession Number
- ADA276626
Entities
People
- Barbara Crist
- Don J. Smith
- G. A. Cipriani
- Heather Tanner
- J. J. Fuller
- James A. Cercone
- John D Martin
- Larry Mccutchan
- Stephen Goodman
- V. S. Foster
- W. M. Clark