Artificial Neural Networks for Seismic Data Interpretation
Abstract
This is the first Semiannual Technical Summary Report of the MIT lincoln Laboratory Artificial Neural Networks for Seismic Data Interpretation project. Seismic surveillance applications were reviewed and data interpretation tasks were selected for initial neural network experimentation. The selected tasks are estimation of signal arrival time (time picking), labeling of seismic phases, and recognition of typical and atypical events on a regional basis. Basic seismology and surveillance techniques are reviewed in this report and preliminary experimental results are summarized. We are using two types of data. Seismic waveform data with associated parametric information are being provided by SAIC in San Diego, CA. Parametric data for a much larger data set are being obtained by remote access to an on-line data base at the Center for Seismic Studies (CSS) in Arlington, VA. All the data are from NORESS and ARCESS arrays in Scandinavia and were processed by the IMS (Intelligent Monitoring System) regional seismic surveillance system. At the start of the contract SAIC provided an initial waveform data set for exploratory experimentation. While using it, our waveform data requirements and formats were worked out with SAIC and the first installment of waveforms for 50 events has now been received.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 30, 1990
- Accession Number
- ADA239673
Entities
People
- Michael C. Seibert
- Richard T. Lacoss
- Robert K. Cunningham
- Susan R. Curtis
Organizations
- Massachusetts Institute of Technology