Two-Dimensional Signal Processing for Regional Seismic Event Identification.
Abstract
The objective of this research was to develop a methodology for regional seismic event identification which utilizes 2-dimensional signal processing and treats the seismogram as an image. The motivation for this research comes from the observation that the spectral features which distinguish quarry explosions from earthquakes can be seen throughout the entire seismogram. These features, which include spectral banding, are apparent in the time-frequency (TF) representation of the seismogram. We parameterized the TF representation by computing the wavenumber spectrum of the image and fitting this to a stochastic surface model developed to characterize bathymetric surfaces. Five model parameters control image features such as lineations, spatial orientation, characteristics lengths, and roughness of the surface. (MM)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1994
- Accession Number
- ADA292601
Entities
People
- Jay J. Pulli
- Paul S. Dysart