Blind Deconvolution to Improve Classification of Transient Source Signals in Multipath

Abstract

General testing of several blind deconvolution techniques has indicated that one algorithm stands out as being the best choice for transient classification in shallow water. This is Cabrelli's algorithm, which was developed for oil and gas exploration in the seismic community. This work describes the algorithm and explores its applicability. Three aspects of performance were tested using simulation studies. The algorithm's ability to handle complicated transient signals was explored. Its ability to handle complex shallow-water environments was tested. Finally, its ability to perform adequately in noise was investigated. Studies using correlation coefficient comparisons of Cabrelli output to ground truth showed that the algorithm can handle complex environments, although it might not perform as well with complicated transient signals. Classification operating characteristics (CLOC) analysis showed that Cabrelli's algorithm performs adequately in noise.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 13, 2000
Accession Number
ADA377973

Entities

People

  • George B. Smith
  • Lisa A. Pflug
  • Michael K. Broadhead

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Acoustic Propagation
  • Acoustics
  • Algorithms
  • Autocorrelation
  • Bandwidth
  • Classification
  • Coefficients
  • Convolution
  • Equations
  • Frequency
  • Hydrophones
  • Image Restoration
  • Military Research
  • Reflection
  • Shallow Water
  • Underwater Acoustics
  • Water

Readers

  • Acoustical Oceanography.
  • Computational Modeling and Simulation
  • Wave Propagation and Nonlinear Chaotic Dynamics.