Classification of sonar signals using Bayesian networks

Author

Larkin, Michael J.

Author_Institution

Naval Underwater Syst. Center, Newport, RI, USA

Volume

1

fYear

1997

fDate

2-5 Nov. 1997

Firstpage

855

Abstract

A pertinent issue in the development of automated classifiers is the combination of information extracted from the data with a priori information that we may have about the classes in questions. Incorporating prior knowledge is particularly useful if the data is incomplete or imprecise. Bayesian networks offer an ideal representation for the combination of a priori knowledge with data. This paper discusses the implementation of a Bayesian network classifier for the purpose of classifying underwater mines, using knowledge that is known by experts (sonar operators) about the distinguishing characteristics of mines, such as size, shape, shadow and resonance.

Keywords

Bayes methods; expert systems; feature extraction; military computing; object detection; pattern classification; sonar signal processing; unsupervised learning; Bayesian network classifier; a priori information; automated classifiers; expert knowledge; feature extraction; incomplete data; resonance; shadow; shape; size; sonar signals classification; underwater mines; unsupervised methods; Acoustic reflection; Acoustic signal detection; Bayesian methods; Data mining; Object detection; Resonance; Shape; Sonar detection; Underwater tracking; Working environment noise;

fLanguage

English

Publisher

ieee

Conference_Titel

Signals, Systems & Computers, 1997. Conference Record of the Thirty-First Asilomar Conference on

Conference_Location

Pacific Grove, CA, USA

ISSN

1058-6393

Print_ISBN

0-8186-8316-3

Type

conf

DOI

10.1109/ACSSC.1997.680564

Filename

680564