Application of cross-term deleted Wigner representation (CDWR) for sonar target detection/classification

Author

Kadambe, Shubha ; Adali, Tulay

Author_Institution

HRL Labs., Malibu, CA, USA

Volume

1

fYear

1998

fDate

1-4 Nov. 1998

Firstpage

822

Abstract

The application of CDWR for sonar target detection/classification is described. For the detection of targets, the CDWR a time-frequency representation is applied. The selected features are extracted in the CDWR domain. These features are used to train a probabilistic network based classifier to achieve three-way classification of target, no-target and clutter. The experimental details and the results obtained on the data supplied by the US Navy are provided. The target detection results are compared with the detector that is currently being used by the US Navy (NAWCAD). The results indicate that the detector based on CDWR performs better than the algorithm that is currently being used by the US Navy.

Keywords

Wigner distribution; clutter; feature extraction; learning (artificial intelligence); military systems; probability; signal classification; signal representation; sonar detection; time-frequency analysis; NAWCAD; US Navy; algorithm; clutter classification; cross-term deleted Wigner representation; experimental details; no-target classification; probabilistic network based classifier training; sonar target detection/classification; target detection results; time-frequency representation; Acoustic beams; Biomedical signal processing; Data mining; Detectors; Helicopters; Object detection; Signal processing; Sonar applications; Sonar detection; Time frequency analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Signals, Systems & Computers, 1998. Conference Record of the Thirty-Second Asilomar Conference on

Conference_Location

Pacific Grove, CA, USA

ISSN

1058-6393

Print_ISBN

0-7803-5148-7

Type

conf

DOI

10.1109/ACSSC.1998.750975

Filename

750975