Title :
Detection of tonals in lofargrams using connectionist methods
Author :
Weber, D.M. ; Krüger, C.C.
Author_Institution :
Dept. of Electr. & Electron. Eng., Stellenbosch Univ., South Africa
Abstract :
A method for enhancing and detecting tonals detected by passive sonar systems is presented. The detection process is performed in the frequency domain using contour-enhancing neural networks for preprocessing and multi-layer perceptrons for the actual detection. This method enables reliable detection of tonals at a signal to noise ratio of -17 dB, with an accuracy of 95%. The accurate detection of tonals results in a reduction of information to be processed by higher level classification procedures
Keywords :
feedforward neural nets; signal detection; sonar; classification procedures; connectionist methods; contour-enhancing neural networks; frequency domain; lofargrams; multi-layer perceptrons; passive sonar systems; signal to noise ratio; tonals; Acoustic signal detection; Africa; Electronic mail; Frequency domain analysis; History; Neural networks; Signal resolution; Signal to noise ratio; Sonar detection; Speech analysis;
Conference_Titel :
Neural Networks, 1993., IEEE International Conference on
Conference_Location :
San Francisco, CA
Print_ISBN :
0-7803-0999-5
DOI :
10.1109/ICNN.1993.298806
