Adaptive window-length detection of underwater transients using wavelets

Abstract

This paper describes a detection method that adapts to unknown characteristics of the underlying transient signal, such as location, length, and time-frequency content. It applies a set of embedded detectors tuned to a number of signal partitions. The detectors are based on the wavelet theory, whereby two different techniques are examined, one using local Fourier transform and the other using discrete wavelet transform. The detection statistics are computed so as to enable prewhitening of unknown colored noise and to allow for a constant false-alarm rate detection. An adapted segmentation of the signal is next obtained with a goal of finding the largest detection statistics within each segment of the partition. The detectors are tested using several underwater acoustic transient signals buried in ambient sea noise.

Autocorrelation based denoising of manatee vocalizations using the undecimated discrete wavelet transform

“…In the last two decades, the wavelet transform has emerged as an alternative to the short-time Fourier transform ͑STFT͒ for time-frequency analysis of nonstationary signals and has been used in analyzing transient signals in sonar applications ͑Chen et al, 1998;Carevic, 2005͒. In addition, several researchers including Learned and Willsky ͑1995͒, Bailey et al ͑1998͒, Huynh et al ͑1998͒, and recently, Adam ͑2006͒ have implemented the wavelet transform in favor of the STFT in feature extraction, detection, and classification of marine mammal vocalizations.…”

Classification of underwater transient signals using MFCC feature vector

“…The problem of detection and classification of transient signals in the ambient noise is of great importance in areas such as underwater acoustics, seismology, and condition monitoring [1][2][3]. A particular example of this application is an underwater surveillance system, where it is required to determine whether the detected transients are from biologic sources like whale or artificial sources like ships and submarines.…”