A Wavelet Packet Approach to Transient Signal Classification

Abstract

Time–frequency transforms, including wavelet and wavelet packet transforms, are generally acknowledged to be useful for studying non-stationary phenomena and, in particular, have been shown or claimed to be of value in the detection and characterization of transient signals. In many applications time–frequency transforms are simply employed as a visual aid to be used for signal display. Although there have been several studies reported in the literature, there is still considerable work to be done investigating the utility of wavelet and wavelet packet time–frequency transforms for automatic transient signal classification. This paper contributes to this ongoing investigation through the development of a non-parametric wavelet packet feature extraction procedure which identifies features to be used for the classification of transient signals for which explicit signal models are not available or appropriate. Recent literature in this area is devoted to truly ad-hoc, high-dimensional, non-parametric types of classification in which one or more time–frequency transform forms the base from which a large number of features are determined by trial and error. In contrast, the wavelet-packet-based procedure presented in this paper was formulated to systematically adapt to any data dictionary within which several classes must be distinguished. This method is aimed at focusing the information in the data set to find the smallest number of features for robust, reliable classification. The promise of our method is illustrated by performing our procedure on a set of biologically generated underwater acoustic signals. For this example the wavelet-packet-based features obtained by our method yield excellent classification results when used as input for a neural network and a nearest neighbor rule.