Nonstationary signal classification using pseudo power signatures:The matrix SVD approach

摘要

This paper deals with the problem of classification ofnnonstationary signals using signatures which are essentially independentnof the signal length. This independence is a requirement in commonnclassification problems like stratigraphic analysis, which was anmotivation for this research. We achieve this objective by developingnthe notion of an approximation to the continuous wavelet transform,nwhich is separable in the time and scale parameters, and using it tondefine power signatures, which essentially characterize the scale energyndensity, independent of time. We present a simple technique which usesnthe singular value decomposition to compute such an approximation, andndemonstrate through an example how it is used to perform thenclassification. The proposed classification approach has potentialnapplications in areas like moving target detection, object recognition,noil exploration, and speech processing