Classification of acoustic and seismic signals based on the statistics of their wavelet sub-band coefficients

Statistical characterization of acoustic signals is a pre-processing technique aiming at the definition of signal observables, which can be used as input data in the formulation of inverse problems of acoustical oceanography aiming at the estimation of critical parameters of the marine environment. Statistical signal characterization is also a means to classify the signals and compare their properties without reference to any physical model that determines the signal observables. Thus, it can in principle be used for the comparison of the signals of any type, including seismic recordings, thus opening the area to many applications of geophysical monitoring, using signals of any type. The paper summarizes the first attempts to study the efficiency of a signal characterization method based on wavelet transform of the signal at various levels, followed by the statistical description of their sub-band coefficients. It is shown that A-stable symmetric distributions are capable of defining the statistics of these coefficients for signals used in applications of ocean acoustic tomography and geoacoustic inversions. It is further investigated if these distributions are capable of defining the statistics of seismic signals and it is shown by preliminary results, that they can indeed be considered as possible candidates in this respect.