Аннотация:The detecting of infrasonic signals from impulsivesources based on the mathematical model of their wavelet spectrum forms characteristic of signals from impulsive sources (explosions, volcanic activity, and others) is proposed. This model is based on wavelet spectra analysis of infrasonic signals from different sources.Modeling is based on morphological image analysis methods that are invariant to changes in signal recording conditions. The wavelet spectrum of a signal is a function of time and frequency,i.e., it depends on two arguments varying on a rectangular grid, and the value of this function (its module or the module of its real part)is considered as image brightness. The spectrum section corresponding to a signal from a source is approximated by a piecewise constant image, and the geometric form of its spots with the samebrightness determines the model of spectral images of signals from different sources. It is shown that, for different impulsive sources, the characteristic form of these spots is conserved and at the same time it significantly differs from the forms that are characteristic of the wavelet spectra of signals from other sources (microbaroms, mountain associated waves, and auroral infrasonic waves). A morphological method of searching for wavelet spectrum sections of signals that are characteristic of impulsive sources is proposed.