FIELD: measuring technology.
SUBSTANCE: invention relates to the field of measurement technology, namely to extracting a useful signal from an input signal containing both a useful component and noise in various signal filtering systems under conditions of a priori uncertainty. To do this, when filtering the input signal in the frequency range corresponding to the spectrum of the useful component of the signal, for each quantization level in the input signal, the period of the sequence of such sections of the implementation of a random process that are unrelated to the useful component is determined, with the help of which, relative to the initial cross-section of the implementation of a random process, as which the cross-section of the implementation of a random process in the moment of time corresponding to the beginning of the observation of the input signal is determined by such sections of the implementation of a random process, which are not related to the useful component and are used as a marker central to the area (window) masking the time regions of the input signal corresponding to noise, the totality of which forms noise, and the method for subtraction of spectra (spectral subtraction) is implemented in relation to the input signal.
EFFECT: increase in the efficiency of separating the useful component from the input signal containing the useful component and noise, which are in the same frequency range, by reducing or completely eliminating the distortion of the shape of the useful signal simultaneously with an increase in the signal-to-noise ratio.
3 cl, 2 dwg
Authors
Dates
2022-07-29—Published
2021-08-25—Filed