Both Correlation and Morphological Methods of Detecting a Specified Acoustic Signal Propagating Through the Atmosphere
- PDF / 1,576,955 Bytes
- 9 Pages / 547.087 x 737.008 pts Page_size
- 18 Downloads / 150 Views
Pure and Applied Geophysics
Both Correlation and Morphological Methods of Detecting a Specified Acoustic Signal Propagating Through the Atmosphere NADEZDA D. TSYBULSKAYA,1 SERGEY N. KULICHKOV,1,2 ALEXEY I. CHULICHKOV,1,2 VITALIY G. PEREPELKIN1 Abstract—Two methods of detecting quasiperiodic signals— the well-known method of calculating correlation coefficients and the method of a morphological analysis that makes it possible to detect signals with variable phases, which is important in isolating signals at long distances from their sources—are compared. The efficiency of a morphological analysis is in solving the problems of isolating quasiperiodic signals against the background of noise that is comparable to or exceeds (in amplitude) the signal under study. These two methods are compared using subwoofer signals recorded in August 2009 at the Scientific Station of the A.M. Obukhov Institute of Atmospheric Physics. The signal frequencies were chosen within a range of 45–130 Hz and the distance to the source was from 0 to 1255 m. Signals were recorded during night and morning hours, which made it possible to study the potentialities of these methods in isolating a desired signal under both disturbed and undisturbed states of the atmosphere (in the absence or presence of noise of different origins). Keywords: Signal extraction, acoustics, infrasound, morphological analysis, correlation analysis.
PACS: PACS 93.85.Bc, PACS 92.10.Yb, PACS 43.60.?d.
1. Introduction In a number of problems of analyzing infrasound sources, it is necessary to isolate fragments of a signal with a characteristic form determined by its source. Such problems arise in recognizing explosion sources (Campus 2009), detecting atmospheric fronts (Kulichkov et al. 2017), etc. The problem of isolating such signal fragments is hampered by the fact that the
1
Obukhov Institute of Atmospheric Physics, Russian Academy of Sciences, Moscow 109017, Russia. E-mail: [email protected] 2 Moscow State University, Moscow 119991, Russia.
IGOR P. CHUNCHUZOV,1 and
amplitudes and phases of signals propagating through the atmosphere are significantly distorted; moreover, noise, whose amplitude may exceed the amplitude of the signal under study, is superimposed on this signal. Signals are usually recorded by two channels, and correlations between the fragments of the signals recorded by these two channels are found. One of the widely used methods of detecting signal fragments similar in form is the correlation method that allows one to isolate fragments that differ from standard ones by a linear amplitude transformation (Le Pichon and Cansi 2003): the modulus of the coefficient of correlation between two signals f(t) and g(t) reaches its maximum, if f ðtÞ ¼ kgðtÞ. However, the real distortions of signals propagating through the atmosphere are not linear. In such cases, more complicated morphological methods yield good results (Pyt’ev and Chulichkov 2016; Kulichkov et al. 2011; Chulichkov et al. 2019). In this work, both correlation and morphological methods of d
Data Loading...
