Detecting Local Characteristic of Analyzed Signals and Processes Using Wavelet Transformation
Abstract
Purpose: Developing a method of detecting changes in local characteristics of signals and processes using wavelet transformations in the context of concerns of applied problems of information processing and system dynamics analysis. Methods: Using invariance under shift and scaling, along with frequency-time localization. Results: A formalized description is given for the method of detecting local characteristics of arbitraty signals using multiresolution wavelet analysis. The main operations of the signals processing when detecting their local characteristics are the analysis of small-scaled wavelet coefficients and the determination of the signal-describing function correlation interval. Local maxima lines along the scale axis on the time-frequency plain unite the points in which the modules of the coefficients of the wavelet transformation have local maxima. These lines converge to certain points with singularities. Studying how the local wavelet transformation maxima change at different scales help determine the location and nature of the local singularities of the analyzed signals and processes. An example is given of using a wavelet transformation for spectral images of observed objects for choosing the most informative spectral channels in systems of multispectral remote flexing of the Earth. Practical relevance: The presented mathematical description of the proposed method can be a base of processing algorithms for signals and processes. The method also can be used in the processing of data obtained by information collection systems of remote flexing, telemetries or control over technological processes.Published
2015-02-20
How to Cite
Kozinov, I. (2015). Detecting Local Characteristic of Analyzed Signals and Processes Using Wavelet Transformation. Information and Control Systems, (1), 21-28. https://doi.org/10.15217/issn1684-8853.2015.1.21
Issue
Section
Information processing and control