Sliding Space-Frequency Processing of Finite Two-Dimensional Real Discrete Signals Based on the Fourier Transform with Variable Parameters

Abstract

In the field of Fourier processing of finite signals, three main directions of scientific research have been identified: Fourier processing of one-dimensional finite signals - processing of scalar functions of a scalar argument, Fourier processing of two-dimensional finite signals - processing of scalar functions of a vector argument, multichannel Fourier processing - processing of vector functions of a scalar argument. As part of the creation of a generalized theory of Fourier processing of finite signals, the authors proposed: the theory of spectral analysis of discrete signals at finite intervals in the bases of parametric exponential functions and the theory of two-dimensional digital signal processing in Fourier bases with variable parameters. The developed theories, generalizing the theory of Fourier processing of one-dimensional and two-dimensional signals, are based: on the introduction of new concepts of the shift of finite discrete signals in one-dimensional and two-dimensional cases and the introduction of new basic Fourier processing systems of discrete signals, which have the properties of multiplicativity, functions in the system. The mathematical apparatus of two-dimensional discrete Fourier transform with variable parameters in matrix and algebraic form is considered. A new method for processing finite two-dimensional real discrete signals in the spatial-frequency domain based on the discrete Fourier transform with variable parameters, the method of sliding spatial-frequency processing, has been introduced. An efficient method and algorithm for fast diagonal sliding spatial-frequency processing of finite two-dimensional real discrete signals based on the discrete Fourier transform with variable parameters has been developed. The estimation of the efficiency and effectiveness of the algorithm of the diagonal sliding two-dimensional discrete Fourier transform with variable parameters from the point of view of computational costs is carried out. As a result of experimental studies on model two-dimensional discrete finite signals, the validity, efficiency and reliability of the proposed method of sliding spatial-frequency processing of finite two-dimensional real discrete signals based on the discrete Fourier transform with variable parameters have been proved. A comparison (from the point of view of computational costs) of the developed method of sliding spatial-frequency processing of finite two-dimensional real discrete signals based on the discrete Fourier transform with variable parameters with the standard method of sliding processing of this type of signals is carried out.