A.V. Bogoslovskiy, E.A. Bogoslovskiy, D.S. Yudakov

In this article the preliminary image processing, in which samples of the impulse response of two-dimensional discrete filter is determined by solving a linear system of algebraic equations whose coefficients are elements of the matrix of the input signal (S-matrix). The latter are normalized coefficients of expansion of the energy spectrum of the input image in two-dimensional Fourier series, or that the same thing, splitting its autocorrelation coefficient. There are several known algorithms like image preprocessing: background suppression algorithm that is based on an analysis of the function processing efficiency, and an algorithm for selecting an object, which can be obtained by considering the filtration of the Wiener-Hopf applicative mixture of «object-background», allowing to provide an image of the object. Both algorithms are based on the function of efficiency, defined as a ratio of energy output and the input image of the object of interest, or background. The aim is to generalize the algorithms for image preprocessing based on the efficiency function, which is solved by introducing into consideration the function of the efficiency of a special type. Designed on the basis of its image preprocessing algorithm combines the advantages above mentioned algorithms and can be called modified. Examples of real image-processing algorithms and all three identified areas of their applicability.