Yu.N. Parshin - Dr. Sc. (Eng.), Professor, Head of Radio Devices Department, Ryazan State Radio Engineering University. E-mail: parshin.y.n@rsreu.ru
A.Yu. Parshin - Assistant, Radio Devices Department, Ryazan State Radio Engineering University. E-mail: alex90fox@gmail.com

Signals and images, obtained by radar, have a complex structure. It is necessary to find a distinguishing parameter for detection of signals and images at the noise background. High quality of obtained radar images provides possibilities of appliance of texture processing algorithms. It allows having great amount of subject in-formation. One of the detection methods is finding and usage of self-similarity properties or fractality in obtained signal. Measure of such properties is a fractal dimension in its different interpretations. Computing algorithms, making dimension estimation, are formed based on this parameter. Dimension estimation is used for comparison with threshold value and making decision about presence or absence of signal or object on image. Algorithms development is closely relate to simulating of different obtained data. One of the models of fractal signal is fractal Brownian motion and for images its two-dimensional analog is used - fractal Brownian surface. Their characteristics .are intensity and dimension , that depends on Hurst exponent. Usage of proposed method of estimation and signal detection provides variety of advantages over classic methods of detection and distinguishing: - given parameter is not depends on signal amplitude. It allows effective detection of low-in-contrast objects at the background or signals with low signal-noise ratio; - appliance of maximum likelihood method provides optimal estimation of fractal signal parameter. Proposed simulation method precise enough conforms to real structures of radar signals. Given expressions are effectively used for radar images processing.