A.N. Katulev, A.A. Khramichev, S.V. Jagolnikov
The method for detection of dynamical objects on the idea of shoring independent sampled statistics in the form of fractal dimension and maximum own meaning of autocorrelation matrix, calculated for current sequence of repeated 2D images about the state of controlled space by passive information means is developed. Estimations for moments of statistics are given. Is stated that maximal own quantity is well information statistics.
The structure of too-channel algorithm is synthesized by criteria of multi-alternative identification of hypothesis with a posteriori accuracy and similar simple hypothesis.
Mathematical expression for statistical decision rules of discovery of object on the 2D image are elaborated. Neyman-Pirson’s criterion is basis for these rules.
In is state by modeling that the application of the method for fractal-statistical detection of the dynamic object on 2D image allows it to increase the efficiency of optical passive means for control of the given space and detection probability of passive aerial objects at to expense of introduction of the additional feature – fractal dimensionality.