Robustness of the Algorithm of Identification of the Type of Dynamic Object Found at the Finite Sequence of 2D Background Frames of the Optoelectron Device
- Autores: Katulev A.1, Sotnikov A.2, Kemaykin V.1, Kozhukhin I.1
-
Afiliações:
- State Technical University
- Joint Supercomputer Center of the Russian Academy of Sciences, Branch of Scientific Research
- Edição: Volume 40, Nº 12 (2019)
- Páginas: 2062-2076
- Seção: Article
- URL: https://journals.rcsi.science/1995-0802/article/view/206455
- DOI: https://doi.org/10.1134/S1995080219120060
- ID: 206455
Citar
Resumo
Here are proposed robustness characteristics of the algorithm of identification of the type of the dynamic object (DO) and the law of probability distribution of identification sufficient statistics, formed by the algorithm under prior uncertainty. The law is applied for verification and validation of the algorithm. Wavelet fractal correlation algorithm (WFCA) implements vectorial criterion of ratio of likelihood functions of simple alternative hypotheses—types of DOs, this criterion being invariant to specific features of DO motion trajectories. The likelihood functions are reconstructed by simulation according to sufficiently representative complexes of implementations of fractal dimensions, energies, wavelet spectra and maximum eigenvalues of biased correlation matrices as functional of the measured coordinates of spatial attitude of various types of real DOs located by the optoelectron device (OED). The simulation proved robustness and high efficiency of the algorithm of identification of the type of DOs.
Palavras-chave
Sobre autores
A. Katulev
State Technical University
Email: kozhukhin@mail.ru
Rússia, Tver, 170023
A. Sotnikov
Joint Supercomputer Center of the Russian Academy of Sciences, Branch of Scientific Research
Autor responsável pela correspondência
Email: asotnikov@jscc.ru
Rússia, Moscow, 119334
V. Kemaykin
State Technical University
Autor responsável pela correspondência
Email: vk-kem@mail.ru
Rússia, Tver, 170023
I. Kozhukhin
State Technical University
Autor responsável pela correspondência
Email: kozhukhin@mail.ru
Rússia, Tver, 170023