Hodovychenko M. Semantic scenarios model based on Petri nets for intelligent video surveillance.

The paper presents the semantic model for representing of scenarios in the video stream by adapting Petri nets to work with video. To be able to take into account the complex temporal relations between objects, a model representation of temporal relations between objects through the use of elements of Petri net for representing temporal logic of relations has been further developed. Semantic model to simulate scenarios based on complex temporal and spatial relationships between individual events and objects has been developed. To work with probabilistic video data, a combined semantic model representations of scenarios based on adapted Petri nets and a probabilistic model of recursive Bayesian filter 20 evaluation and particles were proposed. The resulting combined semantic model allowed to carry out further development of structural and statistical recognition scenarios that could substantially increase the reliability of authentication scenarios in intelligent video surveillance.