Automated video surveillance system based on hierarchical object identification

The structure, data models and implementation of the automated video surveillance system are proposed in this paper. The developed methods and basic information technologies for contour segmentation based on the Canny's method, and the methods of contour inflection points selection with the adjustable detailing allow us to carry out structural and statistical identification of hierarchical objects by their contour inflection points. Using the technologies proposed by the authors, the automated video surveillance system can identify objects with different levels of detailing. This allows us to increase its efficiency under the conditions of limited computing resources.