Detail výsledku
Selected Video-Processing Methods and System Architecture Design
Bartík Vladimír, Ing., Ph.D., UIFS (FIT)
Bartl Vojtěch, Ing., Ph.D., UPGM (FIT)
Fröml Vojtěch, Ing., UIFS (FIT)
Herout Adam, prof. Ing., Ph.D., UPGM (FIT)
Nosko Svetozár, Ing., Ph.D., UPGM (FIT)
Šolony Marek, Ing., Ph.D., UPGM (FIT)
Zemčík Pavel, prof. Dr. Ing., dr. h. c., UAMT (FEKT), UPGM (FIT)
Zendulka Jaroslav, doc. Ing., CSc., UIFS (FIT)
This technical report describes image-processing and classification methods for extracting information from image data in order to increase the automation of processing data from the surveillance cameras in real time. Another considered source of image data is mobile imaging platforms such as drones etc.
Presented research focuses on specific methods for processing images with lower quality under realistic conditions. Selected techniques utilizes the high-dynamic-range (HDR) images in addition to the standard image representation, the development of image information over time, or the reconstruction of 3D information from multiple images. As a result of these procedures, specific methods are presented.
The report further describes the design of a system for managing video-data and metadata and their subsequent use in order to extract context information and detects predefined situation patterns in an online manner. The system architecture part presents key parts of the system, its functionalities and designed API.
@misc{BUT168677,
author="Vítězslav {Beran} and Vladimír {Bartík} and Vojtěch {Bartl} and Vojtěch {Fröml} and Adam {Herout} and Svetozár {Nosko} and Marek {Šolony} and Pavel {Zemčík} and Jaroslav {Zendulka}",
title="Selected Video-Processing Methods and System Architecture Design",
year="2017",
pages="1--15",
address="Technical report of project no. VI20172020068 - 2017 - Brno",
url="https://www.fit.vut.cz/research/publication/12391/"
}