Thesis Details
Surveillance Video Search
With the growing number of video recordings produced by security cameras, the demand for systems that are able to search them is growing as well. This work examines such systems and methods that are behind them.The introduction of this thesis describes the scheme of surveillance video search systems together with the methods that these systems use to store information they obtain during the video analysis. Algorithms for object detection (YOLO) and object tracking (DeepSort) are also introduced. These algorithms are then used in a system created for the practical part of this thesis.The end of the thesis describes the created system, which uses the trajectories of detected objects in searched video recordings. To specify the searched events, the proposed query language within this work is used. This language consists of so-called search blocks, the composition of which can be used to define situations such as: "a person got out of a car" or "a car stopped in a parking space".
surveillance video search, object detection, object tracking, object trajectory-based video search, homography
Bařina David, Ing., Ph.D. (DCGM FIT BUT), člen
Burget Radek, doc. Ing., Ph.D. (DIFS FIT BUT), člen
Holík Lukáš, doc. Mgr., Ph.D. (DITS FIT BUT), člen
Jaroš Jiří, doc. Ing., Ph.D. (DCSY FIT BUT), člen
@bachelorsthesis{FITBT22439,
author = "Luk\'{a}\v{s} Piwowarski",
type = "Bachelor's thesis",
title = "Surveillance Video Search",
school = "Brno University of Technology, Faculty of Information Technology",
year = 2020,
location = "Brno, CZ",
language = "english",
url = "https://www.fit.vut.cz/study/thesis/22439/"
}