Home // UBICOMM 2015, The Ninth International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies // View article


Generating Arbitrary View of Vehicles for Human-assisted Automated Vehicle Recognition in Intelligent CCTV Systems

Authors:
Youri Ku
Kim Insu
Kin Choong Yow

Keywords: 3D model reconstruct; drone; generate unseen view of object; image warping; texture map UVW; arbitrary view

Abstract:
Intelligent closed-circuit televisions (CCTV) are CCTV systems that can perform Video Content Analysis (VCA). However, in the area of Automatic Vehicle Recognition, there is still no good algorithm to recognize a car based on its description. In this paper, we propose a novel algorithm that will take an image (or several) of a car, extract special markings (if any) from it, and then texture map it to a 3D model of the same car. With the texture mapped 3D model, we can rotate the car to any arbitrary view point, especially to the view of another CCTV, so that non-sophisticated image matching algorithms can match the image to the actual CCTV feed. We performed experiments on three cars with different body markings and the results show that we can achieve quite realistic images of the car at any arbitrary viewpoint. This system will have significant impact on the use of intelligent CCTVs.

Pages: 89 to 95

Copyright: Copyright (c) IARIA, 2015

Publication date: July 19, 2015

Published in: conference

ISSN: 2308-4278

ISBN: 978-1-61208-418-3

Location: Nice, France

Dates: from July 19, 2015 to July 24, 2015