Super-Resolution for Surveillance Video Via Adaptive Block-matching Registration

DOI

10.2991/aiie-15.2015.1How to use a DOI?

Keywords

video surveillance; block-matching; adaptive registration; super-resolution

Abstract

In recent years, numerous surveillance cameras are widely installed in the streets, residential areas and traffic routes to protect the national public security. Since interested objectives are often far away from the camera, the resolution of these target images are always in low-resolution (LR). It is hard to recognize the LR monitoring targets. Super-resolution (SR) algorithm can improve the resolution of input images effectively. It takes advantages of multi-frame images to reconstruct high-resolution (HR) target image. However, registration errors reduce the quality of reconstructed HR images. In order to solve this problem, an adaptive block-matching registration method based video SR algorithm is proposed in this paper. According to video content adaptive registration parameters improve the accuracy of multi-frame registration and furthermore enhance the quality of reconstructed HR image. Experimental results show that the proposed SR algorithm has better performance both on subjective and objective reconstruction quality than many state-of-art SR algorithms.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - T. Lu
AU  - W. Yang
AU  - Y.J. Wan
PY  - 2015/07
DA  - 2015/07
TI  - Super-Resolution for Surveillance Video Via Adaptive Block-matching Registration
BT  - Proceedings of the 2015 International Conference on Artificial Intelligence and Industrial Engineering
PB  - Atlantis Press
SP  - 1
EP  - 3
SN  - 1951-6851
UR  - https://doi.org/10.2991/aiie-15.2015.1
DO  - 10.2991/aiie-15.2015.1
ID  - Lu2015/07
ER  -