The Application of Real-Time Object Detection on Aerial HD Videos Based on Deep CNN

DOI

10.2991/csece-18.2018.28How to use a DOI?

Keywords

real-time object detection; SSD; HD videos; UAV

Abstract

Efficient real-time object detection in aerial HD videos is an urgent need, with the increasing use of UAV(unmanned aerial vehicles) in various fields. But it is still challenging to detect objects accurately and timely due to its large pixel size and relatively small objects. We use the popular SSD to detect people, cars, boats and so on from real-time HD videos which behave better than traditional methods. Moreover, we improve the deep learning algorithm by tailoring its networks and enlarging its size. We apply this algorithm to the real time detection of UAV aerial HD video and experiments show that our method can realize the organic combination of detection efficiency and detection effect.

Copyright

© 2018, 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  - Guanghui Xu
AU  - Shuai Xie
AU  - Jun Wang
AU  - Guodong Wu
PY  - 2018/02
DA  - 2018/02
TI  - The Application of Real-Time Object Detection on Aerial HD Videos Based on Deep CNN
BT  - Proceedings of the 2018 International Conference on Computer Science, Electronics and Communication Engineering (CSECE 2018)
PB  - Atlantis Press
SP  - 134
EP  - 138
SN  - 2352-538X
UR  - https://doi.org/10.2991/csece-18.2018.28
DO  - 10.2991/csece-18.2018.28
ID  - Xu2018/02
ER  -