Optical flows Clustering Used for Counting Pedestrians

DOI

10.2991/ameii-16.2016.248How to use a DOI?

Keywords

Pedestrian Counting, Pyramid Optical Flow Clustering, Corner Point

Abstract

Traditionally ,pedestrian counting is a manual process. It requires a lot of manpower and material resources, and also generates the possible human error. Therefore, the reality in many cases are in urgent need of automatic pedestrian counting. The most widely deployed methods utilize laser sensors and infrared sensors. However, these methods sometimes fail to count pedestrians correctly when the heights of pedestrians walking together are similar or when the heights do not fall within the presumed range, because such methods depend on the difference in propagation delays of reflected laser pulses or infrared light. Although the methods using multiple infrared sensors can count pedestrians moving various directions, the counting accuracy degrades considerably when the street has much traffic and occlusion occurs frequently. An occlusion is caused by pedestrians interacting with each other when many pedestrians are present. In this paper, we introduce a method based on Pyramid optical flows clustering of corner point to improve the counting accuracy. We also report that using length clustering, angle clustering and original location clustering to enhance the counting accuracy.

Copyright

© 2016, 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  - Ping Pan
AU  - Yujiang Fu
AU  - Fangming Liu
PY  - 2016/04
DA  - 2016/04
TI  - Optical flows Clustering Used for Counting Pedestrians
BT  - Proceedings of the 2nd International Conference on Advances in Mechanical Engineering and Industrial Informatics (AMEII 2016)
PB  - Atlantis Press
SN  - 2352-5401
UR  - https://doi.org/10.2991/ameii-16.2016.248
DO  - 10.2991/ameii-16.2016.248
ID  - Pan2016/04
ER  -