Home // ICONS 2020, The Fifteenth International Conference on Systems // View article

Multiview-Fusion-Based Crowd Density Estimation Method for Dense Crowd

Authors:
Liu Bai
Cheng Wu
Yiming Wang
Feng Xie

Keywords: Crowd gathering safety situation; Video monitoring; Accident analysis and early warning; Traffic safety.

Abstract:
Crowd gathering places are prone to crowd stampede and other public emergencies, resulting in large numbers of casualties and property losses, then, leading to negative social impact. At present, the research on dynamic assessment of crowd gathering safety situation mainly relies on isolated real-time video monitoring, and lacks reliable methods to deal with plenty of video data from different sources, perspectives and granularities. Based on the traffic Internet of things infrastructure, this paper explores the fusion technology of multi-sensor source homogeneous video data. On the basis of the static model of crowd aggregation based on the high-altitude perspective, this paper studies the different source and multi granularity real-time dynamic monitoring video cooperative perception methods in the middle and low altitude and different perspectives. The dynamic scene crowd statistical perception including motion prediction mechanism is used to extract the global coarse-grained motion situation of the crowd from the perspective of high altitude. The multi column convolution depth neural network is used to extract the local fine-grained density features of the crowd with line of sight occlusion in low altitude perspective, thus establishing the holographic model of the temporal and spatial evolution of crowd situation, and proposing a new method of crowd aggregation safety situation assessment. This method is applied to the crowd gathering safety situation assessment of Suzhou city life fountain square, and achieves good results, which provides theoretical support for the safety control of crowd gathering place based on the Internet of things.

Pages: 50 to 55

Copyright: Copyright (c) IARIA, 2020

Publication date: February 23, 2020

Published in: conference

ISSN: 2308-4243

ISBN: 978-1-61208-771-9

Location: Lisbon, Portugal

Dates: from February 23, 2020 to February 27, 2020