Safety monitoring system for tourist scenic spots based on crowd scene type recognition
by Qinqin Dong
International Journal of Security and Networks (IJSN), Vol. 19, No. 3, 2024

Abstract: The tourism industry's rapid growth necessitates addressing the safety of tourist attractions. This study proposes a safety monitoring system based on crowd scene recognition. Using a dual-channel deep convolutional neural network model, the system combines static and dynamic features to identify and monitor potential risks in crowded areas. A comprehensive security monitoring solution is provided by integrating crowd distribution analysis and crowd type recognition using motion feature maps. The proposed density field-based crowd distribution analysis demonstrated superior estimation accuracy across various datasets. The safety monitoring system achieved an average accuracy of 82.35% and 81.32% in heterogeneous and homogeneous crowd scenes, respectively, with average area under the curve of 79.12% and 75.38%. These results significantly outperformed the comparison algorithms, confirming the study's effectiveness. The objective is to design accurate algorithms capable of estimating crowd types and behaviour patterns in complex scenarios, thereby enhancing the safety management and tourism experience at attractions.

Online publication date: Tue, 01-Oct-2024

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Security and Networks (IJSN):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com