Video surveillance-based identification of abnormal behavior of people in key areas

In this paper, from the demand of intelligent security monitoring of gas metering facilities, abnormal behavior monitoring and identification techniques for video terminals in key areas are proposed, mainly including personnel intrusion detection based on Faster-Rcnn deep neural network model, perso...

Full description

Saved in:
Bibliographic Details
Published in2022 International Conference on Machine Learning and Knowledge Engineering (MLKE) pp. 217 - 221
Main Authors Huang, Donghong, Li, Yong, Dong, XinLi, Jie, Hui, Dong, Xiangmin
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.02.2022
Subjects
Deep learning
machine vision
multi-target detection
Neural networks
Object detection
Personnel
Safety
Streaming media
surveillance
Target tracking
video
Online AccessGet full text

Cover

More Information
Summary:In this paper, from the demand of intelligent security monitoring of gas metering facilities, abnormal behavior monitoring and identification techniques for video terminals in key areas are proposed, mainly including personnel intrusion detection based on Faster-Rcnn deep neural network model, personnel wandering detection based on DeepSORT multi-target tracking model, and personnel holding foreign objects detection based on multi-target detection and area overlap. This method was tested in real scenarios with fast and accurate recognition and good results.
DOI:10.1109/MLKE55170.2022.00049