Abstract
An intrusion recognition method based on the combination of one-dimensional convolution neural network (1-D CNN) and DenseNet network is proposed and demonstrated. Assisted with fiber optic distributed acoustic sensing (DAS) system, the average recognition accuracy of 98.4% for six types of events and 2ms processing time were achieved in the field test.
© 2021 The Author(s)
PDF ArticleMore Like This
Shixiong Zhang, Tao He, Kai xiao, Cunzheng Fan, Hao Li, Zhijun Yan, Deming Liu, and Qizhen Sun
JW3A.22 CLEO: Applications and Technology (CLEO:A&T) 2022
Tao He, Yijie Liu, Shixiong Zhang, Zhijun Yan, Deming Liu, and Qizhen Sun
JW1A.119 CLEO: Applications and Technology (CLEO:A&T) 2021
Shixiong Zhang, Tao He, Hao Li, Kai Xiao, Cunzheng Fan, Zhijun Yan, and Qizhen Sun
AW4I.6 CLEO: Applications and Technology (CLEO:A&T) 2023