Loading...
Loading...
Contributors:
Dataset URI:
10.1371/journal.pone.0338205.t001Oil and gas pipeline security is critical to national infrastructure, yet existing monitoring systems often lack the sensitivity and real-time responsiveness required to detect subtle intrusion events. This study presents a novel multimodal sensing and interaction frame-work that integrates phase-sensitive optical time-domain reflectometry (φ-OTDR)–based distributed acoustic sensing (DAS) with an optimized one-dimensional convolutional neural network (1-D CNN) architecture. The approach leverages both raw fiber optic vi-bration signals and carefully selected handcrafted features, enabling robust automatic in-trusion classification across multiple event types including manual tapping, mechanical excavation, and human footsteps. By incorporating transfer learning from publicly avail-able human activity datasets, the model achieves enhanced feature generalization, result-ing in a classification accuracy exceeding 95%. This work demonstrates the potential of combining advanced multimodal sensing technologies with deep learning-based interac-tive analytics for real-time pipeline security monitoring, paving the way for intelligent in-frastructure protection systems. Future efforts will focus on expanding dataset diversity, integrating multi-sensor fusion, and enhancing adaptive interaction capabilities for field deployment.
Keyword(s):
approach leverages; based interac; Biological Sciences not elsewhere classified; bration signals; deep learning.
MeSH Term(s):
Grants and Funding:
No Grants or Funding Data Available
Associated Publication(s):
Supplementary Data:
No Supplementary Data Available
Related Dataset(s):
No Related Dataset Data Available
The National Library of Medicine (NLM) provides information about biomedical datasets. Inclusion on an NLM Web page does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: NLM Web Policies (nih.gov)