Enhancing φ-OTDR Event Classification with a Novel CNN-gMLP Model

This paper introduces a novel CNN-gMLP model for enhancing the classification of events in φ-OTDR systems. Leveraging a publicly available dataset featuring pre-classified φ-OTDR events and baseline SVM and CNN models, we demonstrate the effectiveness of our proposed approach. Our CNN-gMLP model, a novel architecture for this specific application, combines the strengths of convolutional neural networks for feature extraction with the efficient global modeling capabilities of gMLP. Experimental results show that our CNN-gMLP model achieves a 2% accuracy improvement over the baseline CNN model, highlighting its potential for accurate event classification in φ-OTDR systems. This paper provides a comprehensive analysis of the proposed model, including details on the dataset, experimental setup, and evaluation metrics. We discuss the significance of our findings, emphasizing the novelty of using CNN-gMLP for φ-OTDR event classification and its potential impact on the field. Furthermore, we outline potential future research directions, such as exploring alternative model architectures and investigating the generalization capabilities of our approach on diverse datasets. This work contributes to the advancement of φ-OTDR event classification by proposing a novel and effective deep learning model that outperforms existing methods.