Comprehensive Dataset for Event Classification Using Distributed Acoustic Sensing (DAS) Systems

This dataset was collected using a Distributed Acoustic Sensing (DAS) system with phase-sensitive Optical Time-Domain Reflectometry (Φ-OTDR) technology. It includes labeled and unlabeled acoustic signal measurements gathered around a university campus, covering activities such as walking, running, vehicular movement, and potential security threats like fiber manipulation and fence climbing. The data was captured using an Optasense ODH-F DAS interrogator, which monitors signals from a buried single-mode fiber optic cable. The dataset, stored in HDF5 format, serves as a critical resource for training machine learning models aimed at event classification in DAS systems. Each event is identified by power spectral density (PSD) representations and labeled accordingly. This dataset is ideal for researchers developing and validating machine learning algorithms for DAS-based applications, including structural health monitoring and perimeter security.本数据集采用搭载相敏光时域反射仪（phase-sensitive Optical Time-Domain Reflectometry, Φ-OTDR）技术的分布式声学传感（Distributed Acoustic Sensing, DAS）系统采集所得。数据集包含采集自大学校园周边区域的标注与未标注声学信号测量数据，覆盖步行、跑步、车辆通行等日常活动，以及光纤操控、翻越围栏等潜在安全威胁场景。数据采集环节使用Optasense公司ODH-F型DAS解调仪，该设备对埋地单模光缆的传感信号进行监测。本数据集以HDF5格式存储，是训练面向DAS系统事件分类任务的机器学习模型的关键支撑资源。每个事件均通过功率谱密度（power spectral density, PSD）特征进行表征并完成对应标注。本数据集可作为开发并验证面向DAS各类应用场景的机器学习算法的科研人员的理想实验数据资源，涵盖结构健康监测与周界安全等领域。

搜集汇总

背景与挑战

背景概述

该数据集是一个用于分布式声学传感（DAS）系统事件分类的综合性资源，采用相位敏感光时域反射（Φ-OTDR）技术，采集了大学校园周围包括行走、跑步、车辆移动和安全威胁等多种活动的声学信号数据。数据以HDF5格式存储，提供功率谱密度表示和相应标签，适用于训练和验证机器学习模型，支持结构健康监测和周边安全等应用。