Perimeter Security
Click here for datasets about Perimeter Security
一、铁路周界标准化数据集
No.1 Dataset for Railway Perimeter Intrusion Detection
本数据集是铁路周界现场采集的上缆分布式光纤传感（DAS）标准化数据集，完成了栅栏清洗后的全流程数据清洗与事件标注，专为铁路周界安防场景下的入侵事件识别、环境干扰区分与智能监测算法研究构建
This dataset is a standardized Distributed Acoustic Sensing (DAS) dataset collected from railway perimeters. It features a complete workflow of data cleaning and event annotation following raw data preprocessing. The dataset is specifically constructed for research on intrusion event recognition, environmental interference discrimination, and intelligent monitoring algorithms within railway perimeter security scenarios
1. 数据规格与时空特性
1. Data Specifications and Spatio-temporal Characteristics
- 单样本维度：单条数据样本为
3001 × 5维度的 DAS 振动信号矩阵- 时间维度：3001 个时间采样点对应 6 秒的连续监测时长，对应系统采样率约为 500 Hz（
3001 / 6 ≈ 500 Hz），匹配分布式光纤传感系统的常规工程配置 - 空间维度：5 个空间采样通道对应 25 米的光纤监测段，单通道覆盖 5 米监测范围，为沿铁路栅栏上部布设的上缆周界监测区域
- 时间维度：3001 个时间采样点对应 6 秒的连续监测时长，对应系统采样率约为 500 Hz（
- Data Shape：A single data sample is a
3001 × 5DAS vibration signal matrix- Temporal Dimension：3,001 sampling points correspond to a continuous monitoring duration of 6 seconds, yielding a system sampling rate of approximately 500 Hz （
3001 / 6 ≈ 500 Hz）. This aligns with the standard engineering configuration of Distributed Acoustic Sensing (DAS) systems - Spatial Dimension：5 spatial sampling channels correspond to a 25-meter fiber optic monitoring segment. Each channel covers a 5-meter range, representing the overhead perimeter monitoring area deployed along the top of the railway fence
- Temporal Dimension：3,001 sampling points correspond to a continuous monitoring duration of 6 seconds, yielding a system sampling rate of approximately 500 Hz （
2. 事件类别体系
2. Label Hierarchy
数据集按事件类型划分为 7 个类别，全面覆盖铁路周界的环境干扰、常规运营事件与入侵 / 破坏风险事件。
The dataset is classified into seven distinct categories based on event types, providing comprehensive coverage of environmental interference, routine operational events, and high-risk intrusion or sabotage events at railway perimeters
| 类别名称 | 事件类型 | 场景说明 |
|---|---|---|
| 背景 | 环境本底信号 | 无人工扰动的铁路周界日常环境噪声，如环境微振动、系统背景噪声等 |
| 过车 | 运营背景事件 | 列车通过周界监测段时产生的光纤振动信号，为铁路运维场景下的常规背景事件 |
| 风 | 气象干扰事件 | 自然风作用于栅栏 / 光纤产生的振动信号，是周界安防中典型的环境干扰源 |
| 落石 | 地质灾害事件 | 落石撞击周界栅栏产生的振动信号，对应铁路沿线边坡落石等地质灾害风险 |
| 敲击 | 入侵试探事件 | 人工敲击栅栏的模拟入侵动作，为周界入侵的典型前置试探行为 |
| 攀爬 | 主动入侵事件 | 人工攀爬栅栏的入侵行为，是铁路周界安防的核心监测目标 |
| 拆卸 | 设施破坏事件 | 人工拆卸栅栏的破坏行为，对应周界物理设施的安全风险 |
| Category Name | Event Type | Scenario Description |
|---|---|---|
| Background | Ambient Background Signal | Daily environmental noise at the railway perimeter without human disturbance, including ambient micro-vibrations and system noise |
| Train Passage | Operational Background Event | Fiber-optic vibration signals generated as a train passes through the monitoring section; a routine background event in railway operation and maintenance (O&M) |
| Wind | Meteorological Interference Event | Vibration signals caused by natural wind acting on the fence or fiber, serving as a typical source of environmental interference in perimeter security |
| Falling Rocks | Geological Hazard Event | Vibration signals generated by rocks striking the perimeter fence, corresponding to geological risks such as slope rockfalls along railway lines |
| Knocking | Intrusion Probing Event | Simulated intrusion actions involving manual tapping or knocking on the fence, representing typical precursor/probing behavior of an intrusion |
| Climbing | Active Intrusion Event | Human climbing or scaling the fence; the core monitoring target for railway perimeter security |
| Dismantling | Facility Destruction Event | Manual dismantling or tampering with the fence, corresponding to the physical infrastructure security risks of the railway perimeter |
Click here for datasets about Railway Perimeter Intrusion Detection