基于多源信息融合与NB-IoT传输的滑坡灾害智能预警系统设计

Design of an Intelligent Landslide Early Warning System Based on Multi-Source Information Fusion and NB-IoT Transmission

摘要:

针对山区滑坡监测普适性差、成本高的问题,设计了一套基于物联网的嵌入式预警系统。系统以STM32为核心,集成雨量、土壤湿度、水位、振动及倾角多源传感器,构建边坡数据感知层；采用NB-IoT实现低功耗数据至ONENET云平台的实时上报。通过阈值分级与加权平均融合算法,将水文因子(雨量、湿度、水位)与动态因子(振动、倾角)分别加权,结合四级风险评估模型,实现滑坡风险动态分级预警。实验表明：传感器测量误差<7%,传输延迟2.68 s,满足实时性；融合算法在暴雨、倾角突变等场景下能准确触发分级警报,有效降低误报率。系统具有低成本、低功耗、高可靠性等特点,为滑坡防控提供了有效方案。

Abstract:

To address the challenges of poor adaptability and high costs associated with conventional landslide monitoring in mountainous regions, this study designs an embedded early warning system based on the Internet of Things. The system employs an STM32 microcontroller as the core processor and integrates multiple heterogeneous sensors, including those for rainfall, soil moisture, water level, vibration, and tilt angle, thereby establishing a data acquisition layer for slope monitoring. Narrowband Internet of Things technology is utilized to construct a low power remote transmission channel, enabling real time data reporting to the ONENET cloud platform. A data fusion algorithm based on threshold grading and weighted averaging is introduced, which separately weights hydrological factors, namely rainfall, soil moisture, and water level, and dynamic factors, namely vibration and tilt angle. Combined with a four level risk assessment model, the system achieves dynamic and graded early warning of landslide risks. Experimental results demonstrate that sensor measurement errors remain within 7 percent, and the average data transmission delay is 2.68 seconds, satisfying real time monitoring requirements. The fusion algorithm accurately triggers graded alarms under critical scenarios such as heavy rainfall and abrupt tilt changes, effectively reducing the false alarm rate. The proposed system is characterized by low cost, low power consumption, and high reliability, offering an effective solution for landslide prevention and control.