摘要
设计了基于ARM(高级精简指令集机器)嵌入式多传感器信息融合的森林火灾报警系统,该系统主要分为传感器节点、信号传输、总控系统和融合算法4个部分.传感器节点主要由LM-PT100温度、M397666空气湿度和MQ-2烟雾传感器组成;信号传输部分以CC2430为控制芯片,采用ZigBee技术实现无线传感器网络自组网和数据传输;总控系统以嵌入式微处理器2410为硬件平台,结合Linux操作系统和ADS1.2集成开发环境.设计了基于粗糙集、BP神经网络和D-S证据理论相结合的融合算法:粗糙集对数据进行约简,神经网络对约简后的子集进行分类识别,证据理论对每个子集的分类识别结果进行融合决策.运行结果显示:该系统与常规的'望台观测'、'地面巡护'等森林监控方法相比,能以更快的速度和效率预防并扑灭森林火灾.
A warning system for forest fire was designed by ARM (advanced RISC machines) embedded multi-sensor information fusion, which was mainly divided into four parts: sensor nodes, signal transmission, total control system and integration algorithm. The sensor node was mainly designed by the composition of the LM-PT100 temperature sensors, the M397666 humidity sensors and the MQ-2 smoke sensors. Signal transmission part used CC2430 as the control chip, and wireless sensor networks of ad hoc networks and data transmission were designed based on the ZigBee technology. The total control system hardware platform was combined with the Linux operating system and ADS1. 2 integrated development environment with embedded microprocessor 2410. The fusion algorithm was designed based on rough sets, back propagation neural network and D-S evidence theory. And the data reduction was achieved by rough set; the subset was classified by the neural network; the integration and decision-making of the classification and recognition results for each subset was obtained by the evidence theory. The running results show that compared with conventional 'lookout observatory' and 'ground patrol' forest monitoring methods, this system can prevent and fight forest fires at a faster speed and efficiency.
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2013年第2期22-25,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
湖北省自然科学基金资助项目(2010CDB11101)
武汉工程大学研究生教育创新基金资助项目(CX201163)