基于BP神经网络和证据理论融合的火灾探测信息处理

发布时间：2018-02-11 13:29

本文关键词： 火灾自动探测信号处理 Labview虚拟仪器 BP神经网络 D-S证据理论融合不确定度　出处：《天津理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着电子信息科技系统的发展和进步，也随着新型编程软件的开发和应用，近些年来，火灾自动探测系统已逐渐成为一种综合类、多学科、多资源的应用技术。为了更好的对火灾进行预测，火灾自动智能报警系统的创新和信号处理技术的发展与研究成为一种趋势和必然。火灾探测和火灾信息处理的根本目的和处理方法是利用探测器对火灾生成物的测量，，以求尽早尽快的在火灾初期就可以发现火情并降低火灾探测中的漏报率和误报率。火灾探测和火灾信息处理技术促进了一系列新技术、新发明的发展。人们开始将跨学科跨领域的知识应用于火灾探测火灾信息处理过程中。然而，由于技术水平的不足和理论水平的低下，目前人们使用火灾探测系统和火灾信息处理方法还存在种种的问题。例如：目前多使用的多线制或二线制火灾自动报警系统，存在线路复杂、布点困难等缺点；火灾信息处理过程中，人们使用阈值比较法或变化率检测法在火灾信号处理过程中已不能适应复杂的火情状况变化；特定材料和特定环境中的火灾所得到的物理参量与普通火灾相比存在明显的区别，单一的火灾预测信息处理办法已不能满足火灾报警的需求。因此，一种新型结构的火灾报警系统的研制和新的火灾信号处理办法的使用成为目前在火灾处理领域研究的热点。本文将研究基于二总线火灾自动报警系统，并利用BP神经网络与数据融合理论的结合进行火灾信息处理。结合天津理工大学教改项目（Y1311-28）和天津市普通高等学校本科教学质量与教学改革研究计划(B01-0829)开展相关的研究工作，并且根据前人的经验对周围环境进行改造与设计，进行真实火灾试验，并顺利地达到了预期目的和效果。本文论述了火灾信息处理算法的发展历程和不足，对几种常用的火灾信息处理算法进行了详细的介绍，并对其优缺点进行了展开阐述；对数据融合理论进行了介绍，着重对D-S数据融合理论进行说明；提出了不确定度的概念来说明系统和信息处理的稳定性和有效性；对Labview虚拟仪器工具进行详尽的介绍，并且对目前流行的几种基于Labview工具的应用进行了说明和阐述；着重介绍了基于Labview虚拟仪器的BP神经网络;设计进行了两类实验，即火灾监测试验与火灾探测试验，对试验中的硬件与软件部分分别进行了介绍，同时对实验场所和试验材料进行了说明；对实验数据进行了分析，将基于Labview工具的BP神经网络技术和基于D-S证据融合理论的信息处理进行了应用，最后对实验结果做出评价。随着科技的发展与时代的进步，相信还有更多的火灾自动报警系统将会产生，更加先进的智能的信息处理方法将会出现以满足火灾探测与火灾预测的需求。本文在最后对火灾自动报警系统的发展趋势进行了展望，提出了笔者认为的火灾自动报警系统的发展方向，并对信息处理的发展目标进行了预测性说明和阐述。
[Abstract]:With the development and application of the electronic information science and technology system , with the development and application of the new programming software , the fire automatic detection system has become a comprehensive , multi - disciplinary and multi - resource application technology in recent years . In order to better predict the fire , the innovation of the automatic fire alarm system and the development and research of signal processing technology have become a trend and inevitable . The basic purpose and the processing method of fire detection and fire information processing are to use the detector to measure the fire product so as to find out the fire at the early stage of the fire as soon as possible and reduce the false alarm rate and false alarm rate in the fire detection . Fire detection and fire information processing technology has promoted a series of new technologies and the development of new invention . People began to apply cross - disciplinary knowledge to fire detection and fire information processing . However , there are many problems in fire detection system and fire information processing method because of insufficient technical level and low theoretical level . Therefore , the development of a new type of fire alarm system and the use of a new fire signal processing method have become a hot spot in the field of fire treatment . Based on the two - bus fire automatic alarm system , the paper studies the fire information processing by combining BP neural network and data fusion theory . Combined with the teaching reform project of Tianjin University of Science and Technology ( Y1311 - 28 ) and the undergraduate teaching quality of Tianjin University of Science and Technology ( B01 - 0829 ) , the research work is carried out , and the environment is reconstructed and designed according to the experience of the previous person , and the real fire test is carried out , and the expected purpose and effect are achieved successfully . In this paper , the development course and shortage of fire information processing algorithm are discussed , and the advantages and disadvantages of several common fire information processing algorithms are introduced . The theory of data fusion is introduced , and the application of the Labview virtual instrument is introduced . Two kinds of experiments are put forward , namely , fire monitoring test and fire detection test , the hardware and software parts in the test are described . The experimental data is analyzed , and the information processing based on Labview tool and D - S evidence fusion theory are applied . Finally , the experimental results are evaluated . With the development of science and technology and the progress of the times , it is believed that more automatic fire alarm systems will be generated , and the more advanced intelligent information processing method will appear to meet the needs of fire detection and fire prediction . In the end , the development trend of the automatic fire alarm system is forecasted , and the development direction of the fire automatic alarm system is put forward .

【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP18;X932

【参考文献】