智能建筑火灾自动报警系统的分析与设计
发布时间：2018-08-06 10:39
【摘要】：随着科技创新智能化的发展越来越普及,关乎人们日常生活的基本设施的智能化也在快速发展,而智能建筑火灾报警系统就是其中一个体现。提高火灾报警系统的检测效率、灵敏度和可靠性,实现火灾的早期发现和报警,具有一定的实用价值。本论文主要研究了智能建筑火灾自动报警系统的构成原理,介绍了系统有关的基本概念、基本结构、基本性能。在对数据信息识别分析和数字图像处理技术识别等方法研究的基础上,重点研究了图像型火灾现场的信息识别分析方法。经过对图像进行滤波预处理、分割处理、优化处理,消除噪声,获取优化的图像样本后,通过对火焰的面积大小、形态变化和边缘变化等特征信息的提取和检测,进行了一系列火灾识别实验,通过仿真技术,验证火灾信息识别算法的可靠性和有效性。本文在火灾现场图像探测中引入BP神经网络算法,结合了该算法的特点和相关的数学模型函数,对实验进行了构建,给出了实验的具体输入输出单元的设计和神经网络的具体拓扑结构。对大量火灾图像样本和干扰图像样本进行了相关的对照实验。由实验的结果可以表明,基于BP神经网络算法的火灾报警系统相比于传统火灾报警系统有着更加明显的优势,大大的减少了火灾的误报率,提高了火情火灾报警的精准度,未来可以将其广泛运用于现代智能建筑综合体中。
[Abstract]:With the development of science and technology innovation and intelligence, the intelligence of basic facilities related to people's daily life is also developing rapidly, and the intelligent building fire alarm system is one of the embodiment. It is of practical value to improve the detection efficiency, sensitivity and reliability of the fire alarm system and to realize the early detection and alarm of the fire. In this paper, the principle of automatic fire alarm system for intelligent building is studied, and the basic concepts, basic structure and basic performance of the system are introduced. Based on the research of data information recognition and digital image processing technology, this paper focuses on the information recognition and analysis method of image fire scene. After the image is processed by filtering, segmentation, optimization, noise elimination, and the optimized image samples are obtained, the characteristic information such as flame area, shape change and edge change are extracted and detected. A series of fire identification experiments were carried out to verify the reliability and effectiveness of the fire information recognition algorithm. In this paper, BP neural network algorithm is introduced into the fire scene image detection, combining the characteristics of the algorithm and the related mathematical model function, the experiment is constructed. The design of the experimental input-output unit and the specific topology of the neural network are given. A large number of fire image samples and interference image samples were compared with each other. The experimental results show that the fire alarm system based on BP neural network has more obvious advantages than the traditional fire alarm system, greatly reduces the false alarm rate of fire, and improves the accuracy of fire alarm. In the future, it can be widely used in modern intelligent building complex.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41
本文编号：2167474
[Abstract]:With the development of science and technology innovation and intelligence, the intelligence of basic facilities related to people's daily life is also developing rapidly, and the intelligent building fire alarm system is one of the embodiment. It is of practical value to improve the detection efficiency, sensitivity and reliability of the fire alarm system and to realize the early detection and alarm of the fire. In this paper, the principle of automatic fire alarm system for intelligent building is studied, and the basic concepts, basic structure and basic performance of the system are introduced. Based on the research of data information recognition and digital image processing technology, this paper focuses on the information recognition and analysis method of image fire scene. After the image is processed by filtering, segmentation, optimization, noise elimination, and the optimized image samples are obtained, the characteristic information such as flame area, shape change and edge change are extracted and detected. A series of fire identification experiments were carried out to verify the reliability and effectiveness of the fire information recognition algorithm. In this paper, BP neural network algorithm is introduced into the fire scene image detection, combining the characteristics of the algorithm and the related mathematical model function, the experiment is constructed. The design of the experimental input-output unit and the specific topology of the neural network are given. A large number of fire image samples and interference image samples were compared with each other. The experimental results show that the fire alarm system based on BP neural network has more obvious advantages than the traditional fire alarm system, greatly reduces the false alarm rate of fire, and improves the accuracy of fire alarm. In the future, it can be widely used in modern intelligent building complex.
【学位授予单位】：东华理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41
【参考文献】
中国期刊全文数据库 前10条
1 方军利;;火灾自动报警系统在智能建筑中的应用[J];安防科技;2012年03期
2 陈俊海;;智能建筑中火灾自动报警系统研究[J];中国城市经济;2010年09期
3 黄松涛;石章松;张丕旭;;基于无线传感器网络的火灾现场定位技术研究[J];消防科学与技术;2009年02期
4 王建平;范科峰;张素兵;莫玮;;基于神经网络的多等级报警系统[J];计算机系统应用;2009年02期
5 蔡家楣;蔡其星;江颉;;基于遗传神经网络分析的内网用户行为审计系统[J];计算机系统应用;2009年02期
6 刘海霞;隋虎林;谷安永;;火灾报警控制系统通信数据信息识别的研究[J];消防科学与技术;2008年08期
7 董文辉;李宁宁;梅志斌;徐放;;欧美国家火灾报警系统规范介绍及启示[J];消防科学与技术;2008年06期
8 衣光照;赵加勇;毛勤;刘硕;;谈智能建筑的火灾自动报警与消防联动控制[J];科技风;2008年09期
9 李尧;张莉;单欣;;智能建筑中的火灾自动报警系统设计分析[J];安防科技;2008年02期
10 刘翱;马晓茜;;基于微粒群算法和模糊神经网络的火灾探测研究[J];消防科学与技术;2007年06期
，本文编号：2167474
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2167474.html
