【摘要】：隨著科技創(chuàng)新智能化的發(fā)展越來越普及,關(guān)乎人們?nèi)粘Ｉ畹幕驹O施的智能化也在快速發(fā)展,而智能建筑火災報警系統(tǒng)就是其中一個體現(xiàn)。提高火災報警系統(tǒng)的檢測效率、靈敏度和可靠性,實現(xiàn)火災的早期發(fā)現(xiàn)和報警,具有一定的實用價值。本論文主要研究了智能建筑火災自動報警系統(tǒng)的構(gòu)成原理,介紹了系統(tǒng)有關(guān)的基本概念、基本結(jié)構(gòu)、基本性能。在對數(shù)據(jù)信息識別分析和數(shù)字圖像處理技術(shù)識別等方法研究的基礎(chǔ)上,重點研究了圖像型火災現(xiàn)場的信息識別分析方法。經(jīng)過對圖像進行濾波預處理、分割處理、優(yōu)化處理,消除噪聲,獲取優(yōu)化的圖像樣本后,通過對火焰的面積大小、形態(tài)變化和邊緣變化等特征信息的提取和檢測,進行了一系列火災識別實驗,通過仿真技術(shù),驗證火災信息識別算法的可靠性和有效性。本文在火災現(xiàn)場圖像探測中引入BP神經(jīng)網(wǎng)絡算法,結(jié)合了該算法的特點和相關(guān)的數(shù)學模型函數(shù),對實驗進行了構(gòu)建,給出了實驗的具體輸入輸出單元的設計和神經(jīng)網(wǎng)絡的具體拓撲結(jié)構(gòu)。對大量火災圖像樣本和干擾圖像樣本進行了相關(guān)的對照實驗。由實驗的結(jié)果可以表明,基于BP神經(jīng)網(wǎng)絡算法的火災報警系統(tǒng)相比于傳統(tǒng)火災報警系統(tǒng)有著更加明顯的優(yōu)勢,大大的減少了火災的誤報率,提高了火情火災報警的精準度,未來可以將其廣泛運用于現(xiàn)代智能建筑綜合體中。
[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.
【學位授予單位】：東華理工大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP391.41

【參考文獻】

中國期刊全文數(shù)據(jù)庫 前10條

1 方軍利;;火災自動報警系統(tǒng)在智能建筑中的應用[J];安防科技;2012年03期

2 陳俊海;;智能建筑中火災自動報警系統(tǒng)研究[J];中國城市經(jīng)濟;2010年09期

3 黃松濤;石章松;張丕旭;;基于無線傳感器網(wǎng)絡的火災現(xiàn)場定位技術(shù)研究[J];消防科學與技術(shù);2009年02期

4 王建平;范科峰;張素兵;莫瑋;;基于神經(jīng)網(wǎng)絡的多等級報警系統(tǒng)[J];計算機系統(tǒng)應用;2009年02期

5 蔡家楣;蔡其星;江頡;;基于遺傳神經(jīng)網(wǎng)絡分析的內(nèi)網(wǎng)用戶行為審計系統(tǒng)[J];計算機系統(tǒng)應用;2009年02期

6 劉海霞;隋虎林;谷安永;;火災報警控制系統(tǒng)通信數(shù)據(jù)信息識別的研究[J];消防科學與技術(shù);2008年08期

7 董文輝;李寧寧;梅志斌;徐放;;歐美國家火災報警系統(tǒng)規(guī)范介紹及啟示[J];消防科學與技術(shù);2008年06期

8 衣光照;趙加勇;毛勤;劉碩;;談智能建筑的火災自動報警與消防聯(lián)動控制[J];科技風;2008年09期

9 李堯;張莉;單欣;;智能建筑中的火災自動報警系統(tǒng)設計分析[J];安防科技;2008年02期

10 劉翱;馬曉茜;;基于微粒群算法和模糊神經(jīng)網(wǎng)絡的火災探測研究[J];消防科學與技術(shù);2007年06期

，

本文編號：2167474