大空间建筑可视火灾探测系统的设计与实现  

Design and Implementation of Visible Fire Detection System in Large Space Building

作　　者：马宗方[1,2] 程咏梅[1] 王慧琴[2] 张国飞[2] 

机构地区：[1]西北工业大学自动化学院,陕西西安710072 [2]西安建筑科技大学信控学院,陕西西安710055

出　　处：《计算机仿真》2011年第4期283-286,307,共5页Computer Simulation

基　　金：国家自然科学基金重点项目(60634030)

摘　　要：研究失火探测优化系统,传统感温感烟探测器在大空间建筑中探测距离短、误报率高、受环境因素影响大等的缺陷。为克服这些缺陷,并解决目前视频火灾探测的检测率低、系统开发复杂的问题,在嵌入式集成开发环境下设计并实现了火焰多特征识别的可视火灾探测系统。通过融合火焰的颜色、亮度及跳动形态等多特征提取出火焰疑似区域,进行火焰检测和跟踪识别,运用Simulink构建系统算法模型和ETTIC6000完成面向目标板的代码自动生成,简化了系统的开发流程。实验测试结果表明,系统可以实现大空间建筑火灾火焰特征的快速提取与检测,并具有开发周期短的优点。In the applications of fire detection for large space building,the traditional temperature and smoke sensors have many defects,such as short distance of detecting,the high rate of false alarming and great impact by environmental factors.In order to overcome the defects of traditional detectors and and resolve the problems of the low detection rate and the complexity of system development in the current video fire detection,a video fire detection based on multi-features flame recognition is designed and implemented in the embedded integrated development environment.Flame suspected regions are extracted based on the fusion of features of color,brightness and jumping shape for detecting and target tracking.The detection algorithm model is constructed in Simulink and the code is generated automatically by ETTIC6000 in the joint developing environment of Matlab,Simulink and Code Composer Studio.The experiment results show the system not only has good real-time experience and expansibility,but also simplifies the developing process effectively to push the products into market in short time.

关 键 词：视频火灾探测 火焰特征 动态仿真 代码生成 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]