基于支持向量机的智能楼宇火灾报警系统设计

Design of intelligent building fire alarm system based on support vector machine

为提高楼宇火灾报警的精度,文中提出一种基于支持向量机的智能楼宇火灾报警系统。将感知温度以及气体检测等数据模拟量作为输入,采用人工神经网络识别和模糊逻辑技术对智能楼宇火灾报警系统的硬件进行升级设计;现场采集监测区域内的红外图像,并获取火灾红外图像的现场温度、火灾面积以及相对稳定性等红外图像特征,及时提取火灾红外图像的显著动态特征;再将全部特征输入到支持向量机分类器中完成火灾识别,最终达到智能楼宇火灾报警的目的。仿真实验结果表明,所提方法可以准确地提取火灾红外图像特征,并能获取精度更高的报警结果。

In order to improve the accuracy of building fire alarm,an intelligent building fire alarm system based on support vector machine is proposed.The simulated data of perceived temperature and gas detection are taken as inputs.The hardware of intelligent building fire alarm system is upgraded and designed by means of artificial neural network recognition and fuzzy logic technology.The infrared images within the monitoring area are collected on site to obtain infrared image features such as on-site temperature,fire area,and relative stability of the fire infrared image,so as to timely extract significant dynamic features of the fire infrared image.All the features are inputted into the classifier of support vector machine to complete fire recognition,and then achieve the purpose of intelligent building fire alarm.The simulation experimental results show that the proposed method can accurately extract the infrared image features of fire and obtain more accurate alarm results.