[1]何世钊,杨宣访,陈晓娟.支持向量机与BP网络在火灾图像探测上的比较[J].智能系统学报,2011,6(04):339-343.
 HE Shizhao,YANG Xuanfang,CHEN Xiaojuan.Comparisons between a support vector machine and BP neural network for video image fire detection[J].CAAI Transactions on Intelligent Systems,2011,6(04):339-343.
点击复制

支持向量机与BP网络在火灾图像探测上的比较(/HTML)
分享到:

《智能系统学报》[ISSN:1673-4785/CN:23-1538/TP]

卷:
第6卷
期数:
2011年04期
页码:
339-343
栏目:
出版日期:
2011-08-25

文章信息/Info

Title:
Comparisons between a support vector machine and  BP neural network for video image fire detection
文章编号:
1673-4785(2011)04-0339-05
作者:
何世钊杨宣访陈晓娟
海军工程大学 电气与信息工程学院,湖北 武汉 430033
Author(s):
HE Shizhao YANG Xuanfang CHEN Xiaojuan
College of Electrical and Information Engineering, Naval University of Engineering, Wuhan 430033, China
关键词:
火灾探测形状特征支持向量机BP神经网络
Keywords:
fire detection shape features SVM BP neural network
分类号:
TP18
文献标志码:
A
摘要:
针对BP神经网络和支持向量机在火灾探测上存在的理论差别,分别构建了基于此2种方法的火灾图像探测方法.2种方法均依据火焰颜色分布规律实现了目标区域的分离,并将目标区域的形状特征及变化值作为判据.通过对火灾实验样本的训练及识别,2种方法的探测表现得到了比较与分析.实验结果表明基于支持向量机的火灾探测方法具有快速收敛特性及所需较少训练样本的优点.同时,BP神经网络对测试集较少的错判反映出其良好的非线性映射能力,适合求解内部机制复杂的问题.
Abstract:
According to the theoretical differences between a back propagation (BP) network and support vector machine (SVM) in relation to fire detection, two kinds of video image fire detection methods based on a BP network and SVM, respectively, were constructed. Judging from color distribution of the flames, the objective regions were separated in both methods, and their shape features along with the changes in shape features were extracted as criteria. The performance of each method was compared and analyzed after conducting many experiments. The experimental results show that the SVM had a high convergence rate and needed fewer training samples. At the same time, fewer misjudgments of testing samples confirmed that the BP network was more suitable for solving complex internal mechanism problems due to its good mapping capability. 

参考文献/References:

[1]王振华,王慧琴,王燕妮,等.基于视频图像的火灾探测方法[J].世界科技研究与发展, 2008, 30(2): 161164.
 WANG Zhenhua, WANG Huiqin, WANG Yanni, et al. Fire detection in video[J]. World SciTech R&D, 2008, 30(2): 161164.
[2]徐小军,郑健,郭尚芬.火灾图像探测的神经网络方法研究[J].计算机工程与设计, 2008, 29 (13): 34163418.
 XU Xiaojun, ZHEN Jian, GUO Shangfen. Study on neural network method of fire image detection[J].Computer Engineering and Design, 2008, 29 (13): 34163418.
[3]朱世增,党选举.基于相关向量机的非线性动态系统辨识[J].计算机仿真, 2008, 25(6): 103107.
ZHU Shizeng, DANG Xuanju. Nonlinear dynamic system identification based on relevance vector machine[J]. Computer Simulation, 2008, 25(6): 103107.
[4]CHEN T H, WU P H, CHIOU Y C. An early firedetection method based on image processing[C]//International Conference on Image Processing.[S.l.], 2004: 17071710.
[5]刘叔军,盖晓华,樊京,等. Matlab 7.0控制系统应用与实例[M]. 北京:机械工业出版社, 2005: 9.
[6]VAPNIKK V N. The nature of statistical learning theory [M]. New York: Springer Verlag, 1995: 101105.
[7]孙正兴,徐文晖.基于局部SVM分类器的表情识别方法[J].智能系统学报, 2008, 3(5): 455466.
SUN Zhengxing, XU Wenhui. Facial expression recognition based on local SVM classifiers[J]. CAAI Transactions on〖LL〗Intelligent Systems, 2008, 3(5): 455466.
[8]JOHN C P. Training of support vector machines using sequential minimal optimization[C]//Advances in Kernel MethodSupport Vector Learning. Cambridge: MIT Press, 1999: 101105.
[9]KEETHI S S, SHEVADE S K, BHATTACHARYYA C, et al. Improvements to Platt’s SMO algorithm for SVM classifier design[J].Neural Computation, 2001, 13( 3): 637649.
[10]黄豪彩,黄宜坚,杨冠鲁.基于LM算法的神经网络系统辨识[J].组合机床与自动化加工技术, 2003, 1(2): 611. 
HUANG Haocai, HUANG Yijian, YANG Guanlu. Neural network system identification based on LevenbergMarquardt algorithm[J]. Modular Machine Tool & Automatic Manufacturing Technique, 2003, 1(2): 611.

备注/Memo

备注/Memo:
收稿日期: 2010-06-19.
基金项目:国家自然科学基金资助项目(50721063).
通信作者:何世钊.E-mail:heshizhao_chn@126.com.
作者简介:
 何世钊, 男, 1986年生,硕士研究生,主要研究方向为检测技术与自动化装置.
杨宣访,男, 1968年生, 副教授, 硕士生导师.主要研究方向为自动测试、电路故障诊断、电力系统诊断.先后主持、参与军队、海军等多项重点科研和工程项目.获得军队科技进步一等奖1项,二等奖2项,三等奖4项.发表学术论文20余篇,参与编写教材1部.
陈晓娟, 女, 1981年生, 博士研究生, 主要研究方向为数字信号处理.
更新日期/Last Update: 2011-09-30