作者：王韬, 刘奇, 陈洪, 孟川杰

作者单位：四川大学电气信息学院, 四川成都 610065

关键词：CCD; BP神经网络; HIS; 颜色模型

摘要：

基于炉膛火焰温度不同,与之相对应通过CCD摄取的炉膛火焰图像颜色也不同,提出了一种检测炉膛火焰温度的方法。将获取的炉膛火焰图像RGB模型转换为HSI颜色值,用H、S值作为BP神经网络输入,通过样本图像训练后,拟合H、S与温度T的非线性关系,计算得到炉膛火焰温度值。实验表明,计算温度与实际温度良好相符,温度测量方法切实可行。

Temperature measurement method of flame images based on BP neural network

WANG Tao, LIU Qi, CHEN Hong, MENG Chuan-jie

College of Electric Information, Sichuan University, Chengdu 610065, China

Abstract: Based on the fact that different furnace flame images captured by the CCD correspond with the different temperature of the furnace flame, a temperature measurement method of the furnace flame images is presented in this paper.Transforming the RGB color model images captured by the CCD into HSI color model images, training this neural network and approximating the nonlinear relationship between the color and the temperature based on BP neural network, the temperature is calculated.The experimental results prove that the calculating temperature accords with the actual one and show that the method is feasible.

Keywords: CCD; BP neural network; HSI; Color model

2005, 31(2): 50-52  收稿日期: 2004-12-5;收到修改稿日期: 2005-2-23

基金项目: 

作者简介: 

参考文献

[1] 王胜正.施朝健.基于两种颜色空间的颜色选择方法[J].计算机应用与软件, 2004, 21(2).
[2] Rafael C, Gonzalez Richard E,Woods.Digital Image Processing (Second Edition), Publishing House of Electronics Industry, 2003, 3.
[3] 王洪元, 史国栋.人工神经网络技术及其应用[M].北京:中华石化出版社, 2002.
[4] 金丕彦, 芮勇.BP算法各种改进算法的研究和应用[J].南京航空航天大学学报, 1994, 26(S1).
[5] 段晓君.神经网络的函数逼近能力分析[J].模糊系统与数学, 1998, 12(4).
[6] 飞思科技产品研发中心.MATLAB6.5辅助神经网络分析与设计[M].北京:电子工业出版社, 2003, 1.
[7] 瞿东晖, 张立明.多层前馈网络在模式识别中的理论和应用[J].电子学报, 1995, 23(7).
[8] 刘耦耕, 贺素良.BP神经网络结构参数的计算机自动确定[J].计算机工程与应用, 2004, 1(1).