计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (26): 7-10.

• 博士论坛 • 上一篇    下一篇

基于火焰图像特征与GRNN的转炉吹炼状态识别

刘 辉1,张云生2,张印辉3,何自芬3   

  1. 1.昆明理工大学 冶金与能源工程学院,昆明 650093
    2.昆明理工大学 信息工程与自动化学院,昆明 650051
    3.昆明理工大学 机电工程学院,昆明 650093
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-09-11 发布日期:2011-09-11

State recognition of BOF based on flame image features and GRNN

LIU Hui1,ZHANG Yunsheng2,ZHANG Yinhui3,HE Zifen3   

  1. 1.Faculty of Metallurgical and Energy Engineering,Kunming University of Science and Technology,Kunming 650093,China
    2.Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650051,China
    3.Faculty of Mechanical and Electrical Engineering,Kunming University of Science and Technology,Kunming 650093,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-09-11 Published:2011-09-11

摘要: 随着转炉冶炼过程的推进,炉口火焰图像在不同的冶炼阶段呈现较为明显的差别。根据火焰图像判断冶炼所处阶段的问题,其关键在于如何准确提取火焰的主要特征,提出了火焰边缘线不变矩特征,火焰图像Laws纹理能量特征,以及图像色彩特征,并研究了它们的变化过程。最后,利用广义回归神经网络(GRNN)建立图像特征和冶炼阶段之间的分类模型。实验结果表明,该方法能有效进行基于图像识别的转炉冶炼状态判断,具有较高的实用价值。

关键词: 转炉, 火焰图像, 线不变矩, 颜色均值, Laws纹理, 广义回归神经网络

Abstract: With the BOF(Basic Oxyen Furnace) process forward,the flame images are different in each state.For the question that how to judge the state of BOF from flame images,the key problem is how to extract the main features of the flame images.The flame edge invariant line-moments characteristics and flame Laws texture energy characteristics are proposed,the flame color features are researched also.The changing processes of the three main features are studied.Finally,the state recognition model is established by using General Regression Neutral Network(GRNN).The experimental results show that the method is effective to do the state recognition of BOF based on flame image,and has a high practical value.

Key words: Basic Oxyen Furnace(BOF), flame image, invariant line-moments, Laws texture, General Regression Neutral Network(GRNN)