球团回转窑火焰纹理特征与温度模型

为了完成球团回转窑内温度监测,本文对温度影响最大的火焰状况进行研究。首先将彩色图像进行RGB分解成三幅灰度图像,利用灰度共生矩阵研究RGB各通道的纹理参数,如能量、熵、对比度、相关性,与通过黑体炉标定的温度建立对应关系,制成温度模型,用于实时监测球团回转窑内火焰状况,指导生产。在实际测试中,该模型最大误差为5%,能够满足生产要求。

Texture of flame and temperature model in pellet rotary kiln

In order to monitor the internal temperature in pellet rotary kiln,the flame status was studied. First,the color image is decomposed to three gray-level images bases on RGB decomposition. Then the tex-ture parameters in each RGB channel such as energy,entropy,contrast and pertinence were analyzed by utiliz-ing gray-level co-occurrence matrix. Relationship between the texture and the temperature of the blackbody furnace is established and a temperature monitoring model is set up. The model can be used in real-time moni-toring of the flame in the pellet rotary kiln and guide the production processes. In the actual test,the maximum error of this model is about 3%,which can meet the production requirements.