A computer vision system for the characterization and classification of flames in glass furnaces

A computer vision system for dealing with both flame analysis and classification problems is described. The vision system performs classification tasks, assigning each visualized flame to a previously established class of flames. Each flame class corresponds to a specific operating point of the furnace, and this information can be useful not only for monitoring purposes but also for purposes controlling the whole furnace. Two different classifiers were designed and compared. One exploits a Bayesian formulation, and the other is based on neural networks. They system was first tested in a laboratory environment, where different operating conditions were created through the variation of the burners geometry, the number of active burners, and the rate of fuel used. In a second stage, the vision system was tested in an industrial environment. The results obtained are presented and discussed.<
>