Prediction model of AOD furnace based on flame image characteristic

Through analysis the principle of temperature measurement which was based on CCD image sensor, the mathematic model of temperature measurement with image was established. Then according to the character of CCD, the model was amended in bandwidth and gray body assumption. Experimental results show that during their own service of furnace the hit rate with prediction method of the Argon-Oxygen Decarburization refining end-point based on the flame image is higher than that with prediction method of Artificial experience. This way that measuring temperature by image is accurate. Therefore it has very wide application spaces.