Blast furnace chute visual detection based on mechanism model : Part 3: Visual Detection Based on Infrared Images

The present paper is part 3 of a paper series discussing the damage mechanism of blast furnace chute and the detection based on infrared imaging. While part 1 gives an overview on the effect of charge flow impact on blast furnace chute ^[1] and part 2 studies on the effect of high temperature gas flow on blast furnace chute ^[2]. Based on the previous two parts of the study, we can predict the possible position of chute damage before monitoring, narrow the detection range, and lay a foundation for accurate detection. This third part discusses image processing of infrared image of blast furnace. The blast furnace chute is in constant motion, which makes the automatic detection of infrared thermal imaging on top of blast furnace based on image processing become a difficult task. According to different thermodynamic coefficient and temperature, the image of each part of the furnace top is different. The features of damaged chute areas are extracted by means of image processing methods such as median filtering, histogram equalization, thresholding and image multiplication. It is necessary to combine the prediction of mechanism model with visual image to determine whether the abnormal temperature area is consistent with the vulnerable area, otherwise there will be misjudgment.