Research On Algorithm Of Ladle-Tundish Slag Detecting Based On Moment Function Of Ultrared Image

In modern continuous metal cast process, the quality of liquid steel is the promise of that of the steel products, while the slag is the main cause of pollution of liquid steel ,it is very severe to prevent slag from flowing from one process to the next process ,that just is Slag Detecting ,which includes two main phases: converter-ladle slag detecting and ladle-tundish slag detecting. This subject focuses on the later.According to the existing slag detecting method based on infrared image, we came up with our own approach in this dissertation, which paid more attention on processing the infrared image of the surface of the tundish ,and then calculated the character when slag emerges, which made the criterion of the wholesome detection.In this dissertation, reasonable presumption was made on the character of the object's infrared image, and according such presumption, image was weighted ,by which, meaningful signal was intensified, while redundant signal was depressed. Then algorithm for feature extraction based on moment function of images was developed.This dissertation intended mainly on theoretical research, in which we proposed some presumption, extracted some models of movement of slag, did some physical experiments. On the whole, this dissertation made some necessary exploration ,and will have some guiding significance in some extent.