Development Of Ladle Slag Carry-over Detection System Based On Infrared Image Recognition

Slag Carry-over Detection Technology (SCDT) is of important practical significance to steel continuous casting production. It can enhance the purity of molten steel, reduce outlet block, increase molten steel yield ratio, protract the service life of high temperature materials, increase the number of continuous casting furnace and finally effectively improve the casting blank quality. Aiming at the problems of current SCDTs', such as expensive cost, complex installation, short service life, low signal strength, large time delay and not easy to workers labor protection, a kind of SCDT based on infrared image recognition measurement was put forward. Mechanism of infrared image and its key technologies were discussed deeply, and the industrial application system was successfully designed.According to the principle; which means in the same temperature conditions the molten steel and slag radiation coefficient is different, especially in the middle and far infrared band, This article discussed the basic law of infrared radiation deeply, which led to the discussion of the actual object emissivity and infrared thermal imaging system composition, and then basing on the basic parameters of the actual radiation about detected target, we implemented a series of calculations and choosed the infrared detector.But in the process of obtaining infrafed images, because of the external factors (such as dust, detector jitter) and internal factors (such as defects in the performance of the hardware acquisition circuit), the acquired images would be different from the original images in the some extent, So the images need image enhancement. This article dicussed different methods of image enhancement, and compared with experimental resultsAccording to whether it has long outlet protection sleeve or not, this article discussed the two situations. As for having long outlet protection sleeve, it presented slag feature extraction based on steel flow edge detection; as for the other situation, it presented slag feature extraction based on the average change in gray about the bakie fluid level, structural similarity and statistical properties.Finally, this acticle introduced the system software detailedly, including function and design, the choosing of software application platform and development tools and infrared image slag detection algorithm processes. On this basis, it further elaborated the key technologies, such as multithread technology and structure of the Filter Graph. Then it described the software module division in detail and depicted the overall software flow chart.Industrial field experiments proved:this detector system is easy to installation and maintenance, has high degree of automation, long service life and a certain success rate of slag detection, can effectively improve the yield of molten steel and reduce the thickness of the tundish slag deposits.