Continuous Casting Ladle Slag Feature Extraction Method Based on Video Image

李培玉,夏军,王民,梁军,谢清顺
DOI: https://doi.org/10.13228/j.boyuan.issn1001-0963.2012.06.012
2012-01-01
Abstract:A continuous casting ladle slag detection system in which an image recognized technology was used to detect the slag was designed by comparing machine vision with workers eye. The system was used for researching on continuous casting ladle slag feature extraction method. System platform was built and completed video image signal sampling and processing by software and hardware to realize the real-time online monitoring of continuous casting ladle slag process. As for having no 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 fluid level, structural similarity, statistical properties and combined analysis for three features. Industrial field experiments proved: this detecting system is easy to installation and maintenance, has high degree of automation, long service life, low cost, non-contact detection and a certain success rate of slag detection. It can effectively reduce slag amount, improve the yield of molten steel, reduce the thickness of the tundish slag deposits and has a high practical value .
What problem does this paper attempt to address?