【摘要】 钢渣层厚度的判定对于RH精炼过程十分重要，传统判定多由经验丰富的炼钢厂工人进行，但人眼观测的局限性会导致操作的不准确。考虑到钢渣和钢液会在浸渍管外部留下不同颜色的印记，本文利用分水岭算法对浸渍管外部信息进行分割，改变相机的曝光量获取两幅不同的浸渍管前景目标，使用边缘坐标在垂直距离上的移动量来确定钢渣层的厚度。通过实验分析，本文方法能够准确地分割出浸渍管外部钢渣语义区，并推测出钢渣层的厚度。更多还原

【Abstract】 Determination of steel slag layer thickness is very important for RH refining process, which is mainly carried out by experienced steelmaking workers traditionally. However, limitation of human eye observation could lead to inaccurate operation. Considering that steel slag and liquid steel will leave different color marks on the outside of the impregnation tube, this paper used the watershed algorithm to segment the external information of the impregnation tube, changed the exposure of the camera to obtain two different foreground targets of the impregnation tube, and used the movement of the edge coordinates in the vertical distance to determine the thickness of the steel slag layer. Through experimental analysis, this method could accurately segment the semantic region of steel slag outside the impregnation pipe and infer the thickness of steel slag layer.更多还原