【摘要】 在覆盖剂黏着于测温管表面的连铸中包中,基于视觉方法测量钢水液位系统无法通过温度场分析测量钢水液位.为此,本文使用光流分析算法跟踪测温管拔出钢水后表面黏着并向下流动的覆盖剂,检测出覆盖剂在测温管不同位置黏着特性的差异,并从微观角度分析覆盖剂在测温管表面的黏着特性与钢水液位的联系,最终计算出钢水液位.经验证,该方法所测得钢水液位误差在3 mm以内.大量现场试验表明,该方法能较好地解决覆盖剂遮挡测温管表面温度信息的问题,使基于视觉方法的钢水液位测量系统在测温管黏着覆盖剂的异常情况下能准确可靠地测量出钢水液位.更多还原

【Abstract】 In some continuous caster,slag often adheres to the temperature-measuring tube. This phenomenon disturbs the vision-based molten steel level measurement system analyzing the temperature field of the temperature-measuring tube. To solve this problem,the liquid slag adhered on the tube lifted from the molten steel was tracked via optical flow analysis and the adherent characteristic in the different position of the tube was gathered,then the relation between the adherent characteristic and the molten steel level prior to lifting was analyzed,finally the molten steel level in the tundish was calculated. Validated by other methods,the error of this method is less than 3 mm. Lots of on-site experiments indicate that this method may well solve the problem of missing temperature data,and accurately measure the molten steel level in the exceptional case of temperature-measuring tube adhering with liquid slag.更多还原