【摘要】 运用数字图像处理技术 ,对烧结机尾的断面图像进行了处理和分析 ,并结合工艺研究成果、专家经验和常规仪表数据 ,解决烧结矿产量的在线推断问题 .安阳钢铁公司烧结厂的现场实验结果表明 ,通过适当的图像模糊增强 ,采用Marr -Heldreth算法着重对烧成带、燃烧带和生料层进行边缘检测等处理 ,系统达到了接近于 1级的预测精度 .该方法还可用于解决烧结终点的闭环控制 ,以及与纹理和颜色相关的质量指标的预报 .更多还原

【Abstract】 This paper applies digital image processing technique to analyze and process sinter image.Meanwhile,according to the expert’s experience and traditional instrument data,its aim consists in solving the problem of on-line deducing sinter production.With the processing of moderate image fuzzy enhancement,this paper uses Marr-Heldreth algorithm to detect the edges of combustion zone,the buming zone and raw layer etc.The result of Anyang Iron & Steel Company’s experiments indicates that this system has been up to and close to 1-grade predicting precision.Moreover,this research may be used to solve sintering terminal loop control and predict quality indicatrix relating to color and grain.更多还原