论文部分内容阅读
结构图像分析能定量地在线显示工业锌浮选中浮选泡沫的特性。本文用邻近灰度水平相关矩阵来定量地确定泡沫的结构.原矿的变化和载流分析的不精确性,致使图像特征与中期锌回收率之间的相关关系较差.正如自组织神经网络所示的那样,只要能识别锌品位和图像特征之间的短期干扰,那么两者之间就一定存在着一种重要的相关关系。磨矿回路的波动和图像特征的变化这两者之间的关系预示着图像分析可用来作为一种诊断工具。
Structural image analysis can quantitatively show online the characteristics of flotation foam in industrial zinc flotation. In this paper, the structure of the foam is quantitatively determined using the adjacent gray level correlation matrix. Changes in ore and inaccuracies in current-carrying analysis led to poorer correlations between image characteristics and medium-term zinc recovery. As the self-organizing neural network shows, there is an important correlation between the two as long as the short-term interference between the zinc grade and the image features can be identified. The relationship between fluctuations in the grinding circuit and changes in image characteristics indicates that image analysis can be used as a diagnostic tool.