数字图像处理技术在煤泥浮选泡沫图像纹理特征的提取及识别上的应用  

Application of Digital Image Processing Technology in Extracting and Identifying Vein Characteristic of Foam Picture in Coal Slime Flotation

作　　者：刘文礼[1] 路迈西[1] 

机构地区：[1]中国矿业大学北京校区化工与环境工程系,北京100083

出　　处：《选煤技术》2004年第4期78-81,共4页Coal Preparation Technology

基　　金：国家自然科学基金资助项目 (5 9974 0 32 )

摘　　要：将数字图像处理系统应用于煤泥浮选泡沫图像的特征提取和识别 ,引入了空间灰度相关矩阵和邻域灰度相关矩阵来提取泡沫的纹理特性 ,并提取基于这两种算法的一系列特征参数来描述泡沫的结构 ;分析了各泡沫特征参数随浮选时间 (泡沫纹理 )的变化关系 ,定性地指出了各泡沫特征参数与泡沫纹理的相关性 ;并利用自组织神经网络对煤泥浮选泡沫的状态进行了识别 ,分类识别的平均正确率达 76 5 %。Digital image processing system was used for extracting and identifying the characteristic of foam picture in coal slime flotation. Space gray scale related matrix and neighborhood related matrix were employed to extract foam vein characteristic and a series of characteristic parameters based on the two calculation methods for description of foam structure. Analysis is done in the paper on each foam characteristic parameter, which is function of flotation time. The paper also presents the correlation between each foam characteristic parameter and foam vein as well as identification of the state of foam in coal slime flotation by use of self-organizing neuroid with average exactness percentage up to 76.5%.

关 键 词：数字图像处理 煤泥 浮选泡沫 模式识别 自组织特征映射网络 

分 类 号：TD923.7[矿业工程—选矿]