基于纹理特征回转窑熟料烧结状态分类.docVIP

基于纹理特征回转窑熟料烧结状态分类 　　摘要:本文采用灰度共生矩阵方法,利用Fish系数提取出最佳分类位置算子和纹理特征参数,通过对实际回转窑窑头熟料图像分析,发现位置算子为(5,-5)即距离为5、方向为45°下的灰度共生矩阵对应的和平均、逆差距、差异熵、对比度、差方差和熵这六个参数具有较好的区分度,其表面纹理特征能客观地反映其烧结程度,并通过基于C4.5算法实现了过烧、欠烧和正常烧结三种不同状态下的熟料纹理分类,其精度达到了95.65%。同时结合实际工况对熟料纹理进行了分析,给出了各自的变化特点。 　　关键词:回转窑,熟料,纹理,灰度共生矩阵,Fisher 系数,C4.5算法 　　中图分类号:TP2文献标识码: A 国家标准学科分类与代码: 520.6040 　　Classification of Sintered Clinker in Rotary Kiln Based on Texture Features 　　HE Min1,ZHANG Jing1,YAN Min2,CHEN Hua3 　　 (1.College of Electrical and Information Engineering, Hunan University, Changsha, Hunan, 410082, China2.College ofPhysics and Microeletronic Science, Hunan University, Changsha, Hunan, 410082, China 　　3. School of ComputerCommunication, Hunan University, Changsha, Hunan, 410082,China ) 　　Abstract:The texture analysis of the clinker image based on the gray-level co-occurrence matrix is proposed to predict the clinker’s sintered mode. The best position operator and feature sets of the gray-level co-occurrence matrix are extracted by Fisher coefficient. Subsequently, these reduced features are applied by C4.5 to classify these clinker images into three categories: over-sintered, less-sintered and normal-sintered. The experimental results show that six texture features, which are Sum Average、Inverse Difference Moment、Difference Entropy、Contrast、Difference Variance and Entropy of the gray-level co-occurrence matrix under the position operator (5,-5) have the highest degree of discrimination, and the classifier accuracy reaches 95.65% using C4.5 classifier. Finally, the difference between these three kinds of clinker textures is summarized in this paper. 　　Key words:Rotary kiln, Clinker, Texture, Gray-level co-occurrence matrix, Fisher coefficient, C4.5 algorithm 　　 　　1 引言 　　在实际的回转窑生产领域,利用机器视觉方法对窑头熟料特征进行分析,有助于实现窑内实际工况的判断,对提高回转窑的自动控制水平具有重要意义。窑头图像中熟料的休止角反映了熟料运动模式,可用于判断其黏度、带高度[1];熟料填充率变化趋势能够反映窑尾进料量以及窑内温度的异变[2]。除此之外,熟料图像的纹理特征能较好地描述其表面的物理属性(如粗糙度、颗粒度),从而间接反映熟料在窑内的烧结状态。基于纹理这一特征属性进行模式识别在气象学、医学、地理学和工业领域已经得到了广泛应用[3-5]。在常