基于Bayes特征判别的选煤建模与识别  

MODELING AND IDENTIFYING OF DISTINGUISHING COAL BASED ON BAYESIAN CHARACTERISTIC DISCRIMINATING

作　　者：刘冠权[1] 李从东[2] 

机构地区：[1]天津大学管理学院,天津博士副教授300072 [2]暨南大学管理学院,广州教授博士生导师510632

出　　处：《煤炭转化》2006年第4期64-68,共5页Coal Conversion

摘　　要：研究了煤块的Harr小波变换和幅度投影分析,在对原始图像进行特征判别分析的基础上,建立了煤块和煤矸石的统计模型,提出了用Bayes判别函数分选煤块和煤矸石的算法,实现了用计算机视觉系统对样本进行了识别检验.并运用神经网络BP算法进行训练识别,检验其正确性,最后通过实验验证了算法有效性.Coal＇s Harr wavelet conversion and range projection analysis are researched. Based on distinguishing and analyzing of original image, statistical modeling of coal and coal gangue is established, arithmetic of applying Bayesian discriminating to distinguish and select coal and coal gangue is put forward, the method of identifying and verifying swatch by computer vision system is achieved. BP neural network arithmetic is applied to train and identify, check out validity. In the end the arithmetic＇s validity is validated by experimentation.

关 键 词：Bayesian判别 统计建模 图像处理 模式识别 

分 类 号：TD921.1[矿业工程—选矿]