视觉匀度分析的改进梯度指数法及其应用  

Advanced gradient index analysis method for the evenness based on computer vision and its application

作　　者：常发亮[1] 王平[1] 乔谊正[1] 季鸿雁[1] 

机构地区：[1]山东大学控制科学与工程学院,山东济南250061

出　　处：《控制与决策》2004年第10期1125-1128,1132,共5页Control and Decision

基　　金：国家自然科学基金资助项目(60104009);山东省自然科学基金资助项目(Y2001G06).

摘　　要：以纸张为研究对像,利用计算机视觉方法,在分析了标准匀度检测方法的基础上,采用类金字塔结构方差给出了改进的梯度指数纹理分析算法,它既体现了微观匀度特性,又体现了宏观匀度特性.实验结果表明,该方法对于分析纸张匀度是可行的,并具有较高的鲁棒性能.An advanced gradient index texture analysis algorithm based on computer vision methods is proposed with paper as the research object, which uses pyramid-like structure squared variance based on the standard evenness detecting method. This method has both micro evenness characteristics and the macro ones. Experimental results indicate that the method is feasible for analyzing paper evenness and is of higher robustness.

关 键 词：计算机视觉 匀度分析 纹理分析 梯度指数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]