基于粗糙集核计算的遥感图像波段优选算法  

Optimal bands selection of remote sensing image based on core attribute of rough set theory

作　　者：陈敏[1] 

机构地区：[1]福州大学管理工程系,福建福州350002

出　　处：《宁德师专学报（自然科学版）》2006年第4期378-380,384,共4页Journal of Ningde Teachers College(Natural Science)

基　　金：福建省教育厅资助项目(K03013);福建省空间信息中心资助项目(01000301)

摘　　要：针对遥感图像波段选择传统手工方法的不足,应用粗糙集理论,从纹理分析的多个角度选择特征,多方面挖掘数据的关键属性;通过核属性的方法计算达到简化决策表中的条件属性的目的,得到图像分类的关键波段和一些次关键波段;实现了波段的自动选择.该算法比传统方法更客观更简便,为遥感图像波段选择提供了崭新的方法.In this paper, the method of getting optimal bands from remote sensing image is discussed. Firstly, by analyzing the features of texture, it gets some key attributes of data at multi - aspects. Secondly, according to the method of calculating the core, condition attributes of decision table are reduced. Finally, optimal bands and some sub - optimal bands are gotten. By this way it can overcome the defect of traditional method which gets the result by hand. The algorithms is objective and simple, it provides a new approach to bands selection.

关 键 词：遥感图像 粗糙集 差别矩阵 核属性 优选波段 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]