基于模糊证据理论分类的多源遥感信息融合  

MULTIPLE SOURCES DATA FUSION USING FUZZY DEMPSTER-SHAFER EVIDENTIAL REASONING APPROACH TO CLASSIFICATION

作　　者：刘纯平[1] 戴锦芳[2] 钟文[1] 王元全[1] 夏德深[1] 

机构地区：[1]南京理工大学计算机科学系 [2]中国科学院南京地理与湖泊研究所,南京210008

出　　处：《模式识别与人工智能》2003年第2期213-218,共6页Pattern Recognition and Artificial Intelligence

基　　金：江苏省高技术研究资助项目(No.BG2001324-2)

摘　　要：运用模糊理论和证据理论相结合的方法融合来自遥感领域中不同数据源的证据,以实现各种数据所含互补信息的融合,同时减少不同数据间的信息冗余。通过模糊理论的引入,基于模糊证据理论对多源数据的融合可以更好地处理分类中的不确定性。最后以中巴地球资源一号卫星(CBERS-1)的CCD与Landsat5的TM图像进行决策层融合分析,试验结果表明该方法在多源信息融合的遥感图像小目标自动分类中具有可行性和有效性,并有很好的应用前景。Fuzzy and evidential theory method can be applied to fuse the evidences from multi-source information in remote sensing field. The fusion of all kinds commutative compensatory data can be realized and the redundant information of different kinds of data can be cut down using these methods. By introducing the fuzzy theory, the fusion of the multi-source data based on the theory of fuzzy evidential reasoning method can deal better with the uncertainty in the classification than other methods. At last, the analysis of the decision-making layer fusion is done based on the CCD of China Brazil Earth Resource Satellite-! (CBERS-1) and the TM image of Landsat5. The experimental result indicates that this method is feasible and effective to the automatic classification of small target in the remote sensing image based on multi-source data fusion and has a better application foreground.

关 键 词：遥感图像 信息融合 模糊理论 证据理论 信息冗余 数据融合 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]