基于模糊自适应谐振网的多源遥感图像融合方法  

MULTI-SOURCE REMOTE SENSING IMAGE FUSION METHOD BASED ON FUZZY ADAPTIVE RESONANCE NEURAL NETWORK

作　　者：周宇[1] 张黎宁[1] 高文君[1] 

机构地区：[1]南京林业大学信息技术与科学学院,江苏南京210037

出　　处：《计算机应用与软件》2007年第10期84-86,共3页Computer Applications and Software

摘　　要：介绍模糊自适应谐振网在多源遥感图像融合的应用。详细分析模糊自适应谐振网聚类算法的步骤和特点,并比较模糊自适应谐振网和一般模糊极小-极大网的差异。实验证明模糊自适应谐振网的自适应稳定性佳,其聚类速度优于一般模糊极小-极大网,而一般模糊极小-极大网聚类精度较好,对训练区域的依赖性较强。The application of Fuzzy Adaptive Resonance Theory （FART） to multi-source remote sensing image fusion is introduced. The steps and features of FART clustering algorithm are analyzed in detail, and the differences between FART and General Fuzzy Min-Max neural network （GFMM） are compared. The result shows that the adaptive stability and clustering speed of FART are better than those of GFMM. GFMM is better in clustering accuracy, homever more dependent on training region.

关 键 词：模糊自适应谐振网 一般模糊极小-极大网 多源遥感 图像融合 动态聚类 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]