一种基于Contourlet域隐马尔可夫树模型的遥感影像融合方法  

REMOTE SENSING IMAGE FUSION METHOD BASED ON CONTOURLET-DOMAIN HIDDEN MARKOV TREE MODEL

作　　者：邓磊[1] 蒋卫国[2] 陈云浩[1] 李京[2] 崔保山[3] 

机构地区：[1]北京师范大学资源学院,北京100875 [2]民政部教育部减灾与应急管理研究院,北京100875 [3]北京师范大学环境学院,北京100875

出　　处：《红外与毫米波学报》2008年第4期285-289,共5页Journal of Infrared and Millimeter Waves

基　　金：国家高技术研究发展计划(2006AA120102);国家自然科学基金(40671130);北京市自然科学基金(40722016)联合资助项目

摘　　要：在多尺度通用传感器模型的框架下,通过结合Contourlet变换和隐马尔可夫树模型,对融合的Contourlet系数进行最大后验概率估计,建立了一种基于Contourlet域隐马尔可夫树模型的遥感影像融合方法.应用该方法和几种常用融合方法对SAR影像和多光谱影像进行实验和对比,结果表明,该方法在提高融合影像空间细节信息的同时,可以有效地保持多光谱影像的光谱信息.Under the multi-resolution universal sensor model, a remote sensing image fusion method based on Hidden Markov tree model was proposed by applying both Contourlet transform and hidden Markov tree model to gain the fused Contourlet coefficient by Maximum A Posteriori（MAP） estimation. Several commonly-used fusion methods were performed between SAR image and multi-spectral data. The fusion results were compared with that proposed in this paper. The experiment indicates that the spectral characteristics can be preserved effectively when the spatial resolution is enhanced.

关 键 词：图像融合 CONTOURLET变换 隐马尔可夫树模型 合成孔径雷达图像 SPOT图像 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]