基于Contourlet域HMT似然概率的图像融合方法

传统的基于Contourlet变换的图像融合方法大都 忽略了Contourlet系数之间 的相关性,导致特征信息的丢失。本文根据隐马尔可夫树(HMT)模型的两种状态和 3组概率确定能有效捕获尺度间、尺度内的Contourlet系数特性的似然概率,设计了图像融 合规则。实验结果表明,Contourlet域HMT模型应用于图像融合领域,能充分挖掘数据之间 的相关性,为融合图像提取更全面、准确的特征纹理信息。

An image fusion algorithm based on likelihood function of HMT in Contourlet domain

Traditional image fusion method based Contourlet transform always ignores the relationship of the Contourlet coefficients.This paper proposes a novel ima ge fusion method based on Contourlet domain hidden Markov tree(HMT) model.The fusion method can strengthen the relationship among Contourlet coefficients and extract more detailed and exact information from the original images.Firstly,the original i mages are decomposed using Contourlet transform.Secondly,different frequenc y bands have different characteristics,so this paper designs the different fusion rules in different frequencies.The proposed method calculates likelihood function according to the parameters of Contourlet HMT.Finally the fused coefficients are reconstructed to obtain fusion results.Two sets of images are taken as experimen tal data,and subjective and objective standards are used to evaluate the results.Experimental results have verified the simplicity and effectiveness of the method.The results show t hat the proposed method can preserve much more information.The proposed fusion rule based on likelihood function can extract much more and exact characteristics for fused images.And it is an effec tive and feasible algorithm.