基于Curvelet变换多聚焦图像融合  

Multifocus Image Fusion Using Curvelet Transform

作　　者：田闯[1] 刘文波[1] 

机构地区：[1]南京航空航天大学自动化学院,江苏南京210016

出　　处：《计算机技术与发展》2008年第7期29-30,34,共3页Computer Technology and Development

基　　金：江苏省自然科学基金资助项目(BK2001047);航空科学基金资助项目(04D52032)

摘　　要：探讨了可见光多聚焦图像的融合问题,提出了一种基于Curvelet变换的图像融合算法。针对Curvelet分解的不同频率域,分别讨论了低频系数和高频系数的选择原则。在选择高频系数时,通过引入Wronskian行列式从而定义局部区域线性相关度,并根据该线性相关度进行高频系数的选择;在选择低频系数时,直接采用平均法。实验结果表明:文中所给出的融合算法能够得到多个目标聚焦都很清晰的图像。An algorithm suitable for fusion of visible light multifocus images is researched in this paper. This algorithm is base on curvelet transform. According to the different frequency areas decomposed by curvelet transform, the selection principle of the low frequency coefficients and the high frequency coefficients were discussed respectively. In choosing the high frequency coefficient, the concept of the local area linear dependency was defined by introducing Wronskian determinant and a selection principle based on the linear dependency was presented. In choosing the low frequency coefficient, the method of average was employed immediately. The experimental results show that the proposed algorithm can make all targets in the fused images very dear.

关 键 词：图像融合 CURVELET变换 线性相关度 WRONSKIAN行列式 

分 类 号：TN911.73[电子电信—通信与信息系统]