一种新的基于局部小波统计特性的多聚焦图像融合算法  

New Multi-Focus Image Fusion Based on Local Statistic of Wavelet Contrast

作　　者：汪强[1] 刘钢钦[1] 

机构地区：[1]长沙大学计算机科学与技术系,湖南长沙410003

出　　处：《长沙大学学报》2006年第2期72-74,共3页Journal of Changsha University

摘　　要：改进了局部小波对比度的定义.提出一种利用小波变换实现多聚焦图像融合的方法.基于新的对比度的融合过程,由源图像上计算的视觉重要性构造融合图像对应的小波系数,因而融合图像更符合人的视觉特性.仿真结果利用熵、交叉熵、均方根误差对该方法的融合性能进行了评价与分析,实验结果表明,该融合方法充分保留图像的细节信息,并具有良好的稳定性.The definition of the local wavelet contrast is altered. A novel multi - focus image fusion scheme based on multiscale decomposition is presented. According to the new wavelet contrast definition, the wavelet coefficients of the fused image are constructed by using the source image＇ s visual sensitivity. So the result is mere suitable for human vision. With the use of the parameters such as entropy, cross entropy, root mean square error, the performance of the fusion scheme is evaluated and analyzed. The experimental results show that the fusion scheme both preserves the source images＇ details and improves the algorithm＇s stability.

关 键 词：图像融合 小波变换 小波对比度 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]