一种不同分辨率异类图的融合算法

将较高分辨率的图像分解到各个较低分辨率，对最低分辨率的图像采用先加权平均，进而取亮度极大值的融合方法进行融合，并对上述最低分辨率上的融合图像和较高分辨率的图像利用基于Kalman滤波的多分辨图像融合算法进行融合；上述2步重复进行，最后可得最高分辨率下综合所有图像信息的融合图像。实验表明，本文所介绍的图像融合算法是有效的。

Multiresolutional multisource image fusion algorithm

The images with higher resolutions are decomposed into individual images with lower resolutions. The images with the lowest resolution are fused by first weighted averaging and then maximizing their intensity values. The fused image is further fused with the image at the higher resolution by using Kalman filter. The process is carried out successively, and finally the image that fused all the images is achieved at the highest resolution. Two groups of images,i.e.,infrared image and optical image as well as SAR image and optical images are used in image fusion experiments. Experimental results illustrate the feasibility and effectiveness of the algorithm.