Entropy based multi-resolution visible-infrared image fusion

Abstract

Image fusion is a useful technique for merging multi-sensor images to enhance the image information. Multi-resolution image fusion is one class of most popular image fuse algorithms. Many attentions were given to highband coefficients fusion algorithms, but few for baseband. This paper proposed a baseband fusion algorithm based on image information entropy for multi-resolution image fusion. Baseband coefficients of fused image were obtained by weighted sum of two source images and weights for every channel are proportional to their entropy. Experiments were taken on 5 different multi-resolution image fusion algorithms based on 11 pairs of public test visual-infrared images. Experimental results show that entropy based fusion algorithms give better results than ordinary average merge or simply select one channel.