Research On Fusion Algorithm Of Multi-source Images

In recent years,the research of image fusion has played an important role in both military applications and medical applications.The key and difficult point of image fusion is to effectively highlight the target image and enhance the visual effect of the fusion effect map,so that people can get the useful information more intuitively.The fusion algorithm of current research has basically focused on how to use the simple fusion to achieve the above purpose while ignoring the role of image enhancement.This article will conduct the research from two aspects of image enhancement and fusion,to explore a fusion algorithm with better fusion effect.The main contents of this thesis are as follows:1.A variety of image enhancement algorithms are studied.It includes adaptive histogram equalization algorithm,MSRCR algorithm,image dark channel fogging algorithm and so on.An image enhancement algorithm based on the different characteristics of visible and infrared images and an enhancement algorithm of image curve adjustment for glare images are proposed.2.The three levels of image fusion and their framework are introduced.Several classic image fusion algorithms based on pixel-level fusion are studied and analyzed,including Laplacian pyramid fusion algorithm,wavelet pyramid fusion algorithm and NSCT fusion algorithm,and conduct an experimental comparison.Considering the image evaluation index,select the NSCT fusion algorithm for further research.3.Compared the original images with enhanced images.Experimental results show that the image after the image enhancement algorithm has a great effect on the final fusion results,which confirms the effectiveness and necessity of image enhancement algorithm in this thesis.4.With the objective of the image fusion as the starting point,the purpose of image fusion is to highlight the observed objects.Therefore,a NSCT image fusion algorithm based on image saliency detection is proposed.The experimental results show that the algorithm can effectively improve the image fusion effect.