摘要
为了克服传统图像融合算法在权值选择时忽略了图像中一些重要信息,从而造成图像细节丢失和出现模糊现象。首先介绍了梯度域融合的基本原理,并在此原理基础上提出了一种基于梯度域信息图像质量评估因子的图像融合算法,通过对图像质量进行一定评估,再利用梯度域信息对权值进行设计。该算法有效结合多张图像中梯度信息以及图像间在梯度方向相对差值信息。实验结果表明该算法能够充分融合不同图像中的信息,最后融合的图像保持了原始图像的色彩、细节丰富。
The traditional image fusion algorithm ignores some important information in weight selection, which leads to the loss of image details and fuzzy phenomenon. The basic principle of gradient domain fusion is first introduced, and based on this principle, an image fusion algorithm based on gradient domain information image quality assessment factor is presented, and through the evaluation of image quality, the gradient domain information is used to design weights. The algorithm effectively combines the gradient information in multiple images and the relative differences between the images in the gradient direction. Experimental results show that the proposed algorithm can fully integrate the information in different images, and finally the fused image maintains the color of the original image with rich details.
出处
《控制工程》
CSCD
北大核心
2016年第8期1234-1237,共4页
Control Engineering of China
关键词
图像融合
梯度域
评估因子
权值设计
相对差值信息
Image fusion
gradient domain
evaluation factor
weight design
relative difference information
