摘要
针对红外和可见光图像融合结果背景信息不足、对比度较低的问题,提出一种结合引导滤波和快速共现滤波的融合方法。首先,以高斯滤波将源图像分解为细节层和基础层。然后以去除值域滤波器、全局统计共现信息的方式简化共现滤波,形成快速共现滤波,再用其融合细节层;此外,引入窗口因子,用图像大小与窗口因子比值确定引导滤波窗口值,再用其融合基础层。实验结果表明该算法增加了图像背景细节,提高了人物与背景的对比度。主观和客观的实验分析验证了该算法的有效性。
Aiming at the problem of insufficient background information and low contrast of the fusion result of infrared and visible light images,this paper proposed a fusion method combining guided filtering and fast co-occurrence filtering.First,it decomposed source images into a detail layer and a base layer by Gaussian filtering.Then it simplified the co-occurrence filtering by removing the range filter and global statistical co-occurrence information to form a fast co-occurrence filter,and then used it to fuse the detail layer.In addition,it introduced a window factor,and determined window for guided filter by the ratio of the image size and the window factor,then used it to fuse the base layer.Experimental results show that the algorithm increases the background details of the image and improves the contrast between the person and the background.Subjective and objective experimental analysis verifies the effectiveness of the algorithm.
作者
朱文鹏
陈莉
张永新
Zhu Wenpeng;Chen Li;Zhang Yongxin(School of Information Science&Technology,Northwest University,Xi’an 710127,China;School of Information Technology,Luoyang Normal University,Luoyang Henan 471934,China)
出处
《计算机应用研究》
CSCD
北大核心
2021年第2期600-604,610,共6页
Application Research of Computers
基金
中国博士后科学基金项目(2015M582697)。
关键词
图像融合
红外和可见光图像
快速共现滤波
引导滤波
image fusion
infrared and visible images
fast co-occurrence filter
guided filter