基于双变换的红外与可见光图像融合增强  

Infrared and visual image fusion enhancement by dual-transform

作　　者：李艳梅[1,2] 陈雷霆[1] 饶云波[3] 罗建[2] 周骏[1] 

机构地区：[1]电子科技大学计算机科学与工程学院,成都611731 [2]西华师范大学计算机学院,四川南充637002 [3]电子科技大学信息与软件工程学院,成都610054

出　　处：《计算机应用研究》2013年第10期3142-3145,共4页Application Research of Computers

基　　金：国家"863"计划基金资助项目(2007AA01Z322);四川省科技厅基础项目(2010JY0151);四川省科技厅支撑项目(2012GZ0020);四川省教育厅重点资助项目(13ZA0011)

摘　　要：通过研究可见光视觉图像与红外图像序列的相关特性及融合处理中存在的问题,提出一种基于视觉图像预增强处理的曲波与小波双变换融合机制来有效保留边缘信息和图像细节信息。实验结果表明所提出的方法提高了融合图像的视觉感知质量,为遮挡和伪装目标的检测和定位提供了更为有效的空间环境,并且融合增强后的图像具有更高的信噪比和更少的人工痕迹,与当前最新的其他方法相比是一种具有更好融合质量和更高效率的图像/视频融合方法。This paper proposed a novel method by studying the characteristics of visual, infrared image/video and the prob- lems introduced from the fusion processing. The proposed method enhanced the illumination of the visual image firstly, and then used curvelet and wavelet dual-transform to save the edge information and preserve the details of the visual and the infra- red images during the fusion enhancement. Experimental results show that the proposed method enhances human perception for image/video and improves camouflaged target＇ s detection and location in complex environment. It is proved to be a high quali- ty and high efficiency image/video fusion method with the characters of higher PSNR and less artifact compared with the other state-of-the-art methods.

关 键 词：图像增强 多频谱图像融合 双变换 伪装目标 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]