文档介绍：
特征差异驱动的红外偏振与光强图像融合#
安富1,杨风暴1,蔺素珍2,周萧1**
(1. 中北大学信息与通信工程学院,太原 030051;
5
10
15
20
25
30
35
40
2. 中北大学电子测试技术国家重点实验室,太原 030051)
摘要:由于红外偏振与光强成像系统的成像特性差异使得红外偏振与光强图像之间具有较多
互补的差异图像特征,而在原有红外偏振融合算法中只是利用源图像的单一表示与提取的差
异图像特征作为融合测度指标,并且单一差异特征对不确定和随机变化的图像特征信息不能
有效描述,因此就会导致源图像的其它有用信息丢失,不利于融合与识别。所以,本文在分
析源图像差异特征形成机理基础上提出一种基于 NSCT(Nonsubsampled Contourlet
Transform)的红外偏振与光强图像的联合差异特征驱动自适应融合算法。通过实验仿真表
明本文算法比 SVT、小波包和 NSCT 局部能量局部方差更能有效地融合源图像的互补和细
节信息,使其能在实际的目标识别中具有一定的应用价值。
关键词:图像融合;差异特征;红外偏振;红外光强;NSCT
中图分类号:TN911 .73
Feature differences driven fusion of infrared polarization
and intensity images
AN Fu1, YANG FengBao1, LIN SuZhen2, ZHOU Xiao1
(1. Information munication Engineering College, North University of China,
TaiYuan 030051;
2. National Key Laboratory for Electronic Measurement and Technology, North University of
China, TaiYuan 030051)
Abstract: As the imaging feature differences between infrared polarization and intensity imaging
system plementary different image features between infrared polarization and
intensity image, however, using single representation and extraction different image feature of the
source image as fusion measure index in the original infrared polarization fusion algorithms, and
single difference features cannot describe uncertain and random image feature information, thus,
this situation will lead to the loss of other useful information of source image and go against image
fusion and recognition. So this paper puts forward a joint difference feature-driven adaptive fusion
algorithm of infrared polarization and intensity image based on NSCT (Nonsubsampled
Contourlet Transform) in analysis of the source image formation mechanism based on the feature
differences. The simulation results indicate that the proposed algorithm can more effective than
SVT algorithm, wavelet packet algorithm and NSCT local energy and variance algorithm integrate
plementary and detail information of the source image ,which enable it have a ce