摘要
针对传统偏振图像伪彩色融合方法存在的不足,提出了一种基于非负矩阵分解和IHS(Intensity Hue Saturation)颜色模型的图像融合方法.首先将偏振信息解析得到的各偏振参量图像作为原始数据集进行非负矩阵分解,得到三幅特征基图像,这些特征基图像包含了场景的大部分偏振信息;然后将三幅特征基图像经直方图匹配之后,分别映射到IHS颜色模型的三个颜色通道,最后变换到RGB颜色空间,得到融合后的图像.实验结果表明,该方法不仅具有较好的色彩表达能力,而且有效地突出了目标的细节信息,提高了图像的可判读性.
To overcome the drawback of conventional false color fusion methods for polarization images, a image fusion approach based on non-negative factorization and IHS（Intensity Hus Saturation）color model is presented. Firstly, all of the polarization parameter images obtained by polarization information analysis are taken as source data set for non-negative matrix factorization, so as to extract three feature basis images containing most of the polarization information of the object. Then, after histogram matching, these three feature basis images are mapped into three color channels of IHS color model. Finally, the fused image is achieved after the transform from IHS to RGB color model. Experiment results have shown that, the proposed method has better color representation capability, and effectively pops out detailed information of objects, and enhances its readability.
出处
《光子学报》
EI
CAS
CSCD
北大核心
2010年第9期1682-1687,共6页
Acta Photonica Sinica
基金
国家自然科学基金(40876095)资助
关键词
光学偏振
图像融合
非负矩阵分解
IHS颜色模型
Optical polarization
Image fusion
Non-negative matrix factorization
IHS color model
