摘要
将高光谱图像与全色图像融合,所得融合数据对于后续的其它高光谱图像处理非常有帮助。区别于传统方法,针对高光谱图像特点,引入了光谱约束项,改进并建立基于光谱约束的非负矩阵分解(spectral-constrained nonnegative matrix factorization,sc-NMF)。改进后,该模型首先在光谱约束前提下,对高光谱图像进行非负矩阵分解,对分解所得基底进行增强,再重建高光谱图像。这样,所得到的融合图像在空间细节和光谱保持性均有比较好的效果。最后,进行了仿真和实际数据的实验验证,通过主观和客观的评价结果,所改进的融合方法性能较好,比传统方法更适用于高光谱图像融合。
The fusion of hyperspectral image （HSI） and panchromatic image （PI） is a crucial and useful technique. The fused image possesses good spatial and spectral quality, and it is very helpful for the follow-up image processing. By using spectral constrained express, the traditional NMF （nonnegative matrix factorization） was improved, and used it in the hyperspectral image fusion. Firstly, the hyperspectral image was decomposed into basis and weight, then the details of hyperspectral image were sharpened by enhancing the details of the basis with high resolution image. Meanwhile, a spectral constraint function was added in the model to preserve the spectral information. Therefore, the fused image obtained by the proposed fusion model possesses good spatial and spectral information at the same time. At last, the experiments on simulated and real data were done with conventional and the proposed methods. The proposed method behaves better both in visual and objective indices, indicating it is a better choice for HSI fusion.
出处
《红外与激光工程》
EI
CSCD
北大核心
2013年第10期2718-2723,共6页
Infrared and Laser Engineering
基金
国家自然科学基金(61273245
60975003
91120301)
国家重点基础研究发展计划(2010CB327904)
虚拟现实技术与系统国家重点实验室开放基金(BUAA-VR-12KF-07)
教育部新世纪优秀人才支持计划(NCET-11-0775)
北京市自然科学基金(4112036)
关键词
高光谱图像融合
非负矩阵分解
光谱保持
质量评价
hyperspectral image fusion
nonnegative matrix factorization
spectral preservation
quality analysis
