基于非局部联合稀疏模型的高光谱图像超分辨率重建
首发时间:2015-12-07
摘要:本文提出了一种基于非局部联合稀疏模型的高光谱图像超分辨率重建算法。该算法首先利用在线字典学习方法对低空间分辨率的高光谱训练图像集合进行字典训练,获得对应的谱域字典;然后借助同一场景的全色图像,对相似像素点进行联合稀疏表示并重建高分辨率图像;最后利用迭代反向投影图像后处理技术对高分辨率重建图像进行处理,得到重建误差更小,视觉重建质量更高的高分辨率高光谱图像。实验结果表明,本文提出的算法能够在保持图像谱域信息完整的同时,在空域更加有效的重建图像的边缘、纹理等结构特征。
关键词: 计算机应用技术 超分辨率重建 高光谱图像 谱域字典 非局部自相似 联合稀疏模型
For information in English, please click here
Hyperspectral Image Super-resolution Based on Nonlocal Joint Sparse Model
Abstract:In this paper, a hyperspectral image super-resolution method based on nonlocal joint sparse model is proposed. The proposed algorithm trains the spectral dictionary with the low-resolution hyperspectral image via an online dictionary learning method. In addition, the algorithm encodes the similar pixels via a joint sparse model and reconstruct the high resolution image, with the help of a panchromatic image of the same sense. Finally, the algorithm apply iterative back projection operator on the reconstructed image to further reduce the reconstruction error. Experimental results suggest that the proposed algorithm can reconstruct the edge and texture of the image effectively, while keep the spectral information completely.
Keywords: Technology of Computer Application Super-resolution Reconstruction Hyperspectral Image Spectral Dictionary Non-local Self-similarity Joint Sparse Model
论文图表:
引用
No.4663989111227214****
同行评议
共计0人参与
勘误表
基于非局部联合稀疏模型的高光谱图像超分辨率重建
评论
全部评论0/1000