Hyperspectral Snapshot Compressive Imaging with Non-Local Spatial-Spectral Residual Network

被引:5
作者
Yang, Ying [1 ]
Xie, Yong [2 ]
Chen, Xunhao [1 ]
Sun, Yubao [1 ]
机构
[1] Nanjing Univ Informat Sci & Technol, Jiangsu Collaborat Innovat Ctr Atmospher Environm, Jiangsu Key Lab Big Data Anal Technol, Nanjing 210044, Peoples R China
[2] Nanjing Univ Informat Sci & Technol, Sch Geog Sci, Nanjing 210044, Peoples R China
基金
中国国家自然科学基金;
关键词
hyperspectral image; coded aperture snapshot spectral imaging; deep network; non-local spatial-spectral attention; compound loss; SPARSE;
D O I
10.3390/rs13091812
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
Snapshot Compressive Imaging is an emerging technology that is based on compressive sensing theory to achieve high-efficiency hyperspectral data acquisition. The core problem of this technology is how to reconstruct 3D hyperspectral data from the 2D snapshot measurement in a fast and high-quality manner. In this paper, we propose a novel deep network, which consists of the symmetric residual module and the non-local spatial-spectral attention module, to learn the reconstruction mapping in a data-driven way. The symmetric residual module uses symmetric residual connections to improve the potential of interaction between convolution operations and further promotes the fusion of local features. The non-local spatial-spectral attention module is designed to capture the non-local spatial-spectral correlation in the hyperspectral image. Specifically, this module calculates the channel attention matrix to capture the global correlations between all of the spectral channels, and it fuses the channel attention attained feature maps and the spatial attention weighted features as the module output, thus both of the spatial-spectral correlations of hyperspectral images can be fully utilized for reconstruction. In addition, a compound loss, including the reconstruction loss, the measurement loss, and the cosine loss, is designed to guide the end-to-end network learning. We experimentally evaluate the proposed method on simulation and real datasets. The experimental results show that the proposed network outperforms the competing methods in terms of the reconstruction quality and running time.
引用
收藏
页数:18
相关论文
共 50 条
[31]   Accumulating Pyramid Spatial-spectral Collaborative Coding Divergence for Hyperspectral Anomaly Detection [J].
Sun, Hao ;
Zou, Huanxin ;
Zhou, Shilin .
2ND ISPRS INTERNATIONAL CONFERENCE ON COMPUTER VISION IN REMOTE SENSING (CVRS 2015), 2016, 9901
[32]   Spatial-Spectral Multiple Manifold Discriminant Analysis for Dimensionality Reduction of Hyperspectral Imagery [J].
Shi, Guangyao ;
Huang, Hong ;
Liu, Jiamin ;
Li, Zhengying ;
Wang, Lihua .
REMOTE SENSING, 2019, 11 (20)
[33]   An efficient unfolding network with disentangled spatial-spectral representation for hyperspectral image super-resolution [J].
Liu, Denghong ;
Li, Jie ;
Yuan, Qiangqiang ;
Zheng, Li ;
He, Jiang ;
Zhao, Shuheng ;
Xiao, Yi .
INFORMATION FUSION, 2023, 94 :92-111
[34]   Deep Spatial-Spectral Representation Learning for Hyperspectral Image Denoising [J].
Dong, Weisheng ;
Wang, Huan ;
Wu, Fangfang ;
Shi, Guangming ;
Li, Xin .
IEEE TRANSACTIONS ON COMPUTATIONAL IMAGING, 2019, 5 (04) :635-648
[35]   Prior Image Guided Snapshot Compressive Spectral Imaging [J].
Chen, Yurong ;
Wang, Yaonan ;
Zhang, Hui .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2023, 45 (09) :11096-11107
[36]   Spatial-Spectral Kernel Sparse Representation for Hyperspectral Image Classification [J].
Liu, Jianjun ;
Wu, Zebin ;
Wei, Zhihui ;
Xiao, Liang ;
Sun, Le .
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2013, 6 (06) :2462-2471
[37]   Spatial-Spectral Constrained Adaptive Graph for Hyperspectral Image Clustering [J].
Zhu, Xing-Hui ;
Zhou, Yi ;
Yang, Meng-Long ;
Deng, Yang-Jun .
SENSORS, 2022, 22 (15)
[38]   Hyperspectral anomaly detection based on spatial-spectral multichannel autoencoders [J].
Jia S. ;
Liu K. ;
Xu M. ;
Zhu J. .
National Remote Sensing Bulletin, 2024, 28 (01) :55-68
[39]   SPATIAL-SPECTRAL MULTIPLE KERNEL LEARNING FOR HYPERSPECTRAL IMAGE CLASSIFICATION [J].
Gu, Yanfeng ;
Feng, Kai ;
Wang, Hong .
2013 5TH WORKSHOP ON HYPERSPECTRAL IMAGE AND SIGNAL PROCESSING: EVOLUTION IN REMOTE SENSING (WHISPERS), 2013,
[40]   Semisupervised classification for hyperspectral image based on spatial-spectral clustering [J].
Wang, Liguo ;
Yang, Yueshuang ;
Liu, Danfeng .
JOURNAL OF APPLIED REMOTE SENSING, 2015, 9