SSANet: An Adaptive Spectral-Spatial Attention Autoencoder Network for Hyperspectral Unmixing

doi:10.3390/rs15082070

CORC > 烟台海岸带研究所 > 中国科学院烟台海岸带研究所 > 中科院海岸带环境过程与生态修复重点实验室

	SSANet: An Adaptive Spectral-Spatial Attention Autoencoder Network for Hyperspectral Unmixing
	Wang, Jie 3; Xu, Jindong 3; Chong, Qianpeng 3; Liu, Zhaowei 3; Yan, Weiqing 3; Xing, Haihua 2; Xing, Qianguo 1; Ni, Mengying 3
刊名	REMOTE SENSING
	2023-04-01
卷号	15 期号:8 页码:21
关键词	hyperspectral unmixing spectral-spatial attention mechanism deep learning autoencoder
DOI	10.3390/rs15082070
通讯作者	Ni, Mengying(nimengying@ytu.edu.cn)
英文摘要	Convolutional neural-network-based autoencoders, which can integrate the spatial correlation between pixels well, have been broadly used for hyperspectral unmixing and obtained excellent performance. Nevertheless, these methods are hindered in their performance by the fact that they treat all spectral bands and spatial information equally in the unmixing procedure. In this article, we propose an adaptive spectral-spatial attention autoencoder network, called SSANet, to solve the mixing pixel problem of the hyperspectral image. First, we design an adaptive spectral-spatial attention module, which refines spectral-spatial features by sequentially superimposing the spectral attention module and spatial attention module. The spectral attention module is built to select useful spectral bands, and the spatial attention module is designed to filter spatial information. Second, SSANet exploits the geometric properties of endmembers in the hyperspectral image while considering abundance sparsity. We significantly improve the endmember and abundance results by introducing minimum volume and sparsity regularization terms into the loss function. We evaluate the proposed SSANet on one synthetic dataset and four real hyperspectral scenes, i.e., Samson, Jasper Ridge, Houston, and Urban. The results indicate that the proposed SSANet achieved competitive unmixing results compared with several conventional and advanced unmixing approaches with respect to the root mean square error and spectral angle distance.
WOS关键词	SPARSE COMPONENT ANALYSIS
WOS研究方向	Environmental Sciences & Ecology ; Geology ; Remote Sensing ; Imaging Science & Photographic Technology
语种	英语
WOS记录号	WOS:000977424900001
内容类型	期刊论文
源URL	[http://ir.yic.ac.cn/handle/133337/32899]
专题	烟台海岸带研究所_中科院海岸带环境过程与生态修复重点实验室
通讯作者	Ni, Mengying
作者单位	1.Chinese Acad Sci, Yantai Inst Coastal Zone Res, Yantai 264003, Peoples R China 2.Hainan Normal Univ, Sch Informat Sci & Technol, Haikou 571158, Peoples R China 3.Yantai Univ, Sch Comp & Control Engn, Yantai 264005, Peoples R China
推荐引用方式 GB/T 7714	Wang, Jie,Xu, Jindong,Chong, Qianpeng,et al. SSANet: An Adaptive Spectral-Spatial Attention Autoencoder Network for Hyperspectral Unmixing[J]. REMOTE SENSING,2023,15(8):21.
APA	Wang, Jie.,Xu, Jindong.,Chong, Qianpeng.,Liu, Zhaowei.,Yan, Weiqing.,...&Ni, Mengying.(2023).SSANet: An Adaptive Spectral-Spatial Attention Autoencoder Network for Hyperspectral Unmixing.REMOTE SENSING,15(8),21.
MLA	Wang, Jie,et al."SSANet: An Adaptive Spectral-Spatial Attention Autoencoder Network for Hyperspectral Unmixing".REMOTE SENSING 15.8(2023):21.