Journal
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
Volume 60, Issue -, Pages -Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGRS.2021.3066485
Keywords
Feature extraction; Hyperspectral imaging; Convolution; Data mining; Correlation; Kernel; Spatial resolution; Attention mechanism; graph convolution network (GCN); hyperspectral image classification (HIC); similarity measurement
Categories
Funding
- State Key Program of National Natural Science of China [61836009]
- National Natural Science Foundation of China [61772401]
- Fundamental Research Funds for the Central Universities [RW180177]
- Open Fund of Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education, Xidian University [IPIU2019007]
- Natural Resources Scientific Research Project of the Department of the Natural Resources of Hunan Province [201910]
- Zhejiang Lab [2021LC0AB05, 2020LC0PI01]
- Open Fund of State Key Laboratory of Geoinformation Engineering [SKLGIE2018-M-4-3]
Ask authors/readers for more resources
This paper proposes a framework based on a deep attention graph convolutional network (DAGCN) to address the challenges of hyperspectral image classification. By integrating attention mechanism and designing deep graph convolutional networks, deep abstract features are extracted and the internal relationship between HSI data is explored, achieving superior classification results compared to the baselines.
Hyperspectral images (HSIs) have gained high spectral resolution due to recent advances in spectral imaging technologies. This incurs problems, such as an increased data scale and an increased number of bands for HSIs, which results in a complex correlation between different bands. In the applications of remote sensing and earth observation, ground objects represented by each HSI pixel are composed of physical and chemical non-Euclidean structures, and HSI classification (HIC) is becoming a more challenging task. To solve the above problems, we propose a framework based on a deep attention graph convolutional network (DAGCN). Specifically, we first integrate an attention mechanism into the spectral similarity measurement to aggregate similar spectra. Therefore, we propose a new similarity measurement method, i.e., the mixed measurement of a kernel spectral angle mapper and spectral information divergence (KSAM-SID), to aggregate similar spectra. Considering the non-Euclidean structural characteristics of HSIs, we design deep graph convolutional networks (DeepGCNs) as a feature extraction method to extract deep abstract features and explore the internal relationship between HSI data. Finally, we dynamically update the attention graph adjacency matrix to adapt to the changes in each feature graph. Experiments on three standard HSI data sets, namely, the Indian Pines, Pavia University, and Salinas data sets, demonstrate that the DAGCN outperforms the baselines in terms of various evaluation criteria. For example, on the Indian Pines data set, the overall accuracy of the proposed method achieves 98.61x0025; when the training sample is 10x0025;.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available