Abstract
Graph convolutional neural networks (GCNs) have been widely used in hyperspectral images (HSIs) classification for their superiority in processing non-Euclidean structure data. The performance of GCNs relies on the initial graph structure. Most GCN models only utilize spectral information to construct a graph, which is inaccurate because they fail to take the relationship between adjacent nodes into consideration. In addition, due to the over-smooth phenomenon, most GCN models are shallow and unable to extract effective features. To address these issues, a dual graph u-nets is proposed by integrating spatial graph and spectral graph for HSIs classification, denoted by DGU-HSI. To integration the spectral and spatial information, two graphs are constructed for feature extraction. The spectral graph is created by spectral similarity among all pixels where multirange spectral information is retained, and the spatial graph is constructed by exploiting the neighborhood relationship of the center pixel, which describes spatial information. Then, a dual GCN is utilized to extract spatial and spectral graph features simultaneously. To relieve the over-smooth phenomenon, the graph u-nets structure is applied on constructed spectral and spatial graph to extract effective features. Finally, the extracted spectral and spatial features are fused for classification. Experiments conducted on the public datasets demonstrate the effectiveness of the proposed method on HSIs classification.
Highlights
H YPERSPECTRAL images (HSIs) contain hundreds of continuous and narrow spectral channels with rich spatial and spectral information
Hl is the feature of Graph convolutional neural networks (GCNs) in lth layer and W l represents the weighs of lth layer. σ(·)is the activation function (i.e., ReLU)
The proposed DGU-HSI utilizes spectral similarity across different pixels to construct a spectral graph for HSIs classification
Summary
H YPERSPECTRAL images (HSIs) contain hundreds of continuous and narrow spectral channels with rich spatial and spectral information. Better classifiers from machine learning, such as e support vector machine (SVM) [8] and random forest [9], have been applied in HSIs classification to obtain satisfactory classification performance These methods employ spectral information of a single pixel to determine its class label and have the advantage of conceptual simplicity and easy implementation. A common feature of these methods is that CNN can only be conducted on data with a grid-like structure They cannot naturally catch the geometric characteristics variations of different regions in HSIs. Besides, due to the spectral similarity and spectral variability among pixels, the relevant characters of spectral cannot be extracted adequately with a fixed convolutional kernel.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
