Abstract
Hyperspectral imagery (HSI) classification, which attempts to assign hyperspectral pixels with proper labels, has drawn significant attention in various applications. Recently, the graph-based semisupervised learning (SSL) approaches have shown the outstanding ability to handle the situation of limited labeled data for classification task. However, it is hard to construct the pairwise adjacent graph due to the high dimensionality of hyperspectral data. Besides, the graph-based SSL models are usually decided by a single classifier, which fail to effectively learn the complex structures and intrinsic properties of HSI. To address these problems, we propose a novel graph-based SSL classification model for HSI, which is based on random multigraphs construction and ensemble strategy (RMGE). Specifically, the anchor graph (AG) is constructed with spatial–spectral features, which integrates the spatial characteristics extracted by local binary pattern on each selected spectrum, preserving fine structures of local region. In order to enhance the discriminative capability of the classifier and avoid the trivial solution, the maximum entropy regularization is added into adjacent AG model. In addition, to capture the diversity of HSI data effectively, we design the ensemble framework by employing multiple AGs to learn HSI features. Experiments conducted on real hyperspectral datasets indicate that the proposed RMGE shows better performance than that of state-of-the-art approaches.
Highlights
H YPERSPECTRAL imagery (HSI) provides abundant spatial structures and spectral characteristics of the land cover classes, which is obtained by utilizing hyperspectral sensors mounted on different platforms [1], [2]
To handle the defect that anchor graph (AG)-based methods usually ignore the spatial information, Wang et al [47] presented a AG-based method with clustering for HSI data, which fuses the spatial information by using the mean of neighboring pixels to reconstruct center pixel
An efficient graph-based Supervised Learning (SSL) classification model with spatial-spectral features is proposed for HSI, which is based on random multi-graphs and ensemble strategy
Summary
H YPERSPECTRAL imagery (HSI) provides abundant spatial structures and spectral characteristics of the land cover classes, which is obtained by utilizing hyperspectral sensors mounted on different platforms [1], [2]. Lots of supervised approaches [9]–[12] have been developed for HSI task He et al [13] proposed a supervised model for HSI classification with limited training samples, which extracts features exiting in and among samples and learns the label distribution. In the past two decades, the graph-based SSL models have obtained widespread attention [17], [18], which explore the pairwise adjacent graph between pixels and capture the data structure These approaches generally suffer from the high complexity associated with eigenvalue decomposition on graph Laplacian [19]. The Indian Pines contains 16 land-cover types whose distribution is illustrated in Fig. 2 (b). kg OA
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
More From: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.