Abstract:Clustering large hyperspectral images (HSIs) is a very challenging problem because large HSIs have high dimensionality, large spectral variability, and large computational and memory consumption. Recently, sparse subspace clustering (SSC) has achieved remarkable success in HSI clustering. However, most SSC-based methods suffer from the following bottlenecks for large HSIs: 1) high computational consumption and memory space during the construction of the similarity matrix and decomposition of the graph Laplacian matrix and 2) failure to capture the relationships among dictionary atoms, sparse coefficients, and hyperspectral pixels. To address these challenges, we propose a novel algorithm that extends SSC to cocluster large HSIs, called bipartite graph partition with graph nonnegative matrix factorization (BGP-GNMF). Specifically, to fully explore the characteristics of the spectral and spatial contexts in HSIs, we propose a novel superpixel and pixel coclustering framework with bipartite graph partitioning in the joint sparse representation domain, where superpixel-based dictionary atoms are defined as disjoint vertex sets of the bipartite graph and the joint sparsity representation is mapped into the adjacency matrix of the undirected bipartite graph. To overcome the challenges of high computational consumption and large memory space for large HSIs, the bipartite graph partition with orthonormal constrained nonnegative matrix factorization is proposed to simultaneously cluster the structured dictionary atoms and hyperspectral pixels with an indicator matrix. Finally, to exploit the intrinsic geometry of HSIs, we incorporate manifold regularization into the bipartite graph partition to improve final clustering accuracy. The effectiveness and efficiency of the proposed method are verified on three classical HSIs, and the experimental results illustrate the superiority of the proposed method compared with other state-of-the-art HSI clustering methods.

Spectral-Spatial Superpixel Anchor Graph-Based Clustering for Hyperspectral Imagery

Spatial–Spectral Constrained Adaptive Graph for Hyperspectral Image Clustering

Attention in Attention for Hyperspectral with High Spatial Resolution (H) Image Classification

Superpixel-Level Global and Local Similarity Graph-Based Clustering for Large Hyperspectral Images

An Efficient Algorithm for Hyperspectral Image Clustering

Hyperspectral Image Clustering by Superpixel- Based Low-Rank Constrained Bipartite Graph Learning

Graph-Based Structural Deep Spectral-Spatial Clustering for Hyperspectral Image

Superpixel Graph Contrastive Clustering with Semantic-Invariant Augmentations for Hyperspectral Images

Spatial and Cluster Structural Prior-Guided Subspace Clustering for Hyperspectral Image

Fast Spectral Clustering With Anchor Graph for Large Hyperspectral Images.

Deep Spatial-Spectral Subspace Clustering for Hyperspectral Image

Fast Spectral Clustering via Efficient Multilayer Anchor Graph

Hyperspectral Image Classification Based on Superpixel Feature Subdivision and Adaptive Graph Structure

Dual Smooth Graph Convolutional Clustering for Large-Scale Hyperspectral Images

Deep clustering using 3D attention convolutional autoencoder for hyperspectral image analysis

Spectral–Spatial Residual Graph Attention Network for Hyperspectral Image Classification

A Bipartite Graph Partition-Based Coclustering Approach With Graph Nonnegative Matrix Factorization for Large Hyperspectral Images

Incorporating Attention Mechanism and Graph Regularization into Cnns for Hyperspectral Image Classification

A Spectral-Spatial Classification of Hyperspectral Images Based on the Algebraic Multigrid Method and Hierarchical Segmentation Algorithm

Unsupervised Self-Correlated Learning Smoothy Enhanced Locality Preserving Graph Convolution Embedding Clustering for Hyperspectral Images

Spectral-Spatial Joint Noise Estimation for Hyperspectral Images