Unsupervised Classification in Hyperspectral Imagery with Nonlocal Total Variation and Primal-Dual Hybrid Gradient Algorithm

We propose a graph-based nonlocal total variation method (NLTV) for unsupervised classification of hyperspectral images (HSI). The variation problem is solved by the primal-dual hybrid gradient (PDHG) algorithm. By squaring the labeling function and using a stable simplex clustering routine, we can implement an unsupervised clustering method with random initialization. Finally, we speed up the calculation using a $k$-d tree and approximate nearest neighbor search algorithm for calculation of the weight matrix for distances between pixel signatures. The effectiveness of this proposed algorithm is illustrated on both synthetic and real-world HSI, and numerical results show that our algorithm outperform other standard unsupervised clustering methods such as spherical K-means, nonnegative matrix factorization (NMF), and the graph-based Merriman-Bence-Osher (MBO) scheme.