Nearest Neighboring Self-Supervised Learning for Hyperspectral Image Classification

Recently, state-of-the-art classification performance of natural images has been obtained by self-supervised learning (S2L) as it can generate latent features through between different views the same images. However, semantic information similar hardly exploited these S2L-based methods. Consequently, to explore potential S2L samples in hyperspectral image (HSIC), we propose nearest neighboring (N2SSL) method, interacting augmentations reliable pairs (RN2Ps) HSI framework bootstrap your own (BYOL). Specifically, there are four main steps: pretraining spectral spatial residual network (SSRN)-based BYOL, generation (N2Ps), training BYOL based on RN2P, final classification. Experimental results three benchmark HSIs validated that facilitate subsequent Moreover, found trained an un-related be fine-tuned for other with less computational cost and higher accuracy than from scratch. Beyond methodology, present a comprehensive review HSI-related data augmentation (DA), which is meaningful future research HSIs.