Underwater optical images are crucial in marine exploration. However, capturing these images directly often results in color distortion, noise, blurring, and other undesirable effects, all of which originate from the unique physical and chemical properties of underwater environments. Hence, various factors need to be comprehensively considered when processing underwater optical images that are severely degraded under complex lighting conditions. Most existing methods resolve one issue at a time, making it challenging for these isolated techniques to maintain consistency when addressing multiple degradation factors simultaneously, often leading to unsatisfactory visual outcomes. Motivated by the global modeling capability of the Transformer, this paper introduces TFCNet, a complex hybrid-architecture network designed for underwater optical image enhancement and restoration. TFCNet combines the benefits of the Transformer in capturing long-range dependencies with the local feature extraction potential of convolutional neural networks, resulting in enhanced restoration results. Compared with baseline methods, the proposed approach demonstrated consistent improvements, where it achieved minimum gains of 0.3 dB in the PSNR and 0.01 in the SSIM and a 0.8 reduction in the RMSE. TFCNet exhibited a commendable performance in complex underwater optical image enhancement and restoration tasks by effectively rectifying color distortion, eliminating marine snow noise to a certain degree, and restoring blur.

The satellite Internet of Things (SatIoT) enables real-time acquisition and large-scale coverage of hyperspectral imagery, providing essential data support for decision-making in domains such as geological exploration, environmental monitoring, and urban management. Hyperspectral remote sensing classification constitutes a critical component of intelligent applications driven by the SatIoT, yet it faces two major challenges: the massive data volume imposes heavy storage and processing burdens on conventional satellite systems, while dimensionality reduction often compromises classification accuracy; furthermore, mainstream neural network models are constrained by insufficient labeled data and spectral shifts, frequently leading to misclassification of unknown categories and degradation of cross-regional performance. To address these issues, this study proposes an open-set hyperspectral classification method with dual branches of reconstruction and prototype-based classification. Specifically, we build upon an autoencoder. We design a spectral–spatial attention module and an information residual connection module. These modules accurately capture spectral–spatial features. This improves the reconstruction accuracy of known classes. It also adapts to the high-dimensional characteristics of satellite data. Prototype representations of unknown classes are constructed by incorporating classification confidence, enabling effective separation in the feature space and targeted recognition of unknown categories in complex scenarios. By jointly leveraging prototype distance and reconstruction error, the proposed method achieves synergistic improvement in both accurate classification of known classes and reliable detection of unknown ones. Comparative experiments and visualization analyses on three publicly available datasets: Salinas-A, PaviaU, and Dioni-demonstrate that the proposed approach significantly outperforms baseline methods such as MDL4OW and IADMRN in terms of unknown detection rate (UDR), open-set overall accuracy (OpenOA), and open-set F1 score, while on the Salinas-A dataset, the performance gap between closed-set and open-set classification is as small as 1.82%, highlighting superior robustness.