Given the problem of spatial detail loss and spectral feature degradation in hyperspectral images (HSIs) characterized as blur, often caused by noise during image acquisition, and methods of removing blur noise designed on HSIs being insufficient, we propose an HSI reconstruction network based on a Blur–Kernel–Prior (BKP) method and Spectral–Spatial Attention (SSA) strategy for noise removal and reconstruction of HSIs. Specifically, a grouping strategy is designed to segment the HSIs into spectral dimension sub-images, and the BKP module, based on U-Net, learns the spatially adaptive blur kernel to extract and remove blurred features from each sub-image while preserving spatial features with spatial resolution. Subsequently, the SSA block is employed to extract shallow features, details, and edge information using a hybrid 2D–3D convolution from the sub-images, followed by deep feature extraction using a deep ResNet and multi-head attention (MSA) on the merged image to maximize the preservation of spectral dimension information. The $L_1$ loss function, combined with spectral dimension loss and peak signal-to-noise ratio loss, is utilized to constrain and ensure reconstruction accuracy. Experiments on both synthetic and real datasets demonstrate that our method exhibits excellent performance in reconstructing HSIs affected by blurred noise, outperforming existing methods in terms of quantitative quality and recovery of spectral dimension information.

Hyperspectral Images (HSIs) classification plays a vital role in precision agriculture and environmental monitoring. However, real-world HSIs is frequently degraded by complex noise, which compromises both spectral integrity and spatial consistency. Existing deep learning (DL) methods, such as Convolutional Neural Networks (CNNs) and Transformers, often face trade-offs between noise suppression and feature preservation, or suffer from high computational costs. To address these challenges, this paper proposes a noise-robust framework named Wavelet-Manifold-Mamba (WMM). Unlike straightforward modular concatenation, WMM establishes a deep coupling between physical-level signal restoration and geometric-level manifold alignment. Specifically, recognizing that the recursive state propagation in Mamba is highly sensitive to input perturbations, a Multi-scale Wavelet Denoising Block (MWDB) is first employed to suppress high-frequency artifacts via learnable cross-frequency attention, thereby stabilizing the token sequence for subsequent modeling. Complementarily, to rectify the feature drift that persists even after denoising, a Manifold-Regularized Mamba Block (MRMB) leverages a topological loss to enforce noise-invariant manifold embeddings. Furthermore, the architecture incorporates a selective scanning mechanism to capture long-range dependencies with linear complexity. Experiments on the WHU-Hi and SZU-Tree datasets demonstrate that the proposed WMM outperforms state-of-the-art methods in noise-robust classification accuracy.