Hyperspectral target detection remains challenging due to complex nonlinear spectral mixing and insufficient in corporation of spatial context. Existing approaches often address these issues in isolation: bilinear models neglect spatial coherence, kernel-based methods require empirical selection of kernel types and parameters, and many spatial-contextual methods lack explicit modeling of material scattering effects. To address these limitations, we propose a spectral-spatial bilinear decomposition model (SSBDM) that explicitly models scattering-induced nonlinearity and exploits spatial contextual information to enhance target representation under nonlinear mixing. Specifically, a bilinear mixing model with second-order interaction terms is adopted to capture scattering effects beyond linear approximations, providing improved physical interpretability. Moreover, we introduce an adaptive multi-kernel automatic target generation strategy that dynamically fuses multiple kernels based on projection residuals, eliminating manual kernel selection while constructing a high-purity, scene-adaptive background dictionary. Furthermore, we propose dual spectral–spatial regularizations for coefficient estimation, consisting of a superpixel-based graph Laplacian to preserve local spatial homogeneity and mitigate spectral distortion, and a double reweighted sparsity constraint that down-weights low-contributing atoms (particularly second-order interaction atoms) to stabilize bilinear coefficient estimation. This dual design simultaneously enforces spatial coherence and selective sparsity, enhancing robustness and target–background separability under nonlinear mixing. Extensive experiments on four real-world datasets validate the effectiveness and robustness of SSBDM under spectral variability and limited supervision, outperforming state-of-the-art methods. Code will be available at https://github.com/shendb2022/SSBDM.

Benefiting from information complementarity, the fusion of hyperspectral imagery (HSI) and light detection and ranging (LiDAR) data for land cover classification has attracted significant attention in the remote sensing community. However, due to modality imbalance and information redundancy, effectively extracting and integrating complementary knowledge from HSI and LiDAR data remains a major challenge. To address the aforementioned issues, a class semantic prototype guided fusion network (CSPGFNet) is proposed to realize efficient and accurate classification by task-relevant feature mining, fusion, and interaction. Specifically, feature extraction sub-networks with multi-scale and multi-type convolutional structures are designed for each modality to mitigate semantic imbalance caused by inherent dimensional discrepancies. Moreover, a feature fusion-interaction module based on cross-attention mechanism is designed to fuse and interact task-relevant spatial–spectral and elevation information from modalities with class semantic prototype (CSP) as the bridge. Furthermore, a composite loss that incorporates multi-factor constraints is optimized to ensure information complementarity, semantic consistency and task relevance of the whole network. Experimental evaluations on three benchmark datasets demonstrate the effectiveness of the proposed method.