SHSRD: Efficient Conditional Diffusion Model for Single Hyperspectral Image Superresolution

The emergence of deep neural networks has spurred progress in single hyperspectral image (HSI) super-resolution. However, mainstream models prioritize network architecture and optimization, overlooking the limitations of hyperspectral image datasets' small scale and potential data imbalance. Direct training models with such datasets can lead to issues like overfitting. To better solve the above problems from the perspective of dataset, we propose SHSRD, an advanced super-resolution framework specifically designed for hyperspectral images based on diffusion model. It incorporates a spectral information injection module, which selectively introduces diverse spectral information into the model in a conditional manner, thereby enabling efficient spectral information perception. Furthermore, a two-stage training strategy is meticulously devised. Initially, the model undergoes pre-training on a large-scale natural image dataset. Subsequently, leveraging transfer learning techniques, the knowledge acquired from the natural image dataset is seamlessly transferred to the single hyperspectral image super-resolution task. This strategic approach culminates in the realization of high-quality super-resolution for single hyperspectral images, effectively addressing the challenges associated with limited training data in hyperspectral imaging. It also demonstrates significant generalization, with only one pre-training session on a natural image dataset required for scale-specific tasks, allowing for rapid transfer to any small HSI dataset. Extensive experiments on four public datasets demonstrate that SHSRD outperforms state-of-the-art methods.