Dual-Diffusion: Dual Conditional Denoising Diffusion Probabilistic Models for Blind Super-Resolution Reconstruction in RSIs

Previous super-resolution (SR) reconstruction works are always designed on the assumption that the degradation operation is fixed, such as bicubic downsampling. However, as for remote sensing images (RSIs), some unexpected factors can cause the blurred visual performance, such as weather factors and orbit altitude. Blind SR methods are proposed to deal with various degradations. There are two main challenges of blind SR in RSIs: 1) the accurate estimation of degradation kernels and 2) the realistic image generation in the ill-posed problem. To rise to the challenge, we propose a novel blind SR framework based on dual conditional denoising diffusion probabilistic models (DDPMs). In our work, we introduce conditional DDPMs from two aspects: kernel estimation progress and reconstruction progress, named the dual-diffusion. As for kernel estimation progress, conditioned on low-resolution (LR) images, a new DDPM-based kernel predictor is constructed by studying the invertible mapping between the kernel distribution and the latent distribution. As for reconstruction progress, regarding the predicted degradation kernels and LR images as conditional information, we construct a DDPM-based reconstructor to learn the mapping from the LR images to high-resolution (HR) images. Comprehensive experiments show the priority of our proposal compared with state-of-the-art (SOTA) blind SR methods. Source code and supplementary materials are available at https://github.com/Lincoln20030413/DDSR.