Dual-Branch Super-Resolution (DBSR): Lightweight DBSR Network for Enhancing Remote Sensing Images.

The article focuses on the development and evaluation of a lightweight dual-branch super-resolution (DBSR) model designed to enhance the spatial resolution of remote sensing images (RSIs). DBSR integrates a main branch—comprising channel recalibration (CRC), a separable swin transformer (SST), and a spatial refinement module (SRM)—with an auxiliary branch featuring a global information enhancement module (GIEM) to improve both local and global feature extraction while maintaining computational efficiency. Experimental results on multiple RSI and natural image benchmark datasets demonstrate that DBSR achieves superior reconstruction quality and efficiency compared to existing lightweight SR methods, with only 245K parameters. Ablation studies confirm the effectiveness of each module, and practical applications, such as landslide detection, indicate DBSR's potential for real-time remote sensing tasks.