Spatial Super Resolution Technology of FY-3D MWRI With Deep Internal Learning

The data of Microwave Radiation Imager(MWRI) on FY3D satellite suffers from numerous impacts during data acquisition, resulting in a loss in spatial resolution. However, existing methods such as wiener filtering and Lucy Richardson are unable to enhance spatial resolution under complicated deterioration conditions. In this paper, an unsupervised method using the “zero-shot” super-resolution(ZSSR) framework is applied to microwave image processing. Experimental results on simulated images and real MWRI images have demonstrated the validity and effectiveness of the method.