This repository is connected to my master's thesis, Deep learning-based cross-sensor super resolution of satellite images, where I investigate how one of the highest performing GAN-based single-image super-resolution models, ESRGAN, can be adapted to a mutispectral satellite task.
ESRGAN core component code based on https://github.com/peteryuX/esrgan-tf2