Continuous Remote Sensing Image Super-Resolution via Neural Operator Diffusion

March 4, 2025 ยท View on GitHub

This repository is the official implementation of the paper "NeurOp-Diff: Continuous Remote Sensing Image Super-Resolution via Neural Operator Diffusion".

Environment configuration

The codes are based on python 3.11, pytorch 2.4.0 and CUDA version 12.4.

Data preparation

checkpoint

The pre-trained weights for continuous SR can be found at this link

Train

python main.py --config train.yml --exp ./result --doc pth --timesteps [steps] --ni

Test

python test.py --config ./configs/test.yml --model [checkpoint_path] --timesteps [steps]

Demo

python demo.py --config ./configs/test.yml --model [checkpoint_path] --path [image_path]

Acknowledgements

This code is mainly built based on DDIM, SRNO and LIIF