IRSR

0. Introduction

Codes for paper: Chen, L., Tang, R., Anisetti, M., & Yang, X. (2020). A Lightweight Iterative Error Reconstruction Network for Infrared Image Super-Resolution in Smart Grid. Sustainable Cities and Society, 102520.

@article{chen2020lightweight,
  title={A Lightweight Iterative Error Reconstruction Network for Infrared Image Super-Resolution in Smart Grid},
  author={Chen, Lihui and Tang, Rui and Anisetti, Marco and Yang, Xiaomin},
  journal={Sustainable Cities and Society},
  pages={102520},
  year={2020},
  publisher={Elsevier}
}

1. Requirements

python3
tqdm
opencv-python
pytorch(>=1.6)
torchvision
yaml

2. Test

Clone this repository:

git clone https://github.com/Huises/IERN-for-IR-Image-SR.git

Then, cd to IERN-for-IR-Image-SR and run the commands for evaluation on GIR50 and Infrared20 (or your own images) :

python test.py -opt options/test/test.yml #test GIR50 and Infrared20
python test.py -opt options/test/test.yml -lr_path your_img_path # test your own images

Finally, you can find the reconstruction images in ./results.

3. Train

Prepare train set and validation set use ./scripts/Prepare_TrainData_HR_LR.m or ./scripts/Prepare_TrainData_HR_LR.py

Clone this repository:

git clone https://github.com/Huises/IERN-for-IR-Image-SR.git

Open IERN-for-IR-Image-SR/options/train/train.yml. Then, modify image paths for train and validation set

Then，cd to IERN-for-IR-Image-SR and run the commands for evaluation

python train.py -opt options/train/train.yml  # train your own models

Acknowledgements