Skip to content

Latest commit

 

History

History
16 lines (9 loc) · 1.11 KB

README.md

File metadata and controls

16 lines (9 loc) · 1.11 KB

SrSeg

Code for the paper of "Super Resolution Guided Deep Network for Land Cover Classification from Remote Sensing Images"

Note:

  1. The code is based on HRNet (https://github.com/HRNet/HRNet-Semantic-Segmentation/tree/pytorch-v1.1). Compared with original version, many documents in the folder of lib are revised, including all documents in the folder of core, all documents in the folder of datasets, all documents in the folder of models and the utils.py in the folder of utils.Furthermore, the train.py and the seg_*.yaml are also reviesed. The pretrained model is downloaded from https://github.com/HRNet/HRNet-Image-Classification and the HRNet-W48-C is used in this project.
  2. The code is a example for Potsdam.
  3. The code to estimate down-sampling kernel is from KernelGAN.
  4. The code for image super-resolution is from EDSR-PyTorch.
  5. The dataset can be downloaded from https://pan.baidu.com/s/1eboZS8bpqh6ET7Km9vTArg (password:rftx).

I am sorry for my mistake!!! Please replace seg_hrnet.py with seg_hrnet _ssr.py. Thanks for the contribution from Zicheng Zhao in Nanjing University of Science and Technology.