Change is Everywhere
Single-Temporal Supervised Object Change Detection
in Remote Sensing Imagery

by Zhuo Zheng, Ailong Ma, Liangpei Zhang and Yanfei Zhong

[Paper] [Project] [BibTeX]

This is an official implementation of STAR and ChangeStar in our ICCV 2021 paper Change is Everywhere: Single-Temporal Supervised Object Change Detection for High Spatial Resolution Remote Sensing Imagery.

We hope that STAR will serve as a solid baseline and help ease future research in weakly-supervised object change detection.

---

News

• 2021/09/24, ChangeStar has been included in microsoft/torchgeo!
• 2021/08/28, The code is available.
• 2021/07/23, The code will be released soon.
• 2021/07/23, This paper is accepted by ICCV 2021.

Features

• Learning a good change detector from single-temporal supervision.
• Strong baselines for bitemporal and single-temporal supervised change detection.
• A clean codebase for weakly-supervised change detection.
• Support both bitemporal and single-temporal supervised settings

Getting Started

Install EVer

pip install ever-beta==0.2.3

Requirements:

• pytorch >= 1.6.0
• python >=3.6

Prepare Dataset

1. Download xView2 dataset (training set and tier3 set) and LEVIR-CD dataset.

2. Create soft link

ln -s </path/to/xView2> ./xView2
ln -s </path/to/LEVIR-CD> ./LEVIR-CD

Training and Evaluation under Single-Temporal Supervision

bash ./scripts/trainxView2/r50_farseg_changemixin_symmetry.sh

Training and Evaluation under Bitemporal Supervision

bash ./scripts/bisup_levircd/r50_farseg_changemixin.sh

Citation

If you use STAR or ChangeStar (FarSeg) in your research, please cite the following paper:

@inproceedings{zheng2021change,
  title={Change is Everywhere: Single-Temporal Supervised Object Change Detection in Remote Sensing Imagery},
  author={Zheng, Zhuo and Ma, Ailong and Zhang, Liangpei and Zhong, Yanfei},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={15193--15202},
  year={2021}
}

@inproceedings{zheng2020foreground,
  title={Foreground-Aware Relation Network for Geospatial Object Segmentation in High Spatial Resolution Remote Sensing Imagery},
  author={Zheng, Zhuo and Zhong, Yanfei and Wang, Junjue and Ma, Ailong},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  pages={4096--4105},
  year={2020}
}

License

ChangeStar is released under the Apache License 2.0.

Copyright (c) Zhuo Zheng. All rights reserved.