This repository provides the code for the methods and experiments presented in our paper 'Edge-guided Recurrent Convolutional Neural Network for Multi-temporal Remote Sensing Image Building Change Detection'. You can find the PDF of this paper on: https://ieeexplore.ieee.org/document/9524849

If you have any questions, you can send me an email. My mail address is baibeifang@gmail.com.

Download the building change detection dataset.

• LEVIR-CD: https://justchenhao.github.io/LEVIR/
• WHU: https://study.rsgis.whu.edu.cn/pages/download/

In the following experiments, each image in the dataset is pre-cropped into multiple image patches of size 256 × 256.

path to dataset:
                ├─train
                  ├─A
                  ├─B
                  ├─label
                  ├─label_edge
                ├─val
                  ├─A
                  ├─B
                  ├─label
                  ├─label_edge
                ├─test
                  ├─A
                  ├─B
                  ├─label
                  ├─label_edge

generate edges.py

train.py

You can use your own trained model or download our pre-trained model

test.py

If you find this paper useful, please cite:

Beifang Bai, Wei Fu, Ting Lu, and Shutao Li, "Edge-Guided Recurrent Convolutional Neural Network for Multitemporal Remote Sensing Image Building Change Detection," in IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1-13, 2022, Art no. 5610613, doi: 10.1109/TGRS.2021.3106697.