This repository includes codes and dataset for "UTRNet: An Unsupervised Time-Distance-Guided Convolutional Recurrent Network for Change Detection in Irregularly Collected Images", which has been published in IEEE Transactions on Geoscience and Remote Sensing.

The materials in this repository are only for study and research, NOT FOR COMMERCIAL USE.

If this code is helpful for you, please consider citing:
B. Yang, L. Qin, J. Liu and X. Liu, "UTRNet: An Unsupervised Time-Distance-Guided Convolutional Recurrent Network for Change Detection in Irregularly Collected Images," in IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1-16, 2022, Art no. 4410516, doi: 10.1109/TGRS.2022.3174009.

python train.py

python test.py

provides codes for generating lables

Remote sensing images and ground truth map of nine study scenes:
Access from Baidu Cloud
Link: https://pan.baidu.com/s/1guLIaIhC5sOr3rVPWMM7Hg
Password: t4ck
Access from Google Drive
Link: https://drive.google.com/file/d/1ssmjAtJLbrahAMK5ql2fCanuSJT1C4ze/view?usp=sharing