Home

Awesome

Adjacent-Level Feature Cross-Fusion with 3-D CNN for Remote Sensing Image Change Detection

Here, we provide the official pytorch implementation of the paper "Adjacent-Level Feature Cross-Fusion with 3-D CNN for Remote Sensing Image Change Detection". Architecture

Requirements

Dataset Preparation

Data Structure

"""
Change detection data set with pixel-level binary labels;
├─A
├─B
├─label
└─list
  ├─train.txt
  ├─val.txt
  ├─test.txt
"""
A: Images of T1 time
B: Images of T2 time
label: label maps
list: contrains train.txt, val.txt, and test.txt. each fild records the name of image paris (XXX.png).

Data Download

WHU-CD: https://study.rsgis.whu.edu.cn/pages/download/building_dataset.html
LEVIR-CD: https://justchenhao.github.io/LEVIR/
SYSU-CD: https://github.com/liumency/SYSU-CD

Training and Testing

train.py
Test.py

Quantitative Results

image

Qualitative Results

SYSU-result

Licence

The code is released for non-commercial and research purposes only. For commercial purposes, please contact the authors.

Citation

If you find this work interesting in your research, please cite our paper as follow:
@ARTICLE{YeCD,
author={Ye, Yuanxin and Wang, Mengmeng and Zhou, Liang and Lei, Guangyang and Fan, Jianwei and Qin, Yao},
journal={IEEE Transactions on Geoscience and Remote Sensing},
title={Adjacent-Level Feature Cross-Fusion With 3-D CNN for Remote Sensing Image Change Detection},
year={2023},
volume={61},
number={},
pages={1-14},
doi={10.1109/TGRS.2023.3305499}}