High resolution representation‐based Siamese network for remote sensing image change detection
Abstract Change detection of high‐resolution remote sensing images can help to accurately understand the changes in the earth's surface. Advanced methods based on deep features have some limitations, including limited accuracy, poor detection effect, and poor robustness. The main reason is that...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2022-07-01
|
Series: | IET Image Processing |
Online Access: | https://doi.org/10.1049/ipr2.12505 |
_version_ | 1811337233018388480 |
---|---|
author | Zheng Liang Bin Zhu Yaoxuan Zhu |
author_facet | Zheng Liang Bin Zhu Yaoxuan Zhu |
author_sort | Zheng Liang |
collection | DOAJ |
description | Abstract Change detection of high‐resolution remote sensing images can help to accurately understand the changes in the earth's surface. Advanced methods based on deep features have some limitations, including limited accuracy, poor detection effect, and poor robustness. The main reason is that these frameworks have poor feature extraction capabilities, insufficient context aggregation, and inadequate discrimination capabilities. In order to solve these problems, SiHDNet, a Siamese segmentation network based on deep, high‐resolution differential feature interaction, is proposed. Specifically, after the high‐resolution features of the dual‐temporal image are extracted, the difference map is generated through a special fusion module, which contains sufficient and effective change information. Finally, the final binary change map is obtained through the improved spatial pyramid pooling module. Experiments are conducted on the newly released building change detection data set LEVIR‐CD and the challenging remote sensing image change detection data set Google Data Set. Five benchmark methods are chosen. The results of quantitative analysis and qualitative comparison show that SiHDNet is superior to the five benchmark methods. The results of the ablation experiment also verify the effectiveness of this method. |
first_indexed | 2024-04-13T17:51:56Z |
format | Article |
id | doaj.art-e1dcf38bccc544d28357a96ff441a8f8 |
institution | Directory Open Access Journal |
issn | 1751-9659 1751-9667 |
language | English |
last_indexed | 2024-04-13T17:51:56Z |
publishDate | 2022-07-01 |
publisher | Wiley |
record_format | Article |
series | IET Image Processing |
spelling | doaj.art-e1dcf38bccc544d28357a96ff441a8f82022-12-22T02:36:40ZengWileyIET Image Processing1751-96591751-96672022-07-011692506251710.1049/ipr2.12505High resolution representation‐based Siamese network for remote sensing image change detectionZheng Liang0Bin Zhu1Yaoxuan Zhu2Photoelectric Reconnaissance Information Processing Laboratory National University of Defense Technology Hefei Anhui ChinaPhotoelectric Reconnaissance Information Processing Laboratory National University of Defense Technology Hefei Anhui ChinaPhotoelectric Reconnaissance Information Processing Laboratory National University of Defense Technology Hefei Anhui ChinaAbstract Change detection of high‐resolution remote sensing images can help to accurately understand the changes in the earth's surface. Advanced methods based on deep features have some limitations, including limited accuracy, poor detection effect, and poor robustness. The main reason is that these frameworks have poor feature extraction capabilities, insufficient context aggregation, and inadequate discrimination capabilities. In order to solve these problems, SiHDNet, a Siamese segmentation network based on deep, high‐resolution differential feature interaction, is proposed. Specifically, after the high‐resolution features of the dual‐temporal image are extracted, the difference map is generated through a special fusion module, which contains sufficient and effective change information. Finally, the final binary change map is obtained through the improved spatial pyramid pooling module. Experiments are conducted on the newly released building change detection data set LEVIR‐CD and the challenging remote sensing image change detection data set Google Data Set. Five benchmark methods are chosen. The results of quantitative analysis and qualitative comparison show that SiHDNet is superior to the five benchmark methods. The results of the ablation experiment also verify the effectiveness of this method.https://doi.org/10.1049/ipr2.12505 |
spellingShingle | Zheng Liang Bin Zhu Yaoxuan Zhu High resolution representation‐based Siamese network for remote sensing image change detection IET Image Processing |
title | High resolution representation‐based Siamese network for remote sensing image change detection |
title_full | High resolution representation‐based Siamese network for remote sensing image change detection |
title_fullStr | High resolution representation‐based Siamese network for remote sensing image change detection |
title_full_unstemmed | High resolution representation‐based Siamese network for remote sensing image change detection |
title_short | High resolution representation‐based Siamese network for remote sensing image change detection |
title_sort | high resolution representation based siamese network for remote sensing image change detection |
url | https://doi.org/10.1049/ipr2.12505 |
work_keys_str_mv | AT zhengliang highresolutionrepresentationbasedsiamesenetworkforremotesensingimagechangedetection AT binzhu highresolutionrepresentationbasedsiamesenetworkforremotesensingimagechangedetection AT yaoxuanzhu highresolutionrepresentationbasedsiamesenetworkforremotesensingimagechangedetection |