计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (2): 183-190.doi: 10.11896/jsjkx.240400131

• 计算机图形学&多媒体 • 上一篇    下一篇

基于上下文细粒度信息修复的遥感变化检测

杜乾刚1, 彭博2, 池明旻1   

  1. 1 复旦大学计算机科学技术学院 上海 200438
    2 上海海洋大学信息学院 上海 201306
  • 收稿日期:2024-04-17 修回日期:2024-07-19 出版日期:2025-02-15 发布日期:2025-02-17
  • 通讯作者: 池明旻(mmchi@fudan.edu.cn)
  • 作者简介:(qgdu21@m.fudan.edu.cn)
  • 基金资助:
    国家自然科学基金(62171139)

Remote Sensing Change Detection Based on Contextual Fine-grained Information Restoration

DU Qiangang1, PENG Bo2, CHI Mingmin1   

  1. 1 School of Computer Science,Fudan University,Shanghai 200438,China
    2 College of Information Technology,Shanghai Ocean University,Shanghai 201306,China
  • Received:2024-04-17 Revised:2024-07-19 Online:2025-02-15 Published:2025-02-17
  • About author:DU Qiangang,born in 1996,postgra-duate.His main research interests include computer version and vision-language.
    CHI Mingmin,born in 1976,Ph.D,professor,Ph.D supervisor.Her main research interests include computer vison and machine learning.
  • Supported by:
    National Natural Science Foundation of China(62171139).

PDF (PC)

122

摘要： 遥感变化检测在军事和民用领域都发挥着至关重要的作用。然而,由于变化检测图像对的数据采集存在巨大的时空差距,因此存在大量伪变化。现有的变化检测方法基于双流孪生网络学习物体特征,然后通过一系列专有网络进行伪变化消除。然而,这种相互独立的去噪方式缺乏捕捉图像对之间相互依存关系的能力,而且往往由于过度关注去噪设计而导致大量的细粒度信息丢失。所提CFIR 利用图像对的数据结构特征来增强模型学习上下文依赖关系的能力,并弥补丢失的细粒度信息,缓解了细粒度信息丢失的问题。此外,CFIR 采用一种门控机制,消除变化检测任务中的伪变化,并引导网络提取相关的变化特征,缓解了变化检测极端的数据不平衡对模型学习真实变化的影响。CFIR 在多个变化检测基准中表现出了极具竞争力的性能,其中相较于变化检测最先进算法,在LEVIR-CD数据集上F1提高0.21%,IoU提高0.38%;在WHU-CD数据集上F1提高0.99%,IoU提高2.43%。

关键词: 遥感变化检测, 细粒度信息重构, 门控机制, 噪声消除, 有监督学习

Abstract: Remote sensing change detection plays a crucial role in both military and civilian fields.However,there is a large amount of pseudo-change noise due to the huge spatial and temporal gaps in data acquisition of change detection image pairs.Exis-ting change detection methods are based on learning object features from a dual-stream twin network,followed by pseudo-change noise removal via a series of proprietary networks.However,this mutually independent denoising approach lacks the ability to capture the interdependencies between image pairs,and often results in the loss of a large amount of fine-grained information due to excessive focus on the denoising design.The CFIR proposed in this paper mitigates the problem of fine-grained information loss by exploiting the data structure features of the image pairs to augment the model's ability to learn the contextual dependencies and to compensate for the lost fine-grained information.In addition,it employs a gating mechanism that eliminates pseudo-change noise in the change detection task and guides the network to extract relevant change features,mitigating the impact of extreme data imbalance in change detection on the model's ability to learn real changes.CFIR has demonstrated competitive performance in several change detection benchmarks.Compared with the state-of-the-art method,it improves F1 by 0.21% and IoU by 0.38% on the LEVIR-CD dataset,and improves F1 by 0.99% and IoU by 2.43% on the WHU-CD dataset.

Key words: Remote sensing change detection, Fine-grained information reconstruction, Gated mechanisms, Denoise, Supervised learning

中图分类号: 

  • TP181
[1]CHEN H,SHI Z.A Spatial-Temporal Attention-Based Method and a New Dataset for Remote Sensing Image Change Detection[J].Remote Sensing,2020,12(10):1662.
[2]JI S,WEI S,LU M.Fully Convolutional Networks for Multi-source Building Extraction From an Open Aerial and Satellite Imagery Data Set[J].IEEE Transactions on Geoscience and Remote Sensing,2019,57(1):574-586.
[3]LEBEDEV M A,VIZILTER Y V,VYGOLOV O V.et al.Change Detection in Remote Sensing Images Using Conditional Adversarial Networks[J].The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences,2018,XLII-2:565-571.
[4]ZHANG C,YUE P,TAPETE D,et al.A deeply supervisedimage fusion network for change detection in high resolution bi-temporal remote sensing images[J].ISPRS Journal of Photogrammetry and Remote Sensing,2020,166:183-200.
[5]DAUDT R C,LE SAUX B,BOULCH A.Fully ConvolutionalSiamese Networks for Change Detection[J].arXiv:1810.08462v1,2018.
[6]CHEN C P,HSIEH J W,CHENP Y,et al.SARAS-Net:Scale and Relation Aware Siamese Network for Change Detection[J].Proceedings of the AAAI Conference on Artificial Intelligence,2023,37(12):14187-14195.
[7]CHEN H,QI Z,SHI Z.Remote Sensing Image Change Detection With Transformers[J].IEEE Transactions on Geoscience and Remote Sensing,2022,60:1-14.
[8]DING X,GUO Y,DING G,et al.ACNet:Strengthening theKernel Skeletons for Powerful CNN via Asymmetric Convolution Blocks[J].arXiv:1908.03930,2019.
[9]FANG S,LI K,SHAO J,et al.SNUNet-CD:A Densely Connec-ted Siamese Network for Change Detection of VHR Images[J].IEEE Geoscience and Remote Sensing Letters,2022,19:1-5.
[10]CODEGONI A,LOMBARDI G,FERRARI A,TINYCD:A(Not So) Deep Learning Model For Change Detection[J].arXiv:2207.13159,2022.
[11]MA X W,YANG J W,HONG T F,et al.STNet:Spatial andTemporal feature fusion network for change detection in remote sensing images[J].arXiv:2304.11422,2023.
[12]FANG Z L,LI K Y.Changer Feature Interaction is What You Need for Change Detection[J].arXiv.2209.08290,2022.
[13]BANDARA W G C,PATEL V M.A Transformer-Based Sia-mese Network for Change Detection[C]//2022 IEEE International Geoscience and Remote Sensing Symposium(IGARSS 2022).2022:207-210.
[14]LEI T,WANG J,NING H,et al.Difference Enhancement andSpatial-Spectral Nonlocal Network for Change Detection in VHR Remote Sensing Images[J].IEEE Transactions on Geoscience and Remote Sensing,2022,60:1-13.
[15]LI Z,TANG C,WANGL,et al.Remote Sensing Change Detection via Temporal Feature Interaction and Guided Refinement[J].IEEE Transactions on Geoscience and Remote Sensing,2022,60:1-11.
[16]LI Z,YAN C,SUN,Y,et al.A Densely Attentive Refinement Network for Change Detection Based on Very-High-Resolution Bitemporal Remote Sensing Images[J].IEEE Transactions on Geoscience and Remote Sensing,2022,60:1-18.
[17]LIN T Y,DOLLAR P,GIRSHICK R,et al.Feature Pyramid Networks for Object Detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).2017:936-944.
[18]LV Z,WANG F,CUI G,et al.Spatial-Spectral Attention Network Guided With Change Magnitude Image for Land Cover Change Detection Using Remote Sensing Images[J].IEEE Transactionson Geoscience and Remote Sensing,2022,60:1-12.
[19]LYU H,LU H.Learning a transferable change detection methodby Recurrent Neural Network[C]//2016 IEEE International Geoscience and Remote Sensing Symposium(IGARSS).2016:5157-5160.
[20]HE K,ZHANG X,REN S,et al.Deep Residual Learning forImage Recognition[J].arXiv:1512.03385,2015.
[21]RUßWURM M,KORNER M.Temporal Vegetation ModellingUsing Long Short-Term Memory Networks for Crop Identification from Medium-Resolution Multi-spectral Satellite Images[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops(CVPRW).2017:1496-1504.
[22]VARGHESE A,GUBBI J,RAMASWAMY A,et al.Change-Net:A Deep Learning Architecture for Visual Change Detection[C]//Computer Vision－ECCV 2018.2019:129-145.
[23]BAI B,FU W,LU T,et al.Edge-Guided Recurrent ConvolutionalNeural Network for Multitemporal Remote Sensing Image Building Change Detection[J].IEEE Transactions on Geos-cience and Remote Sensing,2022.60:1-13.
[24]YANG B,QIN L,LIU J,et al.IRCNN:An Irregular-Time-Distanced Recurrent Convolutional Neural Network for Change Detection in Satellite Time Series[J].IEEE Geoscience and Remote Sensing Letters,2022,19:1-5.
[25]VASWANI A,SHAZEER N,PARMAR N,et al.Attention isall you need[J].arXiv:1706.03762,2017.
[26]WANG S P,LI Y X,XIE M,et al.Align,Perturb and Decouple:Toward Better Leverage of Difference Information for RSI Change Detection[J].arXiv:2305.18714,2023.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发