Unsupervised Change Detection Using Convolutional-Autoencoder Multiresolution Features

被引:24
|
作者
Bergamasco, Luca [1 ,2 ]
Saha, Sudipan [3 ]
Bovolo, Francesca [2 ]
Bruzzone, Lorenzo [1 ]
机构
[1] Univ Trento, Dept Informat Engn & Comp Sci, I-38123 Trento, Italy
[2] Fdn Bruno Kessler, Ctr Digital Soc, I-38123 Trento, Italy
[3] Tech Univ Munich, Data Sci Earth Observat, D-85521 Ottobrunn, Germany
来源
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING | 2022年 / 60卷
关键词
Feature extraction; Training; Data models; Decoding; Task analysis; Remote sensing; Semantics; Convolutional autoencoder (CAE); deep learning (DL); multitemporal analysis; remote sensing (RS); unsupervised change detection (CD); unsupervised learning; MULTIPLE-CHANGE DETECTION; CHANGE VECTOR ANALYSIS; IMAGES; ADAPTATION; FOREST;
D O I
10.1109/TGRS.2022.3140404
中图分类号
P3 [地球物理学]; P59 [地球化学];
学科分类号
0708 ; 070902 ;
摘要
The use of deep learning (DL) methods for change detection (CD) is currently dominated by supervised models that require a large number of labeled samples. However, these samples are difficult to acquire in the multitemporal case. A possible alternative is leveraging methods that exploit transfer learning for CD by reusing DL models pretrained for other tasks. However, the performance of the transfer-learning-based models decreases as much as the target images differ from the ones used for training the model. To overcome this limit, we propose an unsupervised CD method that exploits multiresolution deep feature maps derived by a convolutional autoencoder (CAE). It automatically learns spatial features from the input during the training phase without requiring any labeled data. The proposed method processes the bitemporal images to obtain and compare multiresolution bitemporal feature maps. These feature maps are then analyzed by a feature-selection technique to select the most discriminant ones. Furthermore, an aggregated multiresolution difference image is computed and used for a detail-preserving multiscale CD. In the context of this CD approach, we propose two alternative strategies to retrieve multiscale reliability maps. We tested the proposed method on bitemporal multispectral images acquired by Landsat-5 and Landsat-8 representing burned areas and Sentinel-2 images representing deforested areas. Results confirm the effectiveness of the proposed CD technique.
引用
收藏
页数:19
相关论文
共 50 条
  • [1] Unsupervised change-detection based on Convolutional-autoencoder Feature Extraction
    Bergamasco, Luca
    Saha, Sudipan
    Bovolo, Francesca
    Bruzzone, Lorenzo
    IMAGE AND SIGNAL PROCESSING FOR REMOTE SENSING XXV, 2019, 11155
  • [2] Unsupervised change detection using hierarchical convolutional autoencoder
    Bergamasco, Luca
    Bovolo, Francesca
    Bruzzone, Lorenzo
    IMAGE AND SIGNAL PROCESSING FOR REMOTE SENSING XXVI, 2020, 11533
  • [3] Detection of Freezing of Gait Using Unsupervised Convolutional Denoising Autoencoder
    Noor, Mohd Halim Mohd
    Nazir, Amril
    Ab Wahab, Mohd Nadhir
    Ling, Jodene Ooi Yen
    IEEE ACCESS, 2021, 9 : 115700 - 115709
  • [4] Unsupervised Machine Anomaly Detection Using Autoencoder and Temporal Convolutional Network
    Li, Zhiyuan
    Sun, Yu
    Yang, Laihao
    Zhao, Zhibin
    Chen, Xuefeng
    IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2022, 71
  • [5] Temporal convolutional autoencoder for unsupervised anomaly detection in time series
    Thill, Markus
    Konen, Wolfgang
    Wang, Hao
    Back, Thomas
    APPLIED SOFT COMPUTING, 2021, 112
  • [6] MULTIRESOLUTION SAR DATA FUSION FOR UNSUPERVISED CHANGE DETECTION
    Moser, Gabriele
    Serpico, Sebastiano B.
    Vernazza, Gianni
    2013 IEEE INTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM (IGARSS), 2013, : 1234 - 1237
  • [7] Unsupervised Anomaly Detection of Industrial Robots Using Sliding-Window Convolutional Variational Autoencoder
    Chen, Tingting
    Liu, Xueping
    Xia, Bizhong
    Wang, Wei
    Lai, Yongzhi
    IEEE ACCESS, 2020, 8 : 47072 - 47081
  • [8] Unsupervised Structural Damage Detection Technique Based on a Deep Convolutional Autoencoder
    Rastin, Zahra
    Ghodrati Amiri, Gholamreza
    Darvishan, Ehsan
    SHOCK AND VIBRATION, 2021, 2021
  • [9] Unsupervised seismic facies classification using deep convolutional autoencoder
    Puzyrev, Vladimir
    Elders, Chris
    GEOPHYSICS, 2022, 87 (04) : IM125 - IM132
  • [10] Unsupervised Subtyping of Cholangiocarcinoma Using a Deep Clustering Convolutional Autoencoder
    Muhammad, Hassan
    Sigel, Carlie S.
    Campanella, Gabriele
    Boerner, Thomas
    Pak, Linda M.
    Buttner, Stefan
    IJzermans, Jan N. M.
    Koerkamp, Bas Groot
    Doukas, Michael
    Jarnagin, William R.
    Simpson, Amber L.
    Fuchs, Thomas J.
    MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION - MICCAI 2019, PT I, 2019, 11764 : 604 - 612