SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher Model
Change detection (CD) is an important yet challenging task in remote sensing. In this article, we underline that the combination of unsupervised and supervised methods in a semisupervised framework improves CD performance. We rely on half-sibling regression for optical change detection (SiROC) as an...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10106115/ |
| _version_ | 1797834840391811072 |
|---|---|
| author | Lukas Kondmann Sudipan Saha Xiao Xiang Zhu |
| author_facet | Lukas Kondmann Sudipan Saha Xiao Xiang Zhu |
| author_sort | Lukas Kondmann |
| collection | DOAJ |
| description | Change detection (CD) is an important yet challenging task in remote sensing. In this article, we underline that the combination of unsupervised and supervised methods in a semisupervised framework improves CD performance. We rely on half-sibling regression for optical change detection (SiROC) as an unsupervised teacher model to generate pseudolabels (PLs) and select only the most confident PLs for pretraining different student models. Our results are robust to three different competitive student models, two semisupervised PL baselines, two benchmark datasets, and a variety of loss functions. While the performance gains are highest with a limited number of labels, a notable effect of PL pretraining persists when more labeled data are used. Further, we outline that the confidence selection of SiROC is indeed effective and that the performance gains generalize to scenes that were not used for PL training. Through the PL pretraining, SemiSiROC allows student models to learn more refined shapes of changes and makes them less sensitive to differences in acquisition conditions. |
| first_indexed | 2024-04-09T14:43:52Z |
| format | Article |
| id | doaj.art-ce515fcf521549ea94c6a5f17e61c8aa |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-04-09T14:43:52Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-ce515fcf521549ea94c6a5f17e61c8aa2023-05-02T23:00:18ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352023-01-01163879389110.1109/JSTARS.2023.326810410106115SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher ModelLukas Kondmann0https://orcid.org/0000-0002-2253-6936Sudipan Saha1https://orcid.org/0000-0002-9440-0720Xiao Xiang Zhu2https://orcid.org/0000-0001-5530-3613Remote Sensing Technology Institute, German Aerospace Center (DLR), Wessling, GermanyYardi School of Artificial Intelligence, Indian Institute of Technology Delhi, New Delhi, IndiaData Science in Earth Observation, Technical University of Munich, Ottobrunn, GermanyChange detection (CD) is an important yet challenging task in remote sensing. In this article, we underline that the combination of unsupervised and supervised methods in a semisupervised framework improves CD performance. We rely on half-sibling regression for optical change detection (SiROC) as an unsupervised teacher model to generate pseudolabels (PLs) and select only the most confident PLs for pretraining different student models. Our results are robust to three different competitive student models, two semisupervised PL baselines, two benchmark datasets, and a variety of loss functions. While the performance gains are highest with a limited number of labels, a notable effect of PL pretraining persists when more labeled data are used. Further, we outline that the confidence selection of SiROC is indeed effective and that the performance gains generalize to scenes that were not used for PL training. Through the PL pretraining, SemiSiROC allows student models to learn more refined shapes of changes and makes them less sensitive to differences in acquisition conditions.https://ieeexplore.ieee.org/document/10106115/Change detection (CD)multitemporaloptical imagessemisupervisedunsupervised |
| spellingShingle | Lukas Kondmann Sudipan Saha Xiao Xiang Zhu SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher Model IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Change detection (CD) multitemporal optical images semisupervised unsupervised |
| title | SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher Model |
| title_full | SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher Model |
| title_fullStr | SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher Model |
| title_full_unstemmed | SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher Model |
| title_short | SemiSiROC: Semisupervised Change Detection With Optical Imagery and an Unsupervised Teacher Model |
| title_sort | semisiroc semisupervised change detection with optical imagery and an unsupervised teacher model |
| topic | Change detection (CD) multitemporal optical images semisupervised unsupervised |
| url | https://ieeexplore.ieee.org/document/10106115/ |
| work_keys_str_mv | AT lukaskondmann semisirocsemisupervisedchangedetectionwithopticalimageryandanunsupervisedteachermodel AT sudipansaha semisirocsemisupervisedchangedetectionwithopticalimageryandanunsupervisedteachermodel AT xiaoxiangzhu semisirocsemisupervisedchangedetectionwithopticalimageryandanunsupervisedteachermodel |