Learning Capsules for SAR Target Recognition
Deep learning has been successfully utilized in synthetic aperture radar (SAR) automatic target recognition tasks and obtained state-of-the-art results. However, current deep learning algorithms do not perform well when SAR images are occluded, noisy, or with a great depression angle variance. This...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9164999/ |
| _version_ | 1831752360895971328 |
|---|---|
| author | Yunrui Guo Zongxu Pan Meiming Wang Ji Wang Wenjing Yang |
| author_facet | Yunrui Guo Zongxu Pan Meiming Wang Ji Wang Wenjing Yang |
| author_sort | Yunrui Guo |
| collection | DOAJ |
| description | Deep learning has been successfully utilized in synthetic aperture radar (SAR) automatic target recognition tasks and obtained state-of-the-art results. However, current deep learning algorithms do not perform well when SAR images are occluded, noisy, or with a great depression angle variance. This article proposes a novel method, SAR capsule network, to achieve the accurate and robust classification of SAR images without significantly increasing network complexity. Specifically, we develop a convolutional neural network extension based on Hinton's capsule network to capture spatial relationships specialized in classification between different entities in a SAR image. The SAR capsules are learned by a vector-based full connection operation instead of the traditional routing process, which not only alleviates the computational burden but also improves recognition accuracy. For occlusion, additive noise, and multiplicative noise tests, SAR capsule network shows superior robustness compared with typical convolution neural networks. When missing training data in a certain aspect angle range or existing a large depression angle variance between training data and test data, the proposed network achieves better performance than the existing works and reveals some competitive advantages in several test scenarios. |
| first_indexed | 2024-12-21T23:00:39Z |
| format | Article |
| id | doaj.art-989da0bb166246f3906c8a002eb903db |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-12-21T23:00:39Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-989da0bb166246f3906c8a002eb903db2022-12-21T18:47:18ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352020-01-01134663467310.1109/JSTARS.2020.30159099164999Learning Capsules for SAR Target RecognitionYunrui Guo0https://orcid.org/0000-0002-1998-528XZongxu Pan1https://orcid.org/0000-0002-5041-3300Meiming Wang2Ji Wang3https://orcid.org/0000-0003-0637-8744Wenjing Yang4https://orcid.org/0000-0002-6997-0406Institute for Quantum Information and State Key Laboratory of High Performance Computing, College of Computer, National University of Defense Technology, Changsha, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaInstitute for Quantum Information and State Key Laboratory of High Performance Computing, College of Computer, National University of Defense Technology, Changsha, ChinaInstitute for Quantum Information and State Key Laboratory of High Performance Computing, College of Computer, National University of Defense Technology, Changsha, ChinaInstitute for Quantum Information and State Key Laboratory of High Performance Computing, College of Computer, National University of Defense Technology, Changsha, ChinaDeep learning has been successfully utilized in synthetic aperture radar (SAR) automatic target recognition tasks and obtained state-of-the-art results. However, current deep learning algorithms do not perform well when SAR images are occluded, noisy, or with a great depression angle variance. This article proposes a novel method, SAR capsule network, to achieve the accurate and robust classification of SAR images without significantly increasing network complexity. Specifically, we develop a convolutional neural network extension based on Hinton's capsule network to capture spatial relationships specialized in classification between different entities in a SAR image. The SAR capsules are learned by a vector-based full connection operation instead of the traditional routing process, which not only alleviates the computational burden but also improves recognition accuracy. For occlusion, additive noise, and multiplicative noise tests, SAR capsule network shows superior robustness compared with typical convolution neural networks. When missing training data in a certain aspect angle range or existing a large depression angle variance between training data and test data, the proposed network achieves better performance than the existing works and reveals some competitive advantages in several test scenarios.https://ieeexplore.ieee.org/document/9164999/Capsule networkconvolutional neural network (CNN)deep learningsynthetic aperture radar (SAR) target recognition |
| spellingShingle | Yunrui Guo Zongxu Pan Meiming Wang Ji Wang Wenjing Yang Learning Capsules for SAR Target Recognition IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Capsule network convolutional neural network (CNN) deep learning synthetic aperture radar (SAR) target recognition |
| title | Learning Capsules for SAR Target Recognition |
| title_full | Learning Capsules for SAR Target Recognition |
| title_fullStr | Learning Capsules for SAR Target Recognition |
| title_full_unstemmed | Learning Capsules for SAR Target Recognition |
| title_short | Learning Capsules for SAR Target Recognition |
| title_sort | learning capsules for sar target recognition |
| topic | Capsule network convolutional neural network (CNN) deep learning synthetic aperture radar (SAR) target recognition |
| url | https://ieeexplore.ieee.org/document/9164999/ |
| work_keys_str_mv | AT yunruiguo learningcapsulesforsartargetrecognition AT zongxupan learningcapsulesforsartargetrecognition AT meimingwang learningcapsulesforsartargetrecognition AT jiwang learningcapsulesforsartargetrecognition AT wenjingyang learningcapsulesforsartargetrecognition |