Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-Resolution
Small ships in remote sensing images have blurred details and are difficult to detect. Existing algorithms usually detect small ships based on predefined anchors with different sizes. However, limited by the number of different sizes, it is difficult for anchor-based methods to match small ships of...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-07-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/15/3530 |
| _version_ | 1797440884668628992 |
|---|---|
| author | Xiaozhu Xie Linhao Li Zhe An Gang Lu Zhiqiang Zhou |
| author_facet | Xiaozhu Xie Linhao Li Zhe An Gang Lu Zhiqiang Zhou |
| author_sort | Xiaozhu Xie |
| collection | DOAJ |
| description | Small ships in remote sensing images have blurred details and are difficult to detect. Existing algorithms usually detect small ships based on predefined anchors with different sizes. However, limited by the number of different sizes, it is difficult for anchor-based methods to match small ships of different sizes and structures during training, as they can easily cause misdetections. In this paper, we propose a hybrid anchor structure to generate region proposals for small ships, so as to take full advantage of both anchor-based methods with high localization accuracy and anchor-free methods with fewer misdetections. To unify the output evaluation and obtain the best output, a label reassignment strategy is proposed, which reassigns the sample labels according to the harmonic intersection-over-union (IoU) before and after regression. In addition, an adaptive feature pyramid structure is proposed to enhance the features of important locations on the feature map, so that the features of small ship targets are more prominent and easier to identify. Moreover, feature super-resolution technology is introduced for the region of interest (RoI) features of small ships to generate super-resolution feature representations with a small computational cost, as well as generative adversarial training to improve the realism of super-resolution features. Based on the super-resolution feature, ship proposals are further classified and regressed by using super-resolution features to obtain more accurate detection results. Detailed ablation and comparison experiments demonstrate the effectiveness of the proposed method. |
| first_indexed | 2024-03-09T12:14:52Z |
| format | Article |
| id | doaj.art-7177d4e8a3c249fdbec698c0af10d1df |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T12:14:52Z |
| publishDate | 2022-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-7177d4e8a3c249fdbec698c0af10d1df2023-11-30T22:47:47ZengMDPI AGRemote Sensing2072-42922022-07-011415353010.3390/rs14153530Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-ResolutionXiaozhu Xie0Linhao Li1Zhe An2Gang Lu3Zhiqiang Zhou4Department of information and communication, Army Academy of Armored Forces, Beijing 100072, ChinaSchool of Automation, Beijing Institute of Technology, Beijing 100081, ChinaState Key Laboratory of Advanced Power Transmission Technology, Global Energy Interconnection Research Institute Co., Ltd., Beijing 102209, ChinaDepartment of information and communication, Army Academy of Armored Forces, Beijing 100072, ChinaSchool of Automation, Beijing Institute of Technology, Beijing 100081, ChinaSmall ships in remote sensing images have blurred details and are difficult to detect. Existing algorithms usually detect small ships based on predefined anchors with different sizes. However, limited by the number of different sizes, it is difficult for anchor-based methods to match small ships of different sizes and structures during training, as they can easily cause misdetections. In this paper, we propose a hybrid anchor structure to generate region proposals for small ships, so as to take full advantage of both anchor-based methods with high localization accuracy and anchor-free methods with fewer misdetections. To unify the output evaluation and obtain the best output, a label reassignment strategy is proposed, which reassigns the sample labels according to the harmonic intersection-over-union (IoU) before and after regression. In addition, an adaptive feature pyramid structure is proposed to enhance the features of important locations on the feature map, so that the features of small ship targets are more prominent and easier to identify. Moreover, feature super-resolution technology is introduced for the region of interest (RoI) features of small ships to generate super-resolution feature representations with a small computational cost, as well as generative adversarial training to improve the realism of super-resolution features. Based on the super-resolution feature, ship proposals are further classified and regressed by using super-resolution features to obtain more accurate detection results. Detailed ablation and comparison experiments demonstrate the effectiveness of the proposed method.https://www.mdpi.com/2072-4292/14/15/3530ship detectionhybrid anchor structurefeature super-resolutiongenerative adversarial training |
| spellingShingle | Xiaozhu Xie Linhao Li Zhe An Gang Lu Zhiqiang Zhou Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-Resolution Remote Sensing ship detection hybrid anchor structure feature super-resolution generative adversarial training |
| title | Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-Resolution |
| title_full | Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-Resolution |
| title_fullStr | Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-Resolution |
| title_full_unstemmed | Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-Resolution |
| title_short | Small Ship Detection Based on Hybrid Anchor Structure and Feature Super-Resolution |
| title_sort | small ship detection based on hybrid anchor structure and feature super resolution |
| topic | ship detection hybrid anchor structure feature super-resolution generative adversarial training |
| url | https://www.mdpi.com/2072-4292/14/15/3530 |
| work_keys_str_mv | AT xiaozhuxie smallshipdetectionbasedonhybridanchorstructureandfeaturesuperresolution AT linhaoli smallshipdetectionbasedonhybridanchorstructureandfeaturesuperresolution AT zhean smallshipdetectionbasedonhybridanchorstructureandfeaturesuperresolution AT ganglu smallshipdetectionbasedonhybridanchorstructureandfeaturesuperresolution AT zhiqiangzhou smallshipdetectionbasedonhybridanchorstructureandfeaturesuperresolution |