H2Det: A High-Speed and High-Accurate Ship Detector in SAR Images
Synthetic aperture radar (SAR) sensor is a vital platform for ship detection whose accuracy and speed are usually difficult to balance. An urgent problem to be solved is how to achieve high-speed detection while maintaining high-accurate. To address this problem, we propose a high-speed and high-acc...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9628058/ |
| _version_ | 1819239093792407552 |
|---|---|
| author | Mingming Zhu Guoping Hu Hao Zhou Shiqiang Wang |
| author_facet | Mingming Zhu Guoping Hu Hao Zhou Shiqiang Wang |
| author_sort | Mingming Zhu |
| collection | DOAJ |
| description | Synthetic aperture radar (SAR) sensor is a vital platform for ship detection whose accuracy and speed are usually difficult to balance. An urgent problem to be solved is how to achieve high-speed detection while maintaining high-accurate. To address this problem, we propose a high-speed and high-accurate detector (H2Det) in SAR images. For one thing, we adopt fewer convolutional layers, CSP module, and rectangle filling to ensure model high-speed. For another, we propose spatial pyramid pooling, bottom-up path augmentation, and mosaic data augmentation to ensure model high-accurate. To establish an optimal H2Det, we conduct comparative studies on SAR ship detection dataset (SSDD). Moreover, we verify the effectiveness of these modules mentioned above through ablation studies. The experimental results on SSDD demonstrate that both accuracy and speed of the proposed method outperform other state-of-the-art methods and references. In addition, the strong migration ability of the proposed H2Det is shown on high-resolution SAR images dataset. |
| first_indexed | 2024-12-23T13:46:39Z |
| format | Article |
| id | doaj.art-4924cd9979264ec7bc12200343627b08 |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-12-23T13:46:39Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-4924cd9979264ec7bc12200343627b082022-12-21T17:44:42ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352021-01-0114124551246610.1109/JSTARS.2021.31311629628058H2Det: A High-Speed and High-Accurate Ship Detector in SAR ImagesMingming Zhu0https://orcid.org/0000-0002-2526-3531Guoping Hu1https://orcid.org/0000-0001-9369-2410Hao Zhou2https://orcid.org/0000-0002-1012-3042Shiqiang Wang3https://orcid.org/0000-0002-5783-7173Air Force Engineering University, Xi'an, ChinaAir Force Engineering University, Xi'an, ChinaAir Force Engineering University, Xi'an, ChinaAir Force Engineering University, Xi'an, ChinaSynthetic aperture radar (SAR) sensor is a vital platform for ship detection whose accuracy and speed are usually difficult to balance. An urgent problem to be solved is how to achieve high-speed detection while maintaining high-accurate. To address this problem, we propose a high-speed and high-accurate detector (H2Det) in SAR images. For one thing, we adopt fewer convolutional layers, CSP module, and rectangle filling to ensure model high-speed. For another, we propose spatial pyramid pooling, bottom-up path augmentation, and mosaic data augmentation to ensure model high-accurate. To establish an optimal H2Det, we conduct comparative studies on SAR ship detection dataset (SSDD). Moreover, we verify the effectiveness of these modules mentioned above through ablation studies. The experimental results on SSDD demonstrate that both accuracy and speed of the proposed method outperform other state-of-the-art methods and references. In addition, the strong migration ability of the proposed H2Det is shown on high-resolution SAR images dataset.https://ieeexplore.ieee.org/document/9628058/Convolutional neural network (CNN)high-resolution SAR images dataset (HRSID)high-speedSAR ship detection dataset (SSDD)ship detectionsynthetic aperture radar (SAR) |
| spellingShingle | Mingming Zhu Guoping Hu Hao Zhou Shiqiang Wang H2Det: A High-Speed and High-Accurate Ship Detector in SAR Images IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Convolutional neural network (CNN) high-resolution SAR images dataset (HRSID) high-speed SAR ship detection dataset (SSDD) ship detection synthetic aperture radar (SAR) |
| title | H2Det: A High-Speed and High-Accurate Ship Detector in SAR Images |
| title_full | H2Det: A High-Speed and High-Accurate Ship Detector in SAR Images |
| title_fullStr | H2Det: A High-Speed and High-Accurate Ship Detector in SAR Images |
| title_full_unstemmed | H2Det: A High-Speed and High-Accurate Ship Detector in SAR Images |
| title_short | H2Det: A High-Speed and High-Accurate Ship Detector in SAR Images |
| title_sort | h2det a high speed and high accurate ship detector in sar images |
| topic | Convolutional neural network (CNN) high-resolution SAR images dataset (HRSID) high-speed SAR ship detection dataset (SSDD) ship detection synthetic aperture radar (SAR) |
| url | https://ieeexplore.ieee.org/document/9628058/ |
| work_keys_str_mv | AT mingmingzhu h2detahighspeedandhighaccurateshipdetectorinsarimages AT guopinghu h2detahighspeedandhighaccurateshipdetectorinsarimages AT haozhou h2detahighspeedandhighaccurateshipdetectorinsarimages AT shiqiangwang h2detahighspeedandhighaccurateshipdetectorinsarimages |