A Large Ship Detection Method Based on Component Model in SAR Images
Large ship targets in synthetic aperture radar (SAR) images have characteristics, such as large image proportions, rich features, and large feature differences in a single target. Existing multiscale ship detection algorithms for SAR data employ multiscale feature pyramids or anchor-free extractors...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10787532/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832582389991211008 |
|---|---|
| author | Tiancheng Dong Taoyang Wang Xuefei Li Jianzhi Hong Maoqiang Jing Tong Wei |
| author_facet | Tiancheng Dong Taoyang Wang Xuefei Li Jianzhi Hong Maoqiang Jing Tong Wei |
| author_sort | Tiancheng Dong |
| collection | DOAJ |
| description | Large ship targets in synthetic aperture radar (SAR) images have characteristics, such as large image proportions, rich features, and large feature differences in a single target. Existing multiscale ship detection algorithms for SAR data employ multiscale feature pyramids or anchor-free extractors to capture features from large ship targets. However, the significant feature variation due to the internal structure and reflection of ships makes it difficult for current extractors to provide a consistent feature description, often resulting in fragmented detection outcomes for ship targets. This article decomposes the large ship detection problem into the detection of individual ship components (including tail, hull, and head), proposing a novel component-based detection method for large ship targets in SAR images. The proposed method enhances the network's efficiency in feature propagation and aggregation across different layers using the generalized efficient layer aggregation network (GELAN) structure. Following the feature extraction of GELAN, a multilevel multipooling channel attention is integrated to optimize the feature extraction structure in a hierarchical manner. The method also incorporates environmental features around the target to strengthen the association between different ship components. The detected ship components are connected using a topological relationship algorithm based on the component structure, culminating in the generation of ship target detection results. Experiments on the large ship component model dataset constructed for this article demonstrate significant improvements in the proposed algorithm over the preoptimized YOLOv8. The experimental results demonstrate that our method achieves promising detection performance when compared with the current state-of-the-art you only look once series algorithms and multiscale SAR ship detection algorithms. The algorithm also effectively avoided noticeable loss or false detection of small ship targets present in the dataset. |
| format | Article |
| id | doaj-art-932dd1ca3df4442da1b2c94b93fcdaaf |
| institution | Kabale University |
| issn | 1939-1404 2151-1535 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj-art-932dd1ca3df4442da1b2c94b93fcdaaf2025-01-30T00:00:12ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01184108412310.1109/JSTARS.2024.351489810787532A Large Ship Detection Method Based on Component Model in SAR ImagesTiancheng Dong0https://orcid.org/0009-0008-4277-1192Taoyang Wang1https://orcid.org/0000-0002-6014-5354Xuefei Li2https://orcid.org/0000-0001-8572-3272Jianzhi Hong3https://orcid.org/0000-0003-4059-7131Maoqiang Jing4https://orcid.org/0000-0002-7378-9394Tong Wei5State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, ChinaSchool of Computer Science, Wuhan University, Wuhan, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, ChinaState Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, ChinaLarge ship targets in synthetic aperture radar (SAR) images have characteristics, such as large image proportions, rich features, and large feature differences in a single target. Existing multiscale ship detection algorithms for SAR data employ multiscale feature pyramids or anchor-free extractors to capture features from large ship targets. However, the significant feature variation due to the internal structure and reflection of ships makes it difficult for current extractors to provide a consistent feature description, often resulting in fragmented detection outcomes for ship targets. This article decomposes the large ship detection problem into the detection of individual ship components (including tail, hull, and head), proposing a novel component-based detection method for large ship targets in SAR images. The proposed method enhances the network's efficiency in feature propagation and aggregation across different layers using the generalized efficient layer aggregation network (GELAN) structure. Following the feature extraction of GELAN, a multilevel multipooling channel attention is integrated to optimize the feature extraction structure in a hierarchical manner. The method also incorporates environmental features around the target to strengthen the association between different ship components. The detected ship components are connected using a topological relationship algorithm based on the component structure, culminating in the generation of ship target detection results. Experiments on the large ship component model dataset constructed for this article demonstrate significant improvements in the proposed algorithm over the preoptimized YOLOv8. The experimental results demonstrate that our method achieves promising detection performance when compared with the current state-of-the-art you only look once series algorithms and multiscale SAR ship detection algorithms. The algorithm also effectively avoided noticeable loss or false detection of small ship targets present in the dataset.https://ieeexplore.ieee.org/document/10787532/Component modellarge ship targetsship target detectionsynthetic aperture radar (SAR)you only look once (YOLO) |
| spellingShingle | Tiancheng Dong Taoyang Wang Xuefei Li Jianzhi Hong Maoqiang Jing Tong Wei A Large Ship Detection Method Based on Component Model in SAR Images IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Component model large ship targets ship target detection synthetic aperture radar (SAR) you only look once (YOLO) |
| title | A Large Ship Detection Method Based on Component Model in SAR Images |
| title_full | A Large Ship Detection Method Based on Component Model in SAR Images |
| title_fullStr | A Large Ship Detection Method Based on Component Model in SAR Images |
| title_full_unstemmed | A Large Ship Detection Method Based on Component Model in SAR Images |
| title_short | A Large Ship Detection Method Based on Component Model in SAR Images |
| title_sort | large ship detection method based on component model in sar images |
| topic | Component model large ship targets ship target detection synthetic aperture radar (SAR) you only look once (YOLO) |
| url | https://ieeexplore.ieee.org/document/10787532/ |
| work_keys_str_mv | AT tianchengdong alargeshipdetectionmethodbasedoncomponentmodelinsarimages AT taoyangwang alargeshipdetectionmethodbasedoncomponentmodelinsarimages AT xuefeili alargeshipdetectionmethodbasedoncomponentmodelinsarimages AT jianzhihong alargeshipdetectionmethodbasedoncomponentmodelinsarimages AT maoqiangjing alargeshipdetectionmethodbasedoncomponentmodelinsarimages AT tongwei alargeshipdetectionmethodbasedoncomponentmodelinsarimages AT tianchengdong largeshipdetectionmethodbasedoncomponentmodelinsarimages AT taoyangwang largeshipdetectionmethodbasedoncomponentmodelinsarimages AT xuefeili largeshipdetectionmethodbasedoncomponentmodelinsarimages AT jianzhihong largeshipdetectionmethodbasedoncomponentmodelinsarimages AT maoqiangjing largeshipdetectionmethodbasedoncomponentmodelinsarimages AT tongwei largeshipdetectionmethodbasedoncomponentmodelinsarimages |